vvEPA
EPA Document No.
EPA-815-R-24-002
Economic Analysis for the
Final Per- and Polyfluoroalkyl Substances
National Primary Drinking Water Regulation
Appendices
-------
FINAL RULE
APRIL 2024
Economic Analysis for the Final Per- and Polyfluoroalkyl Substances National
Primary Drinking Water Regulation Appendices
Prepared by:
U.S. Environmental Protection Agency
Office of Water
Office of Ground Water and Drinking Water
Standards and Risk Management Division
Washington, DC 20460
EPA Document Number: EPA-815-R-24-002
APRIL 2024
-------
FINAL RULE
APRIL 2024
Disclaimer
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
Final PFAS Rule Economic Analysis
i
April 2024
-------
FINAL RULE APRIL 2024
Contents
Appendix A. Framework of Bayesian Hierarchical Markov Chain Monte Carlo
Occurrence Model A-l
A. IData Selection A-l
A.2Conceptual Model Structure A-4
A.3Model Implementation A-7
Appendix B. Affected Population B-l
Appendix C. Cost Analysis Results C-l
C.lPWS-Level Cost Details C-l
C.l.l Mean Annual Cost for all Community Water Systems C-l
C.1.2 Mean Annual Cost for all Non-Transient Non-Community Water Systems .... C-5
C.1.3 Mean Annual Cost for Community Water Systems that Treat or Change
Water Source C-10
C.1.4 Mean Annual Cost for Non-Transient Non-Community Water Systems that
Treat or Change Water Source
C.1.5 Distribution of Small Community Water System Costs
C.1.6 Distribution of Small Non-Community Non-Transient Water System Costs.
C.1.7 Distribution of Small Community Water System Costs that Treat or Change
Water Source
C.1.8 Distribution of Small Non-Community Water Non-Transient System Costs
that Treat or Change Water Source C-30
C.2Household-Level Cost Details C-34
C.2.1 Household Costs for all Community Water Systems C-34
C.2.2 Household Costs for Community Water Systems that Treat or Change Water
Source C-39
Appendix D. PFOA and PFOS Serum Concentration-Birth Weight Relationship D-l
D. 1 Weight of Evidence of Birth Weight Effects D-l
D.2Review of Available Meta-Analyses D-l
D.3Exposure-Response Functions Based on Epidemiological Studies D-9
Appendix E. Effects of Reduced Birth Weight on Infant Mortality E-l
E. 1 Birth Weight-Mortality Relationship E-l
E.2 Basis for Updated Birth Weight-Mortality Relationship E-3
E.3 Development of the Analytical Dataset E-5
E.3.1 Data Sources E-5
E.3.2 Dataset Development E-6
E.3.3 Identification of Infant Mortality Risk Factors E-6
E.4 Development of Variables E-7
E.5 Summary Statistics E-ll
E.6Estimation Methods E-l3
Final PFAS Rule Economic Analysis ii April 2024
C-14
C-18
C-22
C-26
-------
FINAL RULE
APRIL 2024
E.7Results and Discussion E-14
E.7.1 Mortality Regression Models E-14
E.7.2 Comparison to Prior Studies E-21
E.8 Limitations and Uncertainties E-21
Appendix F. Serum Cholesterol Dose Response Functions F-l
F.l Data Sources F-l
F.l.l Literature Review and Studies Identification for the Meta-Analysis F-l
F.l.2 Assessment of Study Applicability to the Meta-Analysis F-2
F. 2 Meta-Analy si s F - 8
F.3 Extraction of Slope Values for TC and HDLC F-8
F.4 Methods and Key Assumptions F-9
F.4.1 Slope Estimation for PFOA F-10
F.4.2 Slope Estimation for PFOS F-16
F.4.3 Sensitivity Analyses F-22
F.4.4 Limitations and Uncertainties F-22
Appendix G. CVD Benefits Model Details and Input Data G-l
G. 1 Model Overview and Notation G-l
G.2Hard CVD Event Incidence Estimation G-8
G.2.1 Probability of the First Hard CVD Event G-8
G.2.2 Prevalence of Past Hard CVD Events G-l 1
G.2.3 Distribution of Fatal and Non-Fatal First Hard CVD Events G-13
G.2.4 Post-Acute CVD Mortality G-16
G.2.5 Survivors of the First Hard CVD Event at Ages 40-65 G-l 8
G.2.6 Survivors of the First Hard CVD Event at Ages 66+ G-20
G.3Detailed CVD Model Calculations G-21
G.3.1 Baseline Recurrent Calculations Without Explicit Treatment of the CVD
Population G-22
G.3.2 Baseline Recurrent Calculations with Explicit Treatment of the CVD
Population G-22
G.3.3 Regulatory Alternative Recurrent Calculations with Explicit Treatment of
the CVD Population G-25
G.3.4 Recurrent Estimation of Post-Acute CVD Mortality G-27
G.3.5 Risk Reduction Calculations G-28
G.4ASCVD Model Validation G-29
G.5CVD Model Inputs G-30
Appendix H. Cancer Benefits Model Details and Input Data H-l
H. IDetails on the Cancer Life Table Approach H-l
H. 1.1 Evolution of Model Population (B,A) under Baseline Pollutant Exposure H-3
H. 1.2 Evolution of Model Population (B,A) under the Regulatory Alternative
Pollutant Exposure H-6
H. 1.3 Health Effects and Benefits Attributable to the Regulatory Alternatives H-6
Final PFAS Rule Economic Analysis
April 2024
-------
FINAL RULE
APRIL 2024
H.2Cancer Life Table Model Input Data H-7
H.3Baseline Kidney Cancer Statistics H-9
H.4Baseline Bladder Cancer Statistics H-36
H.5Baseline Liver Cancer Statistics H-42
H.6RCC Valuation Data H-63
Appendix I. Trihalomethane Co-Removal Model Details and Analysis 1-1
I.1 Data Analysis 1-1
1.2 Discussion of Other Models 1-5
1.3 THM4 Reduction Results 1-6
1.4 Sampling Points from the Fourth Six Year Review Plants with Granular Activated
Carbon Treatment 1-0
Appendix J. Value of a Statistical Life Updating J-l
Appendix K. Benefits Sensitivity Analyses K-l
K. 1 Overview of the Hypothetical Exposure Reduction K-l
K.2Estimation of Blood Serum PFOA, PFOS, and PFNA K-2
K.3CVD Sensitivity Analyses K-4
K.4Birth Weight Sensitivity Analyses K-8
K.6RCC Sensitivity Analyses K-l 1
Appendix L. Uncertainty Characterization Details and Input Data L-l
L.l Cost Analysis Uncertainty Characterization L-l
L.l.l Total Organic Carbon Concentration Uncertainty L-l
L.l.2 Compliance Technology Unit Cost Curve Selection Uncertainty L-l
L.2 Benefits Analysis Uncertainty Characterization L-2
L.2.1 Exposure-Response Function Uncertainty L-3
L.2.2 Population Attributable Fraction Uncertainty L-0
Appendix M. Environmental Justice M-l
M. 1 Demographic Profile of Category 4 and 5 PWS Service Areas M-l
M.2 Exposure Analysis Results M-5
M.2.1 Baseline Scenario M-5
M.2.2Hypothetical Regulatory Scenario #1: UCMR 5 MRLs M-9
M.2.3Hypothetical Regulatory Scenario #2: 10.0 ppt M-10
Appendix N. Supplemental Cost Analyses N-l
N.lCost Analysis for Very Large Systems N-l
N.2Hazardous Waste Disposal Cost Impacts N-2
N.3National Level Sensitivity Analysis of Incremental Treatment Cost of PFNA, PFBS
and HFPO-DA N-4
N.4National Level Sensitivity Analysis Considering PFNA and HFPO-DA MCLs N-6
Appendix O. Supplemental Benefits Analyses O-l
0.1 Supplemental Liver Cancer Analysis O-l
0.1.1 Overview of the Liver Cancer Risk Reduction Analysis O-l
Final PFAS Rule Economic Analysis iv April 2024
-------
FINAL RULE
APRIL 2024
0.1.2 Liver Cancer Exposure-Response Modeling 0-2
0.1.3 Estimation of Liver Cancer Risk Reductions 0-3
0.1.4 Valuation of Liver Cancer Risk Reductions 0-4
0.1.5 Results 0-7
0.1.6 Limitations and Uncertainties 0-8
0.2Supplemental Analysis Using Willingness to Pay for Cancer Morbidity Risk
Reductions 0-11
Appendix P. Additional Model Outputs P-l
P. 1 Total Estimated Benefits and Costs P-2
P.2 National Annualized Costs P-4
P.3 National Annualized Benefits P-8
P.3.1 National Birth Weight Benefits P-10
P.3.2 National CVD Benefits P-12
P.3.3 National RCC Benefits P-l4
P.3.4 National Bladder Cancer Benefits P-16
P.4 Comparison of Costs and Benefits P-l 8
P.5 Benefits Sensitivity Analyses P-23
P.6 Supplemental Cost Analyses P-26
P.7 Supplemental Benefits Analyses P-27
P.8 Undiscounted Benefits and Costs P-29
Appendix Q. Appendix References Q-l
Final PFAS Rule Economic Analysis v April 2024
-------
FINAL RULE
APRIL 2024
List of Tables
Table A-l: System and Sample Counts for Contributions to the Supplemental State Dataset
by State A-3
Table B-l: Summary of Inputs and Data Sources Used to Estimate Affected Population B-2
Table C-l: Mean Annualized Cost per CWSs, Final Rule (PFOA and PFOS MCLs of 4.0
ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) (Commercial Cost
of Capital, $2022) C-l
Table C-2: Mean Annualized Cost per CWSs, Option la (PFOA and PFOS MCLs of 4.0
ppt) (Commercial Cost of Capital, $2022) C-2
Table C-3: Mean Annualized Cost per CWSs, Option lb (PFOA and PFOS MCLs of 5.0
ppt) (Commercial Cost of Capital, $2022) C-3
Table C-4: Mean Annualized Cost per CWSs, Option lc (PFOA and PFOS MCLs of 10.0
ppt) (Commercial Cost of Capital, $2022) C-4
Table C-5: Mean Annualized Cost per NTNCWS, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) (Commercial
Cost of Capital, $2022) C-5
Table C-6: Mean Annualized Cost per NTNCWS, Option la (PFOA and PFOS MCLs of
4.0 ppt) (Commercial Cost of Capital, $2022) C-6
Table C-7: Mean Annualized Cost per NTNCWS, Option lb (PFOA and PFOS MCLs of
5.0 ppt) (Commercial Cost of Capital, $2022) C-8
Table C-8: Mean Annualized Cost per NTNCWS, Option lc (PFOA and PFOS MCLs of
10.0 ppt) (Commercial Cost of Capital, $2022) C-9
Table C-9: Mean Annualized Cost per CWSs that Treat or Change Water Source, Final
Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt
each and HI of 1) (Commercial Cost of Capital, $2022) C-10
Table C-10: Mean Annualized Cost per CWSs that Treat or Change Water Source, Option
la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022) C-l 1
Table C-l 1: Mean Annualized Cost per CWSs that Treat or Change Water Source, Option
lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022) C-12
Table C-12: Mean Annualized Cost per CWSs that Treat or Change Water Source, Option
lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022) C-13
Table C-13: Mean Annualized Cost per NTNCWSs that Treat or Change Water Source,
Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of
10 ppt each and HI of 1) (Commercial Cost of Capital, $2022) C-14
Table C-14: Mean Annualized Cost per NTNCWSs that Treat or Change Water Source,
Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022) C-15
Final PFAS Rule Economic Analysis
vi
April 2024
-------
FINAL RULE
APRIL 2024
Table C-15: Mean Annualized Cost per NTNCWSs that Treat or Change Water Source,
Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022) C-16
Table C-16: Mean Annualized Cost per NTNCWSs that Treat or Change Water Source,
Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022) C-17
Table C-17: Distribution of Annualized Cost for Small CWSs, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022) C-18
Table C-18: Distribution of Annualized Cost for Small CWSs, Option la (PFOA and PFOS
MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022) C-19
Table C-19: Distribution of Annualized Cost for Small CWSs, Option lb (PFOA and PFOS
MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022) C-20
Table C-20: Distribution of Annualized Cost for Small CWSs, Option lc (PFOA and PFOS
MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022) C-21
Table C-21: Distribution of Annualized Cost for Small NTNCWSs, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022) C-22
Table C-22: Distribution of Annualized Cost for Small NTNCWSs, Option la (PFOA and
PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022) C-23
Table C-23: Distribution of Annualized Cost for Small NTNCWSs, Option lb (PFOA and
PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022) C-24
Table C-24: Distribution of Annualized Cost for Small NTNCWSs, Option lc (PFOA and
PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022) C-25
Table C-25: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA
MCLs of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022) C-26
Table C-26: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital,
$2022) C-27
Table C-27: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital,
$2022) C-28
Table C-28: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital,
$2022) C-29
Table C-29: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-
DA MCLs of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022) C-30
Final PFAS Rule Economic Analysis
vii
April 2024
-------
FINAL RULE
APRIL 2024
Table C-30: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital,
$2022) C-31
Table C-31: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of
Capital, $2022) C-32
Table C-32: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of
Capital, $2022) C-33
Table C-33: Mean Annualized Cost per Household in CWSs, Final Rule (PFOA and PFOS
MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022) C-34
Table C-34: Mean Annualized Cost per Household in CWSs, Option la (PFOA and PFOS
MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022) C-35
Table C-35: Mean Annualized Cost per Household in CWSs, Option lb (PFOA and PFOS
MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022) C-36
Table C-36: Mean Annualized Cost per Household in CWSs, Option lc (PFOA and PFOS
MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022) C-38
Table C-37: Mean Annualized Cost per Household in CWSs that Treat or Change Water
Source, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA
MCLs of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022) C-39
Table C-38: Mean Annualized Cost per Household in CWSs that Treat or Change Water
Source, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital,
$2022) C-40
Table C-39: Mean Annualized Cost per Household in CWSs that Treat or Change Water
Source, Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital,
$2022) C-41
Table C-40: Mean Annualized Cost per Household in CWSs that Treat or Change Water
Source, Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital,
$2022) C-42
Table D-l: Data Sources for PFOA/PFOS Meta-Analyses of Birth Weight Effects D-4
Table E-l: Comparison of Sample Means for Singletons between the 1989 Natality-
Mortality Detail File and the Combined 2016-2018 Period/Cohort Linked Birth-Infant
Death Data Files E-4
Table E-2: Variables Used in Singleton Mortality Regression Analysis E-8
Table E-3: Maternal and Infant Characteristics of the Study Population E-12
Table E-4: Odds Ratios and Marginal Effects for the Non-Hispanic Black, Non-Hispanic
White, and Hispanic Mortality Regression Models E-l8
Final PFAS Rule Economic Analysis viii April 2024
-------
FINAL RULE
APRIL 2024
Table E-5: Comparison of Ma et al. (2010) and the EPA Analysis E-21
Table E-6: Limitations and Uncertainties in the Analysis of the Birth Weight-Mortality
Relationship E-22
Table F-l: Studies Selected for Inclusion in the Meta-Analyses F-5
Table F-2: Results forPFOA Meta-Analyses F-l 1
Table F-3: Results for PFOS Meta-Analyses F-17
Table F-4: Limitations and Uncertainties in the Analysis of the Serum Cholesterol Dose
Response Functions F-23
Table G-l: CVD Life Table Model Elements and Notation Summary G-6
Table G-2: ASCVD Model Coefficients G-9
Table G-3: Estimated Past Hard CVD Event Prevalence per 100,000 G-12
Table G-4: Estimated First Hard CVD Event Incidence and Distribution by CVD Event
Type G-l 3
Table G-5: Probability of Hospital Death for a Hard CVD Event G-15
Table G-6: Estimated Distribution of Fatal and Non-Fatal First Hard CVD Events G-l 6
Table G-l. Post-Acute All-Cause Mortality After the First Myocardial Infarction G-l8
Table G-8: Post-Acute Mortality After the First Myocardial Infarction G-l9
Table G-9: Post-Acute CVD Mortality Following the First Myocardial Infarction and First
Ischemic Stroke in the Population Aged 66 Years or Older G-21
Table G-10: A Mapping of CVD Model Calculations by Initial Cohort Age, Current Cohort
Age, and Estimation Type G-22
Table G-l 1: Summary of ASCVD Model Validation G-30
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model G-30
Table H-l: Health Risk Model Variable Definitions H-2
Table H-2: Summary of Data Sources Used in Cancer Lifetime Risk Models H-7
Table H-3: Summary of Baseline Kidney Cancer Incidence Data Used in the Model H-10
Table H-4: Summary of Race/Ethnicity-Specific Baseline Kidney Cancer Incidence Data
Used in the Model H-l 1
Table H-5: Summary of Relative and Absolute Kidney Cancer Survival Used in the Model. H-14
Table H-6: Summary of Race/Ethni city-Specific Relative and Absolute Kidney Cancer
Survival Used in the Model H-l8
Table H-7: Summary of All-Cause and Kidney Cancer Mortality Data Used in the Model.... H-31
Final PFAS Rule Economic Analysis
ix
April 2024
-------
FINAL RULE
APRIL 2024
Table H-8: Summary of Race/Ethnicity-Specific All-Cause and Kidney Cancer Mortality
Data Used in the Model H-32
Table H-9: Summary of Baseline Bladder Cancer Incidence Data Used in the Model H-36
Table H-10: Summary of Relative and Absolute Bladder Cancer Survival Used in the
Model 11-38
Table H-l 1: Summary of All-Cause and Bladder Cancer Mortality Data Used in the Model. H-41
Table H-12: Summary of Baseline Liver Cancer Incidence Data Used in the Model H-42
Table H-13: Summary of Race/Ethni city-Specific Baseline Liver Cancer Incidence Data
Used in the Model H-43
Table H-14: Summary of Relative Liver Cancer Survival Used in the Model H-45
Table H-15: Summary of Race/Ethni city-Specific Relative Liver Cancer Survival Used in
the Model 11-48
Table H-16: Summary of All-Cause and Liver Cancer Mortality Data Used in the Model H-58
Table H-l 7: Summary of Race/Ethni city-Specific All-Cause and Liver Cancer Mortality
Data Used in the Model H-59
Table H-l8: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs H-64
Table 1-1: ICR TSD Predictions for ATHM4 Based on Disinfectant 1-7
Table 1-2: ICR TSD Predictions for ATHM4 for V2 Year GAC Replacement Based on
Disinfectant Type, EBCT, and Source Water Type 1-8
Table 1-3: ICR TSD Predictions for ATHM4 for One Year GAC Replacement Based on
Disinfectant Type, EBCT, and Source Water Type 1-9
Table 1-4: ICR TSD Predictions for ATHM4 for 1 V2 Year GAC Replacement Based on
Disinfectant Type, EBCT, and Source Water Type 1-10
Table 1-5: ICR TSD Predictions for ATHM4 for Two Year GAC Replacement Based on
Disinfectant Type, EBCT, and Source Water Type 1-11
Table 1-6: Sampling Point IDs for each PWSID were Extracted and Matched for the Years
that Represent Before/After GAC Treatment (Example: PWSID AL0000577) 1-0
Table J-l: Estimated Value of Statistical Life Series J-2
Table J-2: Summary of Inputs and Data Sources Used for Valuation J-4
Table K-l: Overview of Hypothetical Exposure Reductions K-2
Table K-2: Overview of CVD Exposure-Response Scenarios K-4
Table K-3: Exposure-Response Information for CVD Biomarkers K-5
Table K-4: Summary of CVD Sensitivity Analysis for Hypothetical Exposure Reduction 1
(PFOA+PFOS) K-6
Final PFAS Rule Economic Analysis x April 2024
-------
FINAL RULE
APRIL 2024
Table K-5: Overview of Birth Weight Exposure-Response Scenarios K-8
Table K-6: Exposure-Response Information for Birth Weight K-9
Table K-7: Summary of Birth Weight Sensitivity Analysis K-10
Table K-8: Overview of RCC Exposure-Response Scenarios K-l 1
Table K-9: Exposure-Response Information for RCC K-l 1
Table K-10: Summary of RCC Sensitivity Analysis K-12
Table L-l: Quantified Sources of Uncertainty in Benefits Estimates L-2
Table L-2: Standard Errors and Distributions for Benefits Model Exposure-Response Slope
Factors L-4
Table M-l: Number of Category 4 PWSs and Population Served by Size and State M-2
Table M-2: Number of Category 5 PWSs and Population Served by Size and State M-2
Table M-3: Population Served by Category 4 and 5 PWSs Compared to Percent of U.S.
Population by Demographic Group M-4
Table M-4: Baseline Scenario: Population Served by Category 4 and 5 PWS Service Areas
Above Baseline Thresholds and as a Percent of Total Population Served M-7
Table M-5: Average PFAS Concentrations (ppt) by Demographic Group in the Baseline,
Category 4 and 5 PWS Service Areas M-8
Table M-6: Hypothetical Regulatory Scenario #1: Demographic Breakdown of Population
Served by Category 4 and 5 PWS Service Areas Above UCMR 5 MRL and as a Percent of
Total Population Served M-ll
Table M-7: Reductions in Average PFAS Concentrations (ppt) by Demographic Group in a
Hypothetical Regulatory Scenario with Maximum Contaminant Level at the UCMR 5
MRLs, Category 4 and 5 PWS Service Areas M-12
Table M-8: Hypothetical Regulatory Scenario #2: Demographic Breakdown of Population
Served by Category 4 and 5 PWS Service Areas Above 10.0 ppt and as a Percent of Total
Population Served M-13
Table M-9: Reductions in Average PFAS Concentrations (ppt) by Demographic Group in a
Hypothetical Regulatory Scenario with Maximum Contaminant Level at 10.0 ppt, Category
4 and 5 PWS Service Areas M-14
Table N-l: Characteristics of PWSs Serving a Retail Population Greater than One Million N-l
Table N-2: Annualized PWS Treatment Cost Associated with Non-Hazardous and
Hazardous Residual Management Requirements, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, and HFPO-DA MCLs of 10 ppt each and HI of 1) (Million
$2022) N-3
Final PFAS Rule Economic Analysis
xi
April 2024
-------
FINAL RULE
APRIL 2024
Table N-3. Marginal Mean Annualized Rule Costs Associated with Individual MCLs of 10
ppt each for PFNA, HFPO-DA (Million $2022) N-6
Table 0-1. Estimated Liver Cancer Willingness to Pay Series 0-4
Table 0-2. National Liver Cancer Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt
each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1) 0-7
Table 0-3. Limitations and Uncertainties in the Analysis of Liver Cancer Benefits 0-8
Table 0-4. Estimated Bladder Cancer Willingness to Pay Series 0-11
Table 0-5. National Willingness to Pay-Based RCC Benefits, Final Rule (PFOA and PFOS
MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1) 0-14
Table 0-6. National Willingness to Pay-Based Bladder Cancer Benefits, Final Rule (PFOA
and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA of 10 ppt each and HI of 1) 0-15
Table P-l: Quantified Total National Annualized Benefits, All Options (Million $2022) P-2
Table P-2: Quantified Total National Annualized Costs, All Options (Million $2022) P-3
Table P-3: National Annualized Costs, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each,
PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) (Million $2022) P-4
Table P-4: National Annualized Costs, Option la (PFOA and PFOS MCLs of 4.0 ppt)
(Million $2022) P-5
Table P-5: National Annualized Costs, Option lb (PFOA and PFOS MCLs of 5.0 ppt)
(Million $2022) P-6
Table P-6: National Annualized Costs, Option lc (PFOA and PFOS MCLs of 10.0 ppt)
(Million $2022) P-7
Table P-7: National Annualized Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt
each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) (Million $2022) P-8
Table P-8: National Annualized Benefits, Option la (PFOA and PFOS MCLs of 4.0 ppt)
(Million $2022) P-8
Table P-9: National Annualized Benefits, Option lb (PFOA and PFOS MCLs of 5.0 ppt)
(Million $2022) P-9
Table P-10: National Annualized Benefits, Option lc (PFOA and PFOS MCLs of 10.0 ppt)
(Million $2022) P-9
Table P-l 1: National Birth Weight Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt
each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) P-10
Table P-12: National Birth Weight Benefits, Option la (PFOA and PFOS MCLs of 4.0 ppt).P-10
Table P-13: National Birth Weight Benefits, Option lb (PFOA and PFOS MCLs of 5.0 ppt). P-l 1
Table P-14: National Birth Weight Benefits, Option lc (PFOA and PFOS MCLs of 10.0
ppt) P-l 1
Final PFAS Rule Economic Analysis xii April 2024
-------
FINAL RULE
APRIL 2024
Table P-15: National CVD Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each,
PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) P-12
Table P-16: National CVD Benefits, Option la (PFOA and PFOS MCLs of 4.0 ppt) P-12
Table P-17: National CVD Benefits, Option lb (PFOA and PFOS MCLs of 5.0 ppt) P-13
Table P-18: National CVD Benefits, Option lc (PFOA and PFOS MCLs of 10.0 ppt) P-13
Table P-19: National RCC Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each,
PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) P-14
Table P-20: National RCC Benefits, Option la (PFOA and PFOS MCLs of 4.0 ppt) P-14
Table P-21: National RCC Benefits, Option lb (PFOA and PFOS MCLs of 5.0 ppt) P-15
Table P-22: National RCC Benefits, Option lc (PFOA and PFOS MCLs of 10.0 ppt) P-15
Table P-23: National Bladder Cancer Benefits, Final Rule (PFOA and PFOS MCLs of 4.0
ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) P-16
Table P-24: National Bladder Cancer Benefits, Option la (PFOA and PFOS MCLs of 4.0
ppt) P-16
Table P-25: National Bladder Cancer Benefits, Option lb (PFOA and PFOS MCLs of 5.0
ppt) P-17
Table P-26: National Bladder Cancer Benefits, Option lc (PFOA and PFOS MCLs of 10.0
ppt) P-17
Table P-27: Annualized Quantified National Costs and Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Million $2022) P-18
Table P-28: Annualized Quantified National Costs and Benefits, Option la (PFOA and
PFOS MCLs of 4.0 ppt) (Million $2022) P-21
Table P-29: Annualized Quantified National Costs and Benefits, Option lb (PFOA and
PFOS MCLs of 5.0 ppt) (Million $2022) P-21
Table P-30: Annualized Quantified National Costs and Benefits, Option lc (PFOA and
PFOS MCLs of 10.0 ppt) (Million $2022) P-22
Table P-31: Summary of CVD Sensitivity Analysis for Hypothetical Exposure Reduction 1
(PFOA+PFOS) P-23
Table P-32: Summary of Birth Weight Sensitivity Analysis P-25
Table P-33: Summary of RCC Sensitivity Analysis P-25
Table P-34: Annualized PWS Treatment Cost Associated with Non-Hazardous and
Hazardous Residual Management Requirements, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, and HFPO-DA MCLs of 10 ppt each and HI of 1) (Million
$2022) P-26
Final PFAS Rule Economic Analysis
xiii
April 2024
-------
FINAL RULE
APRIL 2024
Table P-35. National Liver Cancer Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt
each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1) P-27
Table P-36. National Willingness to Pay-Based RCC Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1) P-27
Table P-37. National Willingness to Pay-Based Bladder Cancer Benefits, Final Rule
(PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA of 10 ppt each and HI
of 1) P-28
Table P-38. Quantified Total National Annual Costs, Final Rule (Undiscounted, Million
$2022) P-29
Table P-39. Quantified Total National Annual Benefits, Final Rule (Undiscounted, Million
$2022) P-32
Final PFAS Rule Economic Analysis
xiv
April 2024
-------
FINAL RULE
APRIL 2024
List of Figures
Figure D-l: Results and Confidence Limits from PFOA, PFOS Meta-Analyses: Changes in
BW (grams) per Change in Serum PFAS Levels (ng/mL) D-8
Figure E-l: Comparison of Change in Incidence of Infant Death per 1 g Increase in Birth
Weight by Gestational Age Category and Race/Ethnicity (Deaths per 1,000 Births) E-l5
Figure F-l: Diagram of Literature Retained for Use in the Meta-Analysis and Data Sources.... F-3
Figure F-2: Forest Plots Showing the Beta Coefficients Relating PFOA Concentrations to
TC and HDLC in Each Study Reporting Linear Associations, and Pooled Estimates After
Random-Effects Meta-Analysis F-12
Figure F-3: Filled-in Funnel Plots to Evaluate Publication Bias of the PFOA and TC (Left)
or HDLC (Right) Association in Studies Reporting Linear Associations F-13
Figure F-4: Forest Plots Showing the Beta Coefficients Relating TC and HDLC to PFOA
Concentrations in Each Study, and Pooled Estimates After Random-Effects Meta-Analysis. .F-14
Figure F-5: Filled-in Funnel Plots to Evaluate Publication Bias of the PF OA and TC (Left)
or HDLC (Right) Association F-15
Figure F-6: Forest Plots Showing the Beta Coefficients Relating TC and HDLC to PFOS
Concentrations in Each Study Reporting Linear Associations, and Pooled Estimates After
Random-Effects Meta-Analy si s F-18
Figure F-7: Filled-in Funnel Plots to Evaluate Publication Bias of the PFOS and TC (Left)
or HDLC (Right) Association in Studies Reporting Linear Associations F-19
Figure F-8: Forest Plots Showing the Beta Coefficients Relating PFOS Concentrations to
TC and HDLC in Each Study, and Pooled Estimates After Random-Effects Meta-Analysis...F-20
Figure F-9: Filled-in Funnel Plots to Evaluate Publication Bias of the PFOS and TC (Left)
or HDLC (Right) Association F-21
Figure G-l: Overview of Life Table Calculations in the CVD Model G-3
Figure G-2: CVD Model Calculations Tracking CVD and Non-CVD Subpopulations for a
Specific Current Age of Cohort G-5
Figure 1-1: Example Breakthrough Curve for THM4 from the ICR Dataset with Logistic
Fit Functions Shown 1-2
Figure 1-2: Example Percent Removal Results vs. Time based on Logistic Plots Shown in
Figure 1-1 1-3
Figure 1-3: Mean Percentage Removal (Shaded Area ± 1 Standard Deviation) 1-4
Figure 1-4: Probability Density Function of Concentration Difference at 2 Years of Carbon
Life (Subdivided by TOC level) 1-5
Figure 0-1. Overview of Analysis of Reduced Liver Cancer Risk 0-2
Final PFAS Rule Economic Analysis
xv
April 2024
-------
FINAL RULE APRIL 2024
Figure P-l: Distribution of Estimated Net Quantified Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1;
3 percent Discount Rate; Million $2022) P-19
Figure P-2: Distribution of Estimated Net Quantified Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1;
7 percent Discount Rate; Million $2022) P-20
Final PFAS Rule Economic Analysis xvi April 2024
-------
FINAL RULE
APRIL 2024
Acronyms and Abbreviations
AE Adverse Events
AHRQ Agency for Healthcare Research and Quality
ANGIDX Angina, or Angina Pectoris, As Defined in the Medical Exposure Panel Survey
ASCVD Atherosclerotic Cardiovascular Disease
ATSDR Agency for Toxic Substances and Disease Registry
BE A Bureau of Economic Analysis
BIRTH Birth Characteristics
BLS Bureau of Labor Statistics
BP Blood Pressure
BW Birth Weight
CAGR Compound Annual Growth Rate
CDC Centers for Disease Control and Prevention
CHD Coronary Heart Disease
CHDDX Coronary Heart Disease, as Defined in the Medical Exposure Panel Survey
CHMS Canadian Health Measures Survey
CI Confidence Interval
COI Cost Of Illness
CPI Consumer Price Index
CVD Cardiovascular Disease
DBP Disinfection Byproduct
DS Distribution System
EA Economic Analysis
EBCT Empty Bed Contact Time
EIA Energy Information Administration
EJ Enviromnental Justice
EPA/OST U.S. Enviromnental Protection Agency Office of Science and Technology
EP Entry Point
FIPS Federal Information Processing Standards
GAC Granular Activated Carbon
GDP Gross Domestic Product
GFR Glomerular Filtration Rate
GW Ground Water
HCUP Healthcare Cost and Utilization Project
HDLC High-Density Lipoprotein Cholesterol
HESD Health Effects Support Document
HMO Health Maintenance Organization
ICR Information Collection Request
IR Incidence Ratio
IS Ischemic Stroke
KC Kidney Cancer
LBW Low Birth Weight
LCB Lower Confidence Bound
Final PFAS Rule Economic Analysis
xvii
April 2024
-------
FINAL RULE
MCL
MDEM
MEPS
MIDX
MR
mRCC
MRF
MRL
NCCN
NCHS
NHANES
NPDWR
NVSS
OGWDW
OHRTDX
OLS
OSHA
OW
PAF
PBPK
PDV
PDYPP
PFAS
PFBS
PFHxS
PFHpA
PFNA
PFOA
PFOS
PK
PPPM
PWS
PWSID
QALY
RCC
ROB
RSSCT
SAB
SD
SDWIS
SE
SEER
STRKDX
SW
APRIL 2024
Maximum Contaminant Level
Maternal Demographic and Socioeconomic Characteristics
Medical Expenditure Panel Survey
Heart Attack, or Myocardial Infarction, as Defined in the Medical Exposure Panel Survey
Point of Maximum Residence
Metastatic Renal Cell Carcinoma
Maternal Risk and Risk Mitigation Factors
Minimum Reporting Level
National Comprehensive Cancer Network
National Center for Health Statistics
National Health and Nutrition Examination Survey
National Primary Drinking Water Regulation
National Vital Statistics System
Office Of Ground Water and Drinking Water
Other Kind of Heart Disease or Condition, As Defined in the Medical Exposure Panel Survey
Ordinary Least Squares
Occupational Safety and Health Administration
Office of Water
Population Attributable Fraction
Pharmacologically Based Pharmacokinetic
Present Discounted Value
Personal Disposable Income Per Capita
Per- and Polyfluoroalkyl Substances
Perfluorobutane Sulfonic Acid
Perfluorohexanesulfonic Acid
Perfluoroheptanoic Acid
Perfluorononanoic Acid
Perfluorooctanoic Acid
Perfluorooctanesulfonic Acid
Pharmacokinetic
Per Patient Per Month
Public Water Systems
Public Water System Identification
Quality-Adjusted Life-Years
Renal Cell Carcinoma
Risk of Bias
Rapid Small-Scale Column Tests
Science Advisory Board
Standard Deviation
Safe Drinking Water Information System
Standard Error
Surveillance, Epidemiology, and End Results
Stroke Diagnosis, As Defined in the Medical Exposure Panel Survey
Surface Water
Final PFAS Rule Economic Analysis
xviii
April 2024
-------
FINAL RULE
TC
TOC
THM4
TSD
UCB
UCMR
VSL
WS
WTP
APRIL 2024
Total Cholesterol
Total Organic Carbon
Four Regulated Trihalomethanes
Treatment Study Database
Upper Confidence Bound
Unregulated Contaminant Monitoring Rule
Value of a Statistical Life
Water System Facility Point
Water Treatment Plant
Final PFAS Rule Economic Analysis
xix
April 2024
-------
FINAL RULE
APRIL 2024
Appendix A. Framework of Bayesian Hierarchical
Markov Chain Monte Carlo Occurrence Model
This appendix is adapted from Cadwallader et al. (2022) and details the Bayesian hierarchical
Markov chain Monte Carlo model developed by the EPA to estimate national occurrence of per-
and polyfluoroalkyl substances (PFAS) at public water systems (PWSs) prior to the
implementation of drinking water treatment technologies and under theoretical regulatory
scenarios (Cadwallader et al. (2022). The EPA used the occurrence model to define the universe
of PWSs that could be required to treat their drinking water to reduce PFAS levels under the
regulatory alternatives. The EPA has used similar hierarchical model structures to inform
analyses in previous regulatory actions (U.S. EPA, 2000; U.S. EPA, 2005b).
A.l Data Selection
Data collected for the third Unregulated Contaminant Monitoring Rule (UCMR 3) served as the
primary dataset for this model due to its nationally representative design. While large PWSs
included in UCMR 3 represent a census, not all small PWSs were required to monitor. Rather, a
statistically representative national sample of 800 small PWSs were selected using a population-
weighted stratified random sampling design to select small PWSs with broad geographic
distribution representative of all source water types and size categories (U.S. EPA, 2012).
Because UCMR 3 included only a sample of small systems, there is greater uncertainty in the
occurrence estimates for small systems compared to large systems.
Because there was a relatively small fraction of UCMR 3 samples with PFAS concentrations
reported above minimum reporting levels (MRLs), the EPA incorporated state PFAS monitoring
datasets to supplement UCMR 3 data in the occurrence model. These datasets, which have
generally been collected more recently than UCMR 3, generally have lower reporting limits
because the analytical methods have matured rapidly over the last 10 years, allowing laboratories
to reliably measure PFAS at concentrations approximately 3 and 30 times lower than for UCMR
3. While the model can incorporate results below reporting limits in the fitting process via
cumulative distribution functions, such results are less informative than reported values. Thus,
state datasets using lower reporting limits than those used in UCMR 3 helped to inform the
model through higher fractions of reported values. The introduction of additional state datasets
consisting of samples that were collected more recently than UCMR 3 broadened the temporal
range of data used to fit the model. The EPA anticipates that, if temporal trends are significant,
the addition of more recent state data will only bias the results towards present day.
The EPA collected state occurrence data using broad internet searches1 and downloaded publicly
available monitoring data from state government websites as of May 2023. While comprehensive
information about methods used and reporting was not fully available for all of the state
monitoring programs, the vast majority of the state data incorporated in the occurrence model
were analyzed using EPA-approved PFAS drinking water analysis methods, including EPA
Methods 533, 537, and 537.1. Of these methods, the most commonly used method was EPA
Method 537.1.
1 Search terms included "PFAS", "drinking water", "occurrence", "monitoring", and "state", or a specific state name.
Final PFAS Rule Economic Analysis
A-l
April 2024
-------
FINAL RULE
APRIL 2024
Additionally, if the state data met certain specifications, the EPA assumed that they were
statistically comparable with the UCMR 3 data and could be used to inform the national
occurrence model. In making these determinations, the EPA performed quality assurance on the
state data as they were reported and described online. The implemented quality assurance
procedures included verifying that the data utilized to inform the national model were inclusive
of finished drinking water samples only, reporting or detection limits were available for any
samples reported as below a reporting limit, perfluorooctanoic acid (PFOA),
perfluorooctanesulfonic acid (PFOS), perfluoroheptanoic acid (PFHpA), and
perfluorohexanesulfonic acid (PFHxS) were reported as individual chemical analytes, and
reported state data were for distinct state monitoring efforts (i.e., they were not also a part of
UCMR 3 monitoring). If any of this information could not be verified based on the descriptions
that states provided on their public websites or within the downloadable data, those state data
were not incorporated within the national occurrence model.
Further, the supplemental state data were limited to samples collected from systems that were
also included in UCMR 3. The purpose of this was to prevent biasing the dataset towards states
for which the data from additional PWSs were available and to maintain the nationally
representative set of systems selected for UCMR 3. Using these criteria, 28 states were identified
as having some state monitoring data to be included in fitting the national occurrence model.
These states included: Arizona, California, Colorado, Delaware, Georgia, Idaho, Illinois, Indiana,
Iowa, Kentucky, Maine, Massachusetts, Michigan, Missouri, New Hampshire, New Jersey, New
York, North Carolina, North Dakota, Ohio, Oregon, Pennsylvania, South Carolina, Tennessee,
Vermont, Virginia, West Virginia, and Wisconsin (Arizona Department of Environmental
Quality, 2021; Arizona Department of Environmental Quality, 2023; California Division of
Drinking Water, 2023; Colorado Department of Public Health and Environment, 2020; Delaware
Office of Drinking Water, 2021; Georgia Environmnetal Protection Division, 2020; Idaho
Department of Environmental Quality, 2023; Indiana Department of Environmental
Management, 2023; Kentucky Department for Environmental Protection, 2019; Maine
Department of Environmental Protection, 2020; Maine Department of Health and Human
Services, 2023; Missouri Department of Natural Resources, 2023; New Hampshire Department
of Environmental Services, 2021; New York Department of Health, 2022; North Carolina
Department of Environmental Quality, 2023; North Dakota Department of Environmental
Quality, 2020; North Dakota Department of Environmental Quality, 2021; Oregon Health
Authority, 2022; South Carolina Department of Health and Environmental Control, 2020; South
Carolina Department of Health and Environmental Control, 2023; Tennessee Department of
Environment and Conservation, 2023; Vermont Department of Environmental Conservation,
2023; Virginia Department of Health, 2021; West Virginia Department of Health and Human
Resources, 2023; Wisconsin Department of Natural Resources, 2023). According to state
websites, these state data represent samples collected between March 2016 through May 2023.
The dataset used to fit the model included all data available in the final UCMR 3 dataset for
PFOS, PFOA, PFHpA, and PFHxS2 (U.S. EPA, 2017). This amounted to 36,972 samples each
for PFOS, PFOA, and PFHpA, and 36,971 UCMR 3 samples for PFHxS. Of these four PFAS,
2 PFBS and PFNA were not included in this model because 19 reported values across the country from the primary dataset
(UCMR 3) were insufficient for fitting the national model (Cadwallader et al., 2022).
Final PFAS Rule Economic Analysis
A-2
April 2024
-------
FINAL RULE
APRIL 2024
1,114 samples had results reported at or above the UCMR 3 MRL3. The additional state datasets
included to supplement the UCMR 3 data included 18,091 PFOS samples, 18,082 PFOA
samples, 14,458 PFHpA samples, and 14,906 PFHxS samples collected at systems that were
included in UCMR 3. Of these samples, 7,156 (40%) were reported values for PFOS, 8,257
(46%) were reported values for PFOA, 4,496 (31%) were reported values for PFHpA, and 5,041
(34%) were reported values for PFHxS. The remainder were listed as being below their
respective reporting limits.
Table A-l provides information on the number of systems and samples included in each
supplemental state dataset. Reporting limits in state datasets varied both across and within
datasets but were primarily in the lower single digits in parts per trillion (ppt) for all four PFAS
included in the model, though for some samples the limits reported were as high as the UCMR 3
limits or as low as sub-1 ppt. The particularly low limits associated with some samples may be
associated with method detection limits rather than more conservative reporting limits.
Table A-l: System and Sample Counts for Contributions to the Supplemental
State Dataset by State
State
Systems
Included
PFOS Samples
PFOA Samples
PFHpA
Samples
PFHxS
Samples
AZ
4
202
201
11
11
CA
85
5372
5372
5179
5179
CO
52
95
95
95
95
DE
1
34
34
0
0
GA
1
2
2
2
2
IA
23
88
88
87
87
ID
2
7
7
7
7
IL
122
763
763
756
762
IN
8
10
10
10
10
KY
23
25
25
25
25
MA
128
3445
3446
3445
3445
ME
17
30
30
30
30
MI
61
550
528
491
520
MO
5
11
11
5
5
NC
29
99
99
0
0
ND
5
5
5
5
5
NH
20
323
323
166
318
NJ
148
5053
5061
2775
2776
NY
98
1059
1059
741
743
OH
145
232
232
0
232
OR
4
4
4
4
4
PA
51
91
91
91
91
SC
44
208
208
205
204
3 MRLs under UCMR 3 were as follows: PFOS 40 ppt; PFOA 20 ppt; PFNA 20 ppt; PFHxS 30 ppt; PFHpA 10 ppt; and PFBS 90
ppt.
Final PFAS Rule Economic Analysis A-3 April 2024
-------
FINAL RULE
APRIL 2024
Table A-l: System and Sample Counts for Contributions to the Supplemental
State Dataset by State
State
Systems
Included
PFOS Samples
PFOA Samples
PFHpA
Samples
PFHxS
Samples
TN
1
2
2
2
2
VA
9
14
14
14
14
VT
10
28
28
28
28
WI
59
308
313
284
311
WV
1
31
31
0
0
Abbreviations: PFOA - perfluorooctanoic acid; PFOS - perfluorooctanesulfonic acid; PFHpA - perfluoroheptanoic acid;
PFHxS - Periluorohexane sulfonate.
Further, there were several instances where approximate values were provided in state data when
the sample results were above a method detection limit but below the quantitation limit. In these
cases, the EPA used the reported values assuming that the uncertainty introduced by using these
values would be small in comparison to within-system variability. While certain systems may
have adapted treatment since the time that data were collected, the data included in the
occurrence model represent a best estimate of the current state of occurrence. Note that both
samples with results reported as specific measured concentrations and samples with
concentrations reported as lower than a reporting limit were used to fit the model. While the
latter helps to provide information to the model, samples providing a measured result are much
more informative.
A.2 Conceptual Model Structure
The Bayesian hierarchical model presented here uses log transformed data. Unless otherwise
noted, all of the following discussions, equations, distributions are based upon the use of PFOA,
PFOS, PFHpA, and PFHxS data that have been log transformed with the natural log.
The EPA tested several model variants. These variants all featured a hierarchical structure with a
multivariate normal distribution of system-level means and system-level normal distributions,
which were assumed to have been the parent distributions for the individual sample results. Thus,
for each variant, the EPA assumed lognormality for system-level medians as well as within-
system occurrence. Lognormality is a common assumption for environmental contaminant
concentrations and constitutes a core assumption made here (Lockwood et al., 2001; Ott, 1995).
The exploration of alternative distributions is inhibited by the large fraction of samples found
below their respective reporting limits. Similar Bayesian hierarchical model approaches have
been used in past drinking water occurrence assessments conducted by the EPA and others,
including for arsenic and Cryptosporidium parvum, two contaminants with considerable
occurrence below reporting limits (Crainiceanu et al., 2003; Lockwood et al., 2001; Ott, 1995).
Model variants differed by inclusion of parameters specific to system size (small versus large)
and source water type (ground water versus surface water). These parameters included:
independent correlation matrices, between-system standard deviations (SDs), within-system SDs,
and fixed factor shifts of system-level means. The EPA included fixed factor shifts in model
variants to allow the model to explore whether systems of certain categories (e.g., large or small,
ground water or surface water), might generally appear to have higher or lower concentrations of
Final PFAS Rule Economic Analysis
A-4
April 2024
-------
FINAL RULE
APRIL 2024
each chemical. The EPA compared these model variants using 5-fold cross validation. The EPA
selected the model that performed best in the 5-fold cross validation exercise (described below).
The EPA assumed that system-level means were distributed multivariate normally. This was
done to allow the model to fit and utilize a covariance matrix among system-level means for the
four PFAS included. Before adjustment for system-specific factors, the system-level means for
PFOS, PFOA, PFHpA, and PFHxS were assumed to be distributed as:
Equation A-l:
mUraw.i ~ MVNorm(MU, 2)
Where i is the system index and equal to 1, ..., nsys , nsys is the number of PWSs informing the
model, muraw i is a vector of length 4, with the four values indicating unadjusted system-level
means for PFOS, PFOA, PFHpA, and PFHxS. MU is a vector of length 4 providing the grand
national means for large PWSs, 1 is the covariance matrix for system-level means. 1 is related to
the correlation matrix and between-system standard deviation as shown in Equation A-2.
Equation A-2:
S = diag(aB) * fl * diag(aB)
Where oB is a vector of between-system standard deviations and Q is the correlation matrix of
system-level means for PFOS, PFOA, PFHpA, and PFHxS. For small systems, a fixed factor
shift was then applied to muraw i. This is shown in Equation A-3.
Equation A-3:
mill = muraw j + (bSM * SM[)
Here bSM is a vector of length 4 indicating an adjustment to be added to the unadjusted system
level mean (muraw i) if a system is small. SMt is a binary indicating whether system i is small
(1) or large (0). muL is a vector of length 4, with the four values indicating adjusted system-level
means for PFOS, PFOA, PFHpA, and PFHxS. Samples are then assumed to be normally
distributed according to Equation A-4: if the sample is either from a large system (serving more
than 10,000) or is a PFHpA or PFHxS sample.
Equation A-4:
jijk ~ Norm(muijk,aWjk)
Where y represents sample results and j is a sample index and equal to 1, .. ,,nsamp, where
nsamp is the total number of samples . Here i is the indicator for the system at which the sample
jijk was collected and k is an indicator for the contaminant that y,-yfcis a sample of (i.e., PFOS,
PFOA, PFHpA, or PFHxS). Thus, yijk represents the jth sample of contaminant k collected
from system i. mui k represents the k th element of muL shown in Equation A-3, ow is a vector
of length 4 providing the within-system standard deviation for each chemical included in the
model. Thus owk represents the k th element of ow.
Within-system standard deviations specific to small systems were fit for PFOS and PFOA.
°wsm replaces ow in Equation A-4 when the sample is either PFOS or PFOA collected at a small
Final PFAS Rule Economic Analysis
A-5
April 2024
-------
FINAL RULE
APRIL 2024
(sm) system. Model variants that included within-system standard deviations specific to small
systems for all 4 chemicals as well as no within-system standard deviations specific to small
systems were both included in the cross-validation model comparison, but both were
outperformed by the model presented here. The limited reported values of PFHxS and PFHpA at
small systems relative to PFOS and PFOA made the fitting of within-system standard deviations
specific to small systems highly uncertain for these chemicals and adversely affected the model's
predictive performance. Because of this, the EPA used within-system standard deviations pooled
across both system size categories for PFHxS and PFHpA.
Final PFAS Rule Economic Analysis
A-6
April 2024
-------
FINAL RULE
APRIL 2024
A.3 Model Implementation
The EPA conducted the data import, model setup, and assessment of model output using the R
programming language and the RStudio IDE (R Core Team, 2021; RStudio Team, 2020). The
agency used Rstan to access the Stan probabilistic programming language and execute the model
(Stan Development Team, 2020; Stan Development Team, 2021). The R packages reshape2 and
dplyr were used for data handling (Wickham, 2007; Wickham et al., 2020). The R packages
bayesplot, ggplot, and ggpubr were used for data visualization (Gabry & Mahr, 2020;
Kassambara, 2020; Wickham, 2016).
Stan uses Hamiltonian Monte Carlo No-U-Turn-Sampling for Markov chain Monte Carlo. The
EPA ran models with 4 chains of 5,000 iterations, 2,000 of which were warmup, thinned by 3.
Thinning was used to balance memory limitations with desired effective sample size. Additional
sampler parameters included: adapt delta = 0.95, max treedepth = 12, and seed= 1337. The
EPA used Shinystan (Gabry et al., 2018) to confirm that the effective sample size exceeded
1,000 for all parameters that were not predefined values, such as the diagonal of a correlation
matrix, which is 1 by definition. The EPA also used Shinystan to confirm chain mixing. No
divergent samples were observed.
For samples that were reported values (i.e., observed), the log probability was incremented using
the log of the normal density for the reported value given the system-level mean and within-
system deviation. For samples reporting the result as below the reporting limit rather than an
observed value, the log probability was incremented as the log of the cumulative normal
distribution at the reporting limit given the system-level mean and within-system standard
deviation.
The EPA optimized the model via non-centered parameterization and Cholesky factorization of
the multivariate normal distribution. Additional information on handling of samples below a
reporting limit and model reparameterization are available in the Stan User's Guide sections on
"Censored data" and "Reparameterization", respectively (Stan Development Team, 2021). The
EPA used weakly informative prior distributions. Prior distributions serve to reflect probabilistic
beliefs for model parameters prior to seeing data. The decision to use weakly informative priors
allowed for the improvement of computational efficiency by providing loose guidance towards
sensical values for model parameters without influencing posterior distributions in any
substantive matter.
Final PFAS Rule Economic Analysis
A-7
April 2024
-------
FINAL RULE
APRIL 2024
Appendix B. Affected Population
This appendix describes the data sources used to evaluate the population potentially affected by
human health risk reductions due to reductions in drinking water exposure to PFAS. Table B-l
describes the data elements used to assess the affected population in the EPA's analysis of the
benefits of reducing PFAS levels in drinking water. These elements include the Safe Drinking
Water Information System (SDWIS) 2021 quarter 4 (Q4) dataset (U.S. EPA, 2021b), and U.S.
Census Bureau (2020).
The SDWIS/Fed dataset provides information reported by states on drinking water systems, as
required by the Safe Drinking Water Act. The dataset generally includes information on system
name, identification number (public water system [PWS] ID), the cities or counties served, the
number of people served, the type of system (community, transient, or non-transient), whether
the system operates year-round or seasonally, and characteristics of the system's source water.
The U.S. Census provides detailed county-level population data by 5-year age-range, sex, race,
and ethnicity from 2010 to 2019. The EPA first calculated, for each county, the average
population for each age-range/sex/race/ethnicity cohort over this 10-year period to determine a
"typical-year" demographic distribution for each county. The EPA then calculated the proportion
of each county's population in each age-range/sex/race/ethnicity cohort in each of the 10-years.
Finally, the EPA estimated the proportion of each county's population in each
age/sex/race/ethnicity cohorts by equally distributing the population in each 5-year age-range
equally over the five years.
To determine the population proportions for each PWS, the EPA took the following steps:
1. For PWSs for which the EPA had information on the boundary of the PWS service area (see
Chapter 9):
a. Calculate the population-weighted proportion of the PWS's service area in each
county.
b. Use the values from (a) as weights, along with the county-level age-specific
sex/race/ethnicity population cohort data, to estimate the PWS's population served in
each age/sex/race/ethnicity cohort.
2. For PWSs for which the EPA did not have information on the boundary of the PWS service
area:
a. Developed a crosswalk between the primary SDWIS county name and the county
Federal Information Processing Standards (FIPS) codes used by the US. Census.
b. Used the PWS primary county age/sex/race/ethnicity population cohort data to
determine the PWS's population served in each age/sex/race/ethnicity cohort.
3. For PWSs for which the EPA did not have information on the boundary or the primary
county:
Final PFAS Rule Economic Analysis
B-l
April 2024
-------
FINAL RULE
APRIL 2024
a. Used national age/sex/race/ethnicity population cohort data to determine the PWS's
population served in each age/sex/race/ethnicity cohort.
Table B-l: Summary of Inputs and Data Sources Used to Estimate Affected Population
Data Element Modeled Variability Data Source Notes
Public water system inventory from the EPA's
SDWIS Q4 in 2021. The EPA uses the SDWIS
2021 population data as the initial total
population per PWS.
The original data source contains total
population by race/ethnicity, sex, and 5-year age
groups.
July 1,2019.
Abbreviations: PWS - public water system; SDWIS - Safe Drinking Water Information System.
Final PFAS Rule Economic Analysis B-2 April 2024
Initial Total
Population
Percentage of
Population in a
Demographic
Population
Subgroup
Location: PWS
Age: integer ages 0-84,
85+
Sex: males, females
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black,
Hispanic, other
Location: U.S. counties
SDWIS 2021
(U.S. EPA,
2021b)
U.S. Census
Bureau (2020):
Annual County
Resident
Population
Estimates by
Age,
Sex, Race, and
Hispanic Origin:
April 1, 2010 to
-------
FINAL RULE
APRIL 2024
Appendix C. Cost Analysis Results
This appendix provides additional cost output details. Section C.l provides PWS-level costs by
system type, primary source water, ownership, and system size category. Costs are provided for
all systems as well as for only those systems that must treat or change water source to comply
with the regulatory option. Section C.2 provides estimates of household costs.
C.l PWS-Level Cost Details
Section C.l provides PWS-level costs by system type, primary source water, ownership, and
system size category. Costs are provided for all systems as well as for only those systems that
must treat or change water source to comply with the regulatory option.
C.l.l Mean Annual Cost for all Community Water Systems
Table C-l: Mean Annualized Cost per CWSs, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$939
$1,267
$1,641
Private
Ground
100 to 500
$1,501
$2,061
$2,725
Private
Ground
500 to 1,000
$2,555
$3,691
$5,029
Private
Ground
1,000 to 3,300
$4,422
$6,565
$9,005
Private
Ground
3,300 to 10,000
$9,947
$17,274
$25,321
Private
Ground
10,000 to 50,000
$124,300
$154,480
$187,950
Private
Ground
50,000 to 100,000
$220,850
$408,390
$634,090
Private
Ground
100,000 to 1,000,000
$387,230
$684,490
$1,114,300
Private
Surface
Less than 100
$958
$1,487
$2,089
Private
Surface
100 to 500
$1,486
$2,238
$3,012
Private
Surface
500 to 1,000
$1,970
$3,701
$5,671
Private
Surface
1,000 to 3,300
$3,260
$6,293
$9,746
Private
Surface
3,300 to 10,000
$7,796
$16,964
$28,072
Private
Surface
10,000 to 50,000
$103,870
$132,270
$162,860
Private
Surface
50,000 to 100,000
$288,750
$400,440
$522,370
Private
Surface
100,000 to 1,000,000
$1,785,000
$2,089,900
$2,416,800
Public
Ground
Less than 100
$929
$1,333
$1,792
Public
Ground
100 to 500
$1,701
$2,389
$3,181
Public
Ground
500 to 1,000
$2,844
$4,057
$5,426
Public
Ground
1,000 to 3,300
$5,456
$7,887
$10,578
Public
Ground
3,300 to 10,000
$15,003
$21,291
$27,664
Public
Ground
10,000 to 50,000
$160,790
$176,300
$193,730
Public
Ground
50,000 to 100,000
$329,880
$411,810
$495,280
Public
Ground
100,000 to 1,000,000
$1,185,700
$1,501,800
$1,879,400
Public
Surface
Less than 100
$1,056
$1,667
$2,334
Public
Surface
100 to 500
$1,823
$2,582
$3,463
Final PFAS Rule Economic Analysis
C-l
April 2024
-------
FINAL RULE
APRIL 2024
Table C-l: Mean Annualized Cost per CWSs, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Public
Surface
500 to 1,000
$2,785
$4,196
$5,825
Public
Surface
1,000 to 3,300
$5,442
$7,815
$10,645
Public
Surface
3,300 to 10,000
$15,106
$21,231
$28,498
Public
Surface
10,000 to 50,000
$135,520
$147,870
$160,150
Public
Surface
50,000 to 100,000
$277,760
$320,770
$366,610
Public
Surface
100,000 to 1,000,000
$786,610
$906,230
$1,036,800
Abbreviations: CWS - community water system.
Table C-2: Mean Annualized Cost per CWSs, Option la (PFOA and PFOS MCLs of
4.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$938
$1,266
$1,642
Private
Ground
100 to 500
$1,498
$2,059
$2,730
Private
Ground
500 to 1,000
$2,553
$3,685
$4,977
Private
Ground
1,000 to 3,300
$4,450
$6,553
$9,191
Private
Ground
3,300 to 10,000
$9,615
$17,224
$25,362
Private
Ground
10,000 to 50,000
$122,760
$152,720
$186,520
Private
Ground
50,000 to 100,000
$202,940
$383,200
$611,860
Private
Ground
100,000 to 1,000,000
$381,440
$676,770
$1,097,600
Private
Surface
Less than 100
$958
$1,486
$2,104
Private
Surface
100 to 500
$1,519
$2,235
$3,011
Private
Surface
500 to 1,000
$1,970
$3,695
$5,667
Private
Surface
1,000 to 3,300
$3,260
$6,283
$9,634
Private
Surface
3,300 to 10,000
$8,034
$16,923
$27,898
Private
Surface
10,000 to 50,000
$103,170
$131,580
$164,420
Private
Surface
50,000 to 100,000
$289,620
$398,790
$516,720
Private
Surface
100,000 to 1,000,000
$1,731,300
$2,038,300
$2,366,900
Public
Ground
Less than 100
$928
$1,332
$1,797
Public
Ground
100 to 500
$1,713
$2,387
$3,150
Public
Ground
500 to 1,000
$2,824
$4,052
$5,391
Public
Ground
1,000 to 3,300
$5,587
$7,873
$10,569
Public
Ground
3,300 to 10,000
$14,794
$21,231
$28,281
Public
Ground
10,000 to 50,000
$159,170
$175,200
$192,570
Public
Ground
50,000 to 100,000
$326,650
$408,980
$494,290
Public
Ground
100,000 to 1,000,000
$1,136,300
$1,466,200
$1,854,700
Public
Surface
Less than 100
$1,055
$1,665
$2,364
Final PFAS Rule Economic Analysis
C-2
April 2024
-------
FINAL RULE
APRIL 2024
Table C-2: Mean Annualized Cost per CWSs, Option la (PFOA and PFOS MCLs of
4.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Public
Surface
100 to 500
$1,821
$2,580
$3,461
Public
Surface
500 to 1,000
$2,785
$4,191
$5,823
Public
Surface
1,000 to 3,300
$5,432
$7,805
$10,483
Public
Surface
3,300 to 10,000
$14,773
$21,198
$28,582
Public
Surface
10,000 to 50,000
$135,360
$147,320
$160,520
Public
Surface
50,000 to 100,000
$277,130
$318,760
$362,860
Public
Surface
100,000 to 1,000,000
$779,220
$899,290
$1,031,900
Abbreviations: CWS - community water system.
Table C-3: Mean Annualized Cost per CWSs, Option lb (PFOA and PFOS MCLs of
5.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$718
$967
$1,263
Private
Ground
100 to 500
$1,131
$1,554
$2,044
Private
Ground
500 to 1,000
$1,809
$2,737
$3,777
Private
Ground
1,000 to 3,300
$3,099
$4,810
$6,784
Private
Ground
3,300 to 10,000
$6,652
$12,398
$19,600
Private
Ground
10,000 to 50,000
$92,034
$117,230
$144,560
Private
Ground
50,000 to 100,000
$137,230
$283,530
$460,060
Private
Ground
100,000 to 1,000,000
$233,870
$460,410
$801,970
Private
Surface
Less than 100
$734
$1,157
$1,654
Private
Surface
100 to 500
$1,167
$1,714
$2,353
Private
Surface
500 to 1,000
$1,401
$2,759
$4,482
Private
Surface
1,000 to 3,300
$2,204
$4,595
$7,487
Private
Surface
3,300 to 10,000
$4,846
$12,161
$21,309
Private
Surface
10,000 to 50,000
$77,607
$100,900
$126,000
Private
Surface
50,000 to 100,000
$225,330
$321,250
$428,230
Private
Surface
100,000 to 1,000,000
$1,366,200
$1,648,800
$1,926,000
Public
Ground
Less than 100
$719
$1,013
$1,381
Public
Ground
100 to 500
$1,258
$1,786
$2,387
Public
Ground
500 to 1,000
$2,022
$2,978
$4,065
Public
Ground
1,000 to 3,300
$3,901
$5,734
$7,856
Public
Ground
3,300 to 10,000
$10,432
$15,276
$20,537
Public
Ground
10,000 to 50,000
$124,670
$138,230
$152,390
Public
Ground
50,000 to 100,000
$249,490
$318,880
$391,610
Public
Ground
100,000 to 1,000,000
$933,800
$1,203,700
$1,526,500
Final PFAS Rule Economic Analysis
C-3
April 2024
-------
FINAL RULE
APRIL 2024
Table C-3: Mean Annualized Cost per CWSs, Option lb (PFOA and PFOS MCLs of
5.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Public
Surface
Less than 100
$794
$1,305
$1,869
Public
Surface
100 to 500
$1,339
$1,953
$2,637
Public
Surface
500 to 1,000
$2,039
$3,111
$4,422
Public
Surface
1,000 to 3,300
$3,702
$5,653
$7,798
Public
Surface
3,300 to 10,000
$10,766
$15,438
$21,019
Public
Surface
10,000 to 50,000
$103,590
$113,280
$123,480
Public
Surface
50,000 to 100,000
$202,770
$237,580
$272,480
Public
Surface
100,000 to 1,000,000
$580,900
$680,330
$788,060
Abbreviations: CWS - community water system.
Table C-4: Mean Annualized Cost per CWSs, Option lc (PFOA and PFOS MCLs of
10.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$351
$443
$556
Private
Ground
100 to 500
$510
$673
$870
Private
Ground
500 to 1,000
$710
$1,069
$1,512
Private
Ground
1,000 to 3,300
$1,031
$1,748
$2,638
Private
Ground
3,300 to 10,000
$1,519
$4,104
$7,570
Private
Ground
10,000 to 50,000
$29,721
$42,855
$58,605
Private
Ground
50,000 to 100,000
$23,945
$80,063
$172,570
Private
Ground
100,000 to 1,000,000
$9,293
$78,509
$201,680
Private
Surface
Less than 100
$398
$579
$826
Private
Surface
100 to 500
$554
$800
$1,091
Private
Surface
500 to 1,000
$509
$1,153
$2,015
Private
Surface
1,000 to 3,300
$669
$1,697
$3,195
Private
Surface
3,300 to 10,000
$722
$3,874
$8,528
Private
Surface
10,000 to 50,000
$26,425
$38,368
$52,390
Private
Surface
50,000 to 100,000
$96,644
$154,310
$219,820
Private
Surface
100,000 to 1,000,000
$534,870
$702,270
$888,110
Public
Ground
Less than 100
$339
$463
$631
Public
Ground
100 to 500
$547
$736
$970
Public
Ground
500 to 1,000
$746
$1,106
$1,538
Public
Ground
1,000 to 3,300
$1,321
$1,981
$2,753
Public
Ground
3,300 to 10,000
$2,908
$4,826
$7,007
Public
Ground
10,000 to 50,000
$50,563
$57,131
$64,403
Final PFAS Rule Economic Analysis
C-4
April 2024
-------
FINAL RULE
APRIL 2024
Table C-4: Mean Annualized Cost per CWSs, Option lc (PFOA and PFOS MCLs of
10.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Public
Ground
50,000 to 100,000
$97,827
$134,570
$174,320
Public
Ground
100,000 to 1,000,000
$430,830
$589,690
$757,930
Public
Surface
Less than 100
$408
$605
$906
Public
Surface
100 to 500
$593
$842
$1,148
Public
Surface
500 to 1,000
$743
$1,198
$1,739
Public
Surface
1,000 to 3,300
$1,151
$1,872
$2,741
Public
Surface
3,300 to 10,000
$3,140
$4,891
$7,248
Public
Surface
10,000 to 50,000
$38,452
$43,249
$48,396
Public
Surface
50,000 to 100,000
$63,513
$79,507
$96,985
Public
Surface
100,000 to 1,000,000
$211,710
$257,300
$310,390
Abbreviations: CWS - community water system.
C.1.2 Mean Annual Cost for all Non-Transient Non-Community
Water Systems
Table C-5: Mean Annualized Cost per NTNCWS, Final Rule (PFOA and PFOS MCLs
of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$1,062
$1,425
$1,830
Private
Ground
100 to 500
$1,471
$2,053
$2,687
Private
Ground
500 to 1,000
$2,287
$3,463
$4,914
Private
Ground
1,000 to 3,300
$3,260
$5,798
$8,707
Private
Ground
3,300 to 10,000
$2,630
$14,634
$31,213
Private
Ground
10,000 to 50,000
$255
$83,271
$373,990
Private
Surface
Less than 100
$897
$1,574
$2,399
Private
Surface
100 to 500
$1,298
$2,464
$3,792
Private
Surface
500 to 1,000
$1,171
$4,275
$7,999
Private
Surface
1,000 to 3,300
$2,078
$6,699
$13,201
Private
Surface
3,300 to 10,000
$3,041
$21,674
$47,898
Private
Surface
10,000 to 50,000
$13,117
$105,950
$228,120
Private
Surface
100,000 to 1,000,000
$485
$406,490
$2,626,200
Public
Ground
Less than 100
$1,010
$1,463
$1,960
Public
Ground
100 to 500
$1,604
$2,277
$3,085
Public
Ground
500 to 1,000
$2,197
$3,504
$4,807
Public
Ground
1,000 to 3,300
$3,660
$6,348
$9,589
Final PFAS Rule Economic Analysis C-5 April 2024
-------
FINAL RULE
APRIL 2024
Table C-5: Mean Annualized Cost per NTNCWS, Final Rule (PFOA and PFOS MCLs
of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Public
Ground
3,300 to 10,000
$520
$18,575
$43,571
Public
Ground
10,000 to 50,000
$76,609
$178,600
$318,660
Public
Surface
Less than 100
$649
$1,639
$2,937
Public
Surface
100 to 500
$1,123
$2,706
$4,634
Public
Surface
500 to 1,000
$460
$3,887
$9,499
Public
Surface
1,000 to 3,300
$1,880
$9,134
$19,778
Public
Surface
3,300 to 10,000
$673
$21,796
$53,103
Public
Surface
10,000 to 50,000
$1,058
$116,200
$287,200
Public
Surface
50,000 to 100,000
$320
$164,980
$813,410
Abbreviations: NTNCWS -
non-transient, non-community water systems.
Table C-6: Mean Annualized Cost per NTNCWS, Option la (PFOA and PFOS MCLs
of 4.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$1,060
$1,423
$1,842
Private
Ground
100 to 500
$1,470
$2,050
$2,685
Private
Ground
500 to 1,000
$2,258
$3,458
$4,910
Private
Ground
1,000 to 3,300
$3,257
$5,791
$8,707
Private
Ground
3,300 to 10,000
$2,629
$14,610
$31,213
Private
Ground
10,000 to 50,000
$255
$83,273
$374,330
Private
Surface
Less than 100
$897
$1,573
$2,377
Private
Surface
100 to 500
$1,298
$2,461
$3,874
Private
Surface
500 to 1,000
$1,347
$4,264
$7,990
Private
Surface
1,000 to 3,300
$2,078
$6,683
$13,196
Private
Surface
3,300 to 10,000
$3,022
$21,562
$47,812
Private
Surface
10,000 to 50,000
$11,769
$105,060
$228,110
Private
Surface
100,000 to 1,000,000
$485
$406,000
$2,626,200
Public
Ground
Less than 100
$1,039
$1,461
$1,950
Public
Ground
100 to 500
$1,604
$2,275
$3,037
Public
Ground
500 to 1,000
$2,273
$3,501
$4,961
Public
Ground
1,000 to 3,300
$3,658
$6,339
$9,547
Public
Ground
3,300 to 10,000
$574
$18,542
$43,571
Public
Ground
10,000 to 50,000
$72,552
$177,870
$310,330
Public
Surface
Less than 100
$622
$1,638
$2,937
Public
Surface
100 to 500
$1,121
$2,703
$4,626
Public
Surface
500 to 1,000
$461
$3,880
$9,499
Final PFAS Rule Economic Analysis
C-6
April 2024
-------
FINAL RULE
APRIL 2024
Table C-6: Mean Annualized Cost per NTNCWS, Option la (PFOA and PFOS MCLs
of 4.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Public
Surface
1,000 to 3,300
$1,836
$9,111
$19,774
Public
Surface
3,300 to 10,000
$685
$21,710
$53,501
Public
Surface
10,000 to 50,000
$1,047
$115,670
$287,190
Public
Surface
50,000 to 100,000
$320
$164,520
$819,920
Abbreviations: NTNCWS - non-transient, non-community water systems.
Final PFAS Rule Economic Analysis
C-7
April 2024
-------
FINAL RULE
APRIL 2024
Table C-7: Mean Annualized Cost per NTNCWS, Option lb (PFOA and PFOS MCLs
of 5.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Population Served Size
5th Percentile
Mean
95th
Water
Category
Percentile
Private
Ground
Less than 100
$819
$1,091
$1,430
Private
Ground
100 to 500
$1,095
$1,546
$2,055
Private
Ground
500 to 1,000
$1,588
$2,550
$3,610
Private
Ground
1,000 to 3,300
$2,258
$4,238
$6,734
Private
Ground
3,300 to 10,000
$466
$10,549
$24,325
Private
Ground
10,000 to 50,000
$255
$64,926
$373,930
Private
Surface
Less than 100
$665
$1,230
$1,940
Private
Surface
100 to 500
$940
$1,897
$3,071
Private
Surface
500 to 1,000
$738
$3,205
$6,722
Private
Surface
1,000 to 3,300
$1,304
$4,925
$10,264
Private
Surface
3,300 to 10,000
$1,025
$15,634
$38,552
Private
Surface
10,000 to 50,000
$5,939
$81,784
$192,730
Private
Surface
100,000 to 1,000,000
$485
$281,790
$2,611,200
Public
Ground
Less than 100
$772
$1,116
$1,507
Public
Ground
100 to 500
$1,179
$1,700
$2,332
Public
Ground
500 to 1,000
$1,525
$2,567
$3,719
Public
Ground
1,000 to 3,300
$2,423
$4,607
$7,302
Public
Ground
3,300 to 10,000
$451
$13,317
$35,147
Public
Ground
10,000 to 50,000
$56,010
$143,490
$253,510
Public
Surface
Less than 100
$423
$1,268
$2,433
Public
Surface
100 to 500
$756
$2,052
$3,752
Public
Surface
500 to 1,000
$434
$2,885
$7,980
Public
Surface
1,000 to 3,300
$719
$6,669
$15,289
Public
Surface
3,300 to 10,000
$616
$15,623
$44,557
Public
Surface
10,000 to 50,000
$900
$87,432
$254,310
Public
Surface
50,000 to 100,000
$320
$119,290
$792,860
Abbreviations: NTNCWS - non-transient, non-community water systems.
Final PFAS Rule Economic Analysis
C-8
April 2024
-------
FINAL RULE
APRIL 2024
Table C-8: Mean Annualized Cost per NTNCWS, Option lc (PFOA and PFOS MCLs
of 10.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Population Served Size
5th Percentile
Mean
95th
Water
Category
Percentile
Private
Ground
Less than 100
$396
$503
$626
Private
Ground
100 to 500
$493
$655
$872
Private
Ground
500 to 1,000
$554
$944
$1,453
Private
Ground
1,000 to 3,300
$622
$1,526
$2,711
Private
Ground
3,300 to 10,000
$373
$3,354
$10,888
Private
Ground
10,000 to 50,000
$255
$25,077
$270,500
Private
Surface
Less than 100
$391
$621
$967
Private
Surface
100 to 500
$481
$881
$1,473
Private
Surface
500 to 1,000
$556
$1,357
$3,384
Private
Surface
1,000 to 3,300
$560
$1,776
$5,059
Private
Surface
3,300 to 10,000
$801
$5,493
$19,049
Private
Surface
10,000 to 50,000
$1,537
$33,245
$107,290
Private
Surface
100,000 to 1,000,000
$485
$70,346
$2,253
Public
Ground
Less than 100
$364
$514
$692
Public
Ground
100 to 500
$502
$703
$968
Public
Ground
500 to 1,000
$526
$938
$1,468
Public
Ground
1,000 to 3,300
$620
$1,569
$2,815
Public
Ground
3,300 to 10,000
$392
$4,109
$14,446
Public
Ground
10,000 to 50,000
$1,172
$60,734
$135,230
Public
Surface
Less than 100
$360
$605
$1,214
Public
Surface
100 to 500
$467
$921
$1,850
Public
Surface
500 to 1,000
$406
$1,151
$3,866
Public
Surface
1,000 to 3,300
$624
$2,378
$7,248
Public
Surface
3,300 to 10,000
$565
$5,137
$20,873
Public
Surface
10,000 to 50,000
$799
$31,077
$129,900
Public
Surface
50,000 to 100,000
$320
$28,907
$1,192
Abbreviations: NTNCWS - non-transient, non-community water systems.
Final PFAS Rule Economic Analysis
C-9
April 2024
-------
FINAL RULE
APRIL 2024
C.1.3 Mean Annual Cost for Community Water Systems that
Treat or Change Water Source
Table C-9: Mean Annualized Cost per CWSs that Treat or Change Water Source,
Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs
of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022)
Ownership
Source
Population Served Size
5th Percentile
Mean
95th
Water
Category
Percentile
Private
Ground
Less than 100
$17,069
$18,234
$20,341
Private
Ground
100 to 500
$26,064
$28,544
$32,445
Private
Ground
500 to 1,000
$42,741
$48,767
$55,660
Private
Ground
1,000 to 3,300
$70,091
$82,118
$94,266
Private
Ground
3,300 to 10,000
$154,050
$197,420
$245,690
Private
Ground
10,000 to 50,000
$406,670
$480,910
$558,610
Private
Ground
50,000 to 100,000
$821,410
$1,181,000
$1,638,200
Private
Ground
100,000 to 1,000,000
$832,170
$1,336,100
$2,082,500
Private
Surface
Less than 100
$16,415
$20,279
$24,844
Private
Surface
100 to 500
$26,031
$30,894
$36,481
Private
Surface
500 to 1,000
$38,491
$51,538
$66,115
Private
Surface
1,000 to 3,300
$65,675
$89,941
$119,330
Private
Surface
3,300 to 10,000
$138,490
$204,110
$282,990
Private
Surface
10,000 to 50,000
$462,160
$545,250
$646,910
Private
Surface
50,000 to 100,000
$955,880
$1,221,100
$1,520,900
Private
Surface
100,000 to 1,000,000
$3,434,000
$4,068,900
$4,778,600
Public
Ground
Less than 100
$17,122
$19,489
$22,283
Public
Ground
100 to 500
$30,915
$34,127
$38,672
Public
Ground
500 to 1,000
$50,185
$55,639
$62,416
Public
Ground
1,000 to 3,300
$92,430
$101,270
$111,700
Public
Ground
3,300 to 10,000
$204,820
$227,420
$250,830
Public
Ground
10,000 to 50,000
$536,600
$577,270
$624,280
Public
Ground
50,000 to 100,000
$1,059,500
$1,245,500
$1,432,900
Public
Ground
100,000 to 1,000,000
$3,193,800
$3,953,500
$4,810,300
Public
Surface
Less than 100
$17,258
$21,668
$26,782
Public
Surface
100 to 500
$32,031
$36,806
$42,508
Public
Surface
500 to 1,000
$51,310
$60,222
$69,960
Public
Surface
1,000 to 3,300
$101,320
$113,880
$127,150
Public
Surface
3,300 to 10,000
$245,060
$272,710
$300,950
Public
Surface
10,000 to 50,000
$559,260
$591,960
$627,160
Public
Surface
50,000 to 100,000
$1,045,200
$1,145,300
$1,250,900
Public
Surface
100,000 to 1,000,000
$2,444,300
$2,730,100
$3,038,100
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-10
April 2024
-------
FINAL RULE
APRIL 2024
Table C-10: Mean Annualized Cost per CWSs that Treat or Change Water Source,
Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Population Served Size
5th Percentile
Mean
95th
Water
Category
Percentile
Private
Ground
Less than 100
$17,045
$18,229
$20,315
Private
Ground
100 to 500
$26,062
$28,529
$32,248
Private
Ground
500 to 1,000
$42,668
$48,726
$55,660
Private
Ground
1,000 to 3,300
$70,579
$82,038
$94,258
Private
Ground
3,300 to 10,000
$154,500
$197,170
$245,670
Private
Ground
10,000 to 50,000
$402,560
$476,330
$559,570
Private
Ground
50,000 to 100,000
$753,860
$1,103,900
$1,554,600
Private
Ground
100,000 to 1,000,000
$831,540
$1,321,800
$2,082,000
Private
Surface
Less than 100
$16,329
$20,276
$24,844
Private
Surface
100 to 500
$26,031
$30,878
$36,318
Private
Surface
500 to 1,000
$38,391
$51,502
$66,090
Private
Surface
1,000 to 3,300
$65,916
$89,852
$119,330
Private
Surface
3,300 to 10,000
$138,690
$203,870
$284,260
Private
Surface
10,000 to 50,000
$453,760
$542,640
$644,220
Private
Surface
50,000 to 100,000
$956,050
$1,216,600
$1,515,400
Private
Surface
100,000 to 1,000,000
$3,350,100
$3,968,400
$4,676,400
Public
Ground
Less than 100
$17,105
$19,483
$22,254
Public
Ground
100 to 500
$30,913
$34,114
$38,744
Public
Ground
500 to 1,000
$50,275
$55,606
$62,017
Public
Ground
1,000 to 3,300
$92,102
$101,180
$111,050
Public
Ground
3,300 to 10,000
$203,340
$227,150
$251,310
Public
Ground
10,000 to 50,000
$532,140
$574,040
$621,150
Public
Ground
50,000 to 100,000
$1,054,300
$1,237,700
$1,424,300
Public
Ground
100,000 to 1,000,000
$3,106,900
$3,863,500
$4,706,400
Public
Surface
Less than 100
$17,121
$21,663
$26,702
Public
Surface
100 to 500
$32,031
$36,793
$42,498
Public
Surface
500 to 1,000
$51,387
$60,190
$69,800
Public
Surface
1,000 to 3,300
$101,440
$113,830
$126,210
Public
Surface
3,300 to 10,000
$245,950
$272,540
$300,750
Public
Surface
10,000 to 50,000
$556,580
$590,090
$625,700
Public
Surface
50,000 to 100,000
$1,044,100
$1,138,600
$1,243,900
Public
Surface
100,000 to 1,000,000
$2,437,300
$2,710,800
$3,009,900
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-ll
April 2024
-------
FINAL RULE
APRIL 2024
Table C-ll: Mean Annualized Cost per CWSs that Treat or Change Water Source,
Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Population Served Size
5th Percentile
Mean
95th
Water
Category
Percentile
Private
Ground
Less than 100
$16,944
$18,186
$20,183
Private
Ground
100 to 500
$25,801
$28,381
$32,205
Private
Ground
500 to 1,000
$41,940
$48,235
$55,777
Private
Ground
1,000 to 3,300
$68,304
$80,715
$95,280
Private
Ground
3,300 to 10,000
$143,480
$192,440
$245,560
Private
Ground
10,000 to 50,000
$359,950
$438,390
$526,620
Private
Ground
50,000 to 100,000
$615,410
$948,460
$1,392,700
Private
Ground
100,000 to 1,000,000
$592,580
$1,050,700
$1,666,100
Private
Surface
Less than 100
$15,352
$20,226
$25,790
Private
Surface
100 to 500
$25,187
$30,707
$36,861
Private
Surface
500 to 1,000
$35,312
$50,902
$69,988
Private
Surface
1,000 to 3,300
$61,201
$88,843
$122,710
Private
Surface
3,300 to 10,000
$125,770
$199,450
$288,180
Private
Surface
10,000 to 50,000
$425,910
$517,180
$619,540
Private
Surface
50,000 to 100,000
$903,940
$1,187,900
$1,530,900
Private
Surface
100,000 to 1,000,000
$2,983,000
$3,596,100
$4,295,000
Public
Ground
Less than 100
$16,725
$19,436
$22,825
Public
Ground
100 to 500
$30,519
$33,905
$38,294
Public
Ground
500 to 1,000
$49,063
$55,040
$61,899
Public
Ground
1,000 to 3,300
$90,153
$99,584
$110,890
Public
Ground
3,300 to 10,000
$194,730
$219,840
$245,950
Public
Ground
10,000 to 50,000
$500,080
$541,560
$585,800
Public
Ground
50,000 to 100,000
$984,600
$1,176,400
$1,382,200
Public
Ground
100,000 to 1,000,000
$2,976,200
$3,699,500
$4,589,800
Public
Surface
Less than 100
$16,658
$21,625
$27,893
Public
Surface
100 to 500
$30,778
$36,528
$42,197
Public
Surface
500 to 1,000
$50,009
$59,678
$70,593
Public
Surface
1,000 to 3,300
$97,524
$112,380
$127,200
Public
Surface
3,300 to 10,000
$237,730
$268,390
$300,680
Public
Surface
10,000 to 50,000
$535,660
$569,340
$605,460
Public
Surface
50,000 to 100,000
$964,720
$1,070,500
$1,176,500
Public
Surface
100,000 to 1,000,000
$2,209,200
$2,490,600
$2,798,400
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-12
April 2024
-------
FINAL RULE
APRIL 2024
Table C-12: Mean Annualized Cost per CWSs that Treat or Change Water Source,
Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Population Served Size
5th Percentile
Mean
95th
Water
Category
Percentile
Private
Ground
Less than 100
$16,187
$18,039
$20,295
Private
Ground
100 to 500
$24,624
$27,953
$32,203
Private
Ground
500 to 1,000
$36,548
$46,598
$57,978
Private
Ground
1,000 to 3,300
$56,592
$76,752
$99,883
Private
Ground
3,300 to 10,000
$98,819
$176,800
$276,220
Private
Ground
10,000 to 50,000
$239,310
$324,260
$419,790
Private
Ground
50,000 to 100,000
$174,140
$435,320
$800,260
Private
Ground
100,000 to 1,000,000
$0
$507,980
$1,175,000
Private
Surface
Less than 100
$14,097
$19,438
$31,100
Private
Surface
100 to 500
$21,045
$30,103
$41,645
Private
Surface
500 to 1,000
$0
$46,865
$82,138
Private
Surface
1,000 to 3,300
$38,218
$83,454
$147,830
Private
Surface
3,300 to 10,000
$33,649
$175,380
$342,640
Private
Surface
10,000 to 50,000
$313,190
$418,160
$549,020
Private
Surface
50,000 to 100,000
$739,270
$1,151,400
$1,661,200
Private
Surface
100,000 to 1,000,000
$1,983,800
$2,659,800
$3,430,300
Public
Ground
Less than 100
$15,074
$19,344
$25,334
Public
Ground
100 to 500
$28,359
$33,254
$38,562
Public
Ground
500 to 1,000
$44,109
$53,122
$62,386
Public
Ground
1,000 to 3,300
$81,392
$94,806
$110,780
Public
Ground
3,300 to 10,000
$162,180
$200,800
$243,170
Public
Ground
10,000 to 50,000
$403,050
$446,600
$493,580
Public
Ground
50,000 to 100,000
$800,390
$1,030,600
$1,305,800
Public
Ground
100,000 to 1,000,000
$2,533,900
$3,560,800
$4,827,400
Public
Surface
Less than 100
$13,493
$20,325
$33,607
Public
Surface
100 to 500
$27,077
$35,995
$46,563
Public
Surface
500 to 1,000
$41,856
$58,095
$76,850
Public
Surface
1,000 to 3,300
$82,653
$108,100
$135,510
Public
Surface
3,300 to 10,000
$211,580
$257,220
$310,940
Public
Surface
10,000 to 50,000
$479,020
$522,730
$567,080
Public
Surface
50,000 to 100,000
$710,630
$831,140
$960,360
Public
Surface
100,000 to 1,000,000
$1,681,100
$1,948,700
$2,250,100
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-13
April 2024
-------
FINAL RULE
APRIL 2024
C.1.4 Mean Annual Cost for Non-Transient Non-Community
Water Systems that Treat or Change Water Source
Table C-13: Mean Annualized Cost per NTNCWSs that Treat or Change Water
Source, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA
MCLs of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$18,125
$19,394
$21,513
Private
Ground
100 to 500
$26,112
$28,871
$32,747
Private
Ground
500 to 1,000
$41,173
$48,876
$56,797
Private
Ground
1,000 to 3,300
$61,540
$79,960
$101,090
Private
Ground
3,300 to 10,000
$94,778
$198,840
$334,310
Private
Ground
10,000 to 50,000
$0
$153,320
$747,450
Private
Surface
Less than 100
$14,470
$19,966
$26,745
Private
Surface
100 to 500
$22,702
$31,823
$43,690
Private
Surface
500 to 1,000
$27,919
$52,731
$89,480
Private
Surface
1,000 to 3,300
$46,653
$90,370
$151,200
Private
Surface
3,300 to 10,000
$61,259
$201,560
$362,440
Private
Surface
10,000 to 50,000
$80,420
$344,440
$635,510
Private
Surface
100,000 to 1,000,000
$0
$405,770
$2,626,200
Public
Ground
Less than 100
$17,587
$20,244
$23,461
Public
Ground
100 to 500
$29,859
$33,627
$38,060
Public
Ground
500 to 1,000
$46,021
$54,139
$63,827
Public
Ground
1,000 to 3,300
$73,407
$93,702
$117,030
Public
Ground
3,300 to 10,000
$0
$226,260
$450,430
Public
Ground
10,000 to 50,000
$354,000
$583,990
$886,520
Public
Surface
Less than 100
$14,436
$21,607
$35,667
Public
Surface
100 to 500
$22,321
$35,767
$53,193
Public
Surface
500 to 1,000
$0
$48,161
$103,640
Public
Surface
1,000 to 3,300
$42,909
$109,430
$198,840
Public
Surface
3,300 to 10,000
$0
$218,700
$432,330
Public
Surface
10,000 to 50,000
$0
$496,320
$1,058,900
Public
Surface
50,000 to 100,000
$0
$164,480
$813,410
Abbreviations: NTNCWS - non-transient, non-community water systems.
Final PFAS Rule Economic Analysis
C-14
April 2024
-------
FINAL RULE
APRIL 2024
Table C-14: Mean Annualized Cost per NTNCWSs that Treat or Change Water
Source, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital,
$2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$18,107
$19,382
$21,454
Private
Ground
100 to 500
$26,166
$28,855
$32,634
Private
Ground
500 to 1,000
$41,173
$48,850
$57,065
Private
Ground
1,000 to 3,300
$62,677
$79,915
$99,887
Private
Ground
3,300 to 10,000
$94,778
$198,790
$334,310
Private
Ground
10,000 to 50,000
$0
$153,370
$747,450
Private
Surface
Less than 100
$14,470
$19,961
$27,243
Private
Surface
100 to 500
$22,702
$31,808
$43,690
Private
Surface
500 to 1,000
$28,203
$52,680
$89,480
Private
Surface
1,000 to 3,300
$46,415
$90,272
$156,990
Private
Surface
3,300 to 10,000
$61,841
$201,080
$371,800
Private
Surface
10,000 to 50,000
$80,395
$341,770
$674,000
Private
Surface
100,000 to 1,000,000
$0
$405,290
$2,626,200
Public
Ground
Less than 100
$17,571
$20,235
$23,461
Public
Ground
100 to 500
$29,855
$33,613
$38,060
Public
Ground
500 to 1,000
$46,021
$54,116
$63,827
Public
Ground
1,000 to 3,300
$73,407
$93,646
$118,760
Public
Ground
3,300 to 10,000
$0
$226,060
$450,430
Public
Ground
10,000 to 50,000
$344,310
$581,700
$892,940
Public
Surface
Less than 100
$14,568
$21,588
$34,339
Public
Surface
100 to 500
$22,808
$35,742
$53,625
Public
Surface
500 to 1,000
$0
$48,080
$110,060
Public
Surface
1,000 to 3,300
$46,115
$109,240
$198,840
Public
Surface
3,300 to 10,000
$0
$218,000
$432,330
Public
Surface
10,000 to 50,000
$0
$494,380
$1,058,900
Public
Surface
50,000 to 100,000
$0
$164,030
$813,410
Abbreviations: NTNCWS - non-transient, non-community water systems.
Final PFAS Rule Economic Analysis
C-15
April 2024
-------
FINAL RULE
APRIL 2024
Table C-15: Mean Annualized Cost per NTNCWSs that Treat or Change Water
Source, Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital,
$2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$17,868
$19,332
$21,416
Private
Ground
100 to 500
$25,641
$28,705
$32,721
Private
Ground
500 to 1,000
$39,619
$48,074
$58,246
Private
Ground
1,000 to 3,300
$59,279
$79,088
$103,620
Private
Ground
3,300 to 10,000
$0
$187,240
$353,360
Private
Ground
10,000 to 50,000
$0
$121,770
$747,450
Private
Surface
Less than 100
$14,121
$19,889
$28,660
Private
Surface
100 to 500
$21,471
$31,606
$45,109
Private
Surface
500 to 1,000
$0
$50,236
$96,546
Private
Surface
1,000 to 3,300
$39,623
$87,040
$160,550
Private
Surface
3,300 to 10,000
$0
$185,300
$395,900
Private
Surface
10,000 to 50,000
$49,308
$312,590
$645,540
Private
Surface
100,000 to 1,000,000
$0
$281,060
$2,611,200
Public
Ground
Less than 100
$17,264
$20,174
$23,615
Public
Ground
100 to 500
$29,340
$33,471
$38,413
Public
Ground
500 to 1,000
$44,370
$53,869
$63,845
Public
Ground
1,000 to 3,300
$69,070
$92,799
$121,850
Public
Ground
3,300 to 10,000
$0
$204,330
$450,430
Public
Ground
10,000 to 50,000
$320,530
$564,510
$934,820
Public
Surface
Less than 100
$0
$20,246
$38,340
Public
Surface
100 to 500
$20,146
$34,922
$55,444
Public
Surface
500 to 1,000
$0
$41,467
$112,980
Public
Surface
1,000 to 3,300
$0
$101,530
$204,050
Public
Surface
3,300 to 10,000
$0
$192,000
$449,400
Public
Surface
10,000 to 50,000
$0
$429,680
$1,045,500
Public
Surface
50,000 to 100,000
$0
$118,780
$792,860
Abbreviations: NTNCWS - non-transient, non-community water systems.
Final PFAS Rule Economic Analysis
C-16
April 2024
-------
FINAL RULE
APRIL 2024
Table C-16: Mean Annualized Cost per NTNCWSs that Treat or Change Water
Source, Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital,
$2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$16,701
$19,048
$21,947
Private
Ground
100 to 500
$23,921
$28,137
$33,235
Private
Ground
500 to 1,000
$32,668
$46,118
$63,177
Private
Ground
1,000 to 3,300
$42,426
$74,725
$126,040
Private
Ground
3,300 to 10,000
$0
$111,920
$300,960
Private
Ground
10,000 to 50,000
$0
$48,529
$540,710
Private
Surface
Less than 100
$0
$16,884
$34,969
Private
Surface
100 to 500
$0
$26,855
$57,097
Private
Surface
500 to 1,000
$0
$30,398
$95,051
Private
Surface
1,000 to 3,300
$0
$53,774
$174,080
Private
Surface
3,300 to 10,000
$0
$109,750
$362,260
Private
Surface
10,000 to 50,000
$0
$197,870
$641,740
Private
Surface
100,000 to 1,000,000
$0
$69,622
$0
Public
Ground
Less than 100
$15,092
$19,867
$26,321
Public
Ground
100 to 500
$26,580
$32,932
$39,958
Public
Ground
500 to 1,000
$38,268
$52,281
$70,819
Public
Ground
1,000 to 3,300
$52,381
$89,382
$144,700
Public
Ground
3,300 to 10,000
$0
$105,680
$392,840
Public
Ground
10,000 to 50,000
$0
$475,260
$991,080
Public
Surface
Less than 100
$0
$10,828
$33,753
Public
Surface
100 to 500
$0
$23,143
$62,659
Public
Surface
500 to 1,000
$0
$17,870
$87,218
Public
Surface
1,000 to 3,300
$0
$55,142
$192,870
Public
Surface
3,300 to 10,000
$0
$88,754
$349,690
Public
Surface
10,000 to 50,000
$0
$210,300
$864,930
Public
Surface
50,000 to 100,000
$0
$28,417
$0
Abbreviations: NTNCWS - non-transient, non-community water systems.
Final PFAS Rule Economic Analysis
C-17
April 2024
-------
FINAL RULE
APRIL 2024
C.1.5 Distribution of Small Community Water System Costs
Table C-17: Distribution of Annualized Cost for Small CWSs, Final Rule (PFOA and PFOS
MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$221
$285
$858
Private
Ground
100 to 500
$190
$190
$316
$441
$1,095
Private
Ground
500 to 1,000
$183
$184
$311
$606
$1,549
Private
Ground
1,000 to 3,300
$190
$207
$357
$776
$3,141
Private
Ground
3,300 to 10,000
$185
$281
$458
$964
$12,678
Private
Surface
Less than 100
$250
$250
$323
$429
$1,157
Private
Surface
100 to 500
$283
$283
$417
$534
$1,297
Private
Surface
500 to 1,000
$269
$269
$444
$689
$2,228
Private
Surface
1,000 to 3,300
$259
$259
$433
$751
$2,436
Private
Surface
3,300 to 10,000
$196
$228
$360
$839
$14,797
Public
Ground
Less than 100
$162
$162
$219
$285
$867
Public
Ground
100 to 500
$190
$190
$317
$427
$1,106
Public
Ground
500 to 1,000
$183
$184
$311
$545
$1,387
Public
Ground
1,000 to 3,300
$190
$207
$355
$733
$2,726
Public
Ground
3,300 to 10,000
$185
$289
$494
$980
$16,246
Public
Surface
Less than 100
$250
$250
$351
$449
$1,232
Public
Surface
100 to 500
$283
$283
$414
$519
$1,271
Public
Surface
500 to 1,000
$269
$269
$447
$685
$1,559
Public
Surface
1,000 to 3,300
$259
$259
$433
$727
$1,783
Public
Surface
3,300 to 10,000
$196
$198
$323
$666
$5,027
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-18
April 2024
-------
FINAL RULE
APRIL 2024
Table C-18: Distribution of Annualized Cost for Small CWSs, Option la (PFOA and PFOS
MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$219
$285
$859
Private
Ground
100 to 500
$190
$190
$316
$438
$1,092
Private
Ground
500 to 1,000
$183
$184
$311
$603
$1,541
Private
Ground
1,000 to 3,300
$190
$206
$357
$771
$3,053
Private
Ground
3,300 to 10,000
$185
$281
$456
$959
$12,639
Private
Surface
Less than 100
$250
$250
$322
$429
$1,157
Private
Surface
100 to 500
$283
$283
$415
$534
$1,298
Private
Surface
500 to 1,000
$269
$269
$443
$685
$2,226
Private
Surface
1,000 to 3,300
$259
$259
$433
$747
$2,405
Private
Surface
3,300 to 10,000
$196
$227
$359
$833
$14,755
Public
Ground
Less than 100
$162
$162
$218
$284
$866
Public
Ground
100 to 500
$190
$190
$317
$425
$1,105
Public
Ground
500 to 1,000
$183
$184
$311
$542
$1,396
Public
Ground
1,000 to 3,300
$190
$206
$354
$728
$2,629
Public
Ground
3,300 to 10,000
$185
$289
$491
$974
$16,056
Public
Surface
Less than 100
$250
$250
$350
$449
$1,231
Public
Surface
100 to 500
$283
$283
$413
$518
$1,270
Public
Surface
500 to 1,000
$269
$269
$447
$682
$1,555
Public
Surface
1,000 to 3,300
$259
$259
$433
$723
$1,803
Public
Surface
3,300 to 10,000
$196
$198
$323
$662
$4,974
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-19
April 2024
-------
FINAL RULE
APRIL 2024
Table C-19: Distribution of Annualized Cost for Small CWSs, Option lb (PFOA and PFOS
MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$218
$281
$759
Private
Ground
100 to 500
$190
$190
$316
$379
$939
Private
Ground
500 to 1,000
$183
$184
$310
$554
$1,067
Private
Ground
1,000 to 3,300
$190
$206
$354
$678
$1,360
Private
Ground
3,300 to 10,000
$185
$281
$444
$880
$2,410
Private
Surface
Less than 100
$250
$250
$320
$420
$950
Private
Surface
100 to 500
$283
$283
$413
$484
$1,181
Private
Surface
500 to 1,000
$269
$269
$443
$595
$1,294
Private
Surface
1,000 to 3,300
$259
$259
$433
$676
$1,295
Private
Surface
3,300 to 10,000
$196
$227
$340
$724
$3,021
Public
Ground
Less than 100
$162
$162
$218
$281
$744
Public
Ground
100 to 500
$190
$190
$316
$371
$945
Public
Ground
500 to 1,000
$183
$184
$310
$502
$1,008
Public
Ground
1,000 to 3,300
$190
$206
$353
$649
$1,296
Public
Ground
3,300 to 10,000
$185
$288
$475
$895
$2,274
Public
Surface
Less than 100
$250
$250
$349
$425
$997
Public
Surface
100 to 500
$283
$283
$410
$480
$1,170
Public
Surface
500 to 1,000
$269
$269
$447
$590
$1,212
Public
Surface
1,000 to 3,300
$259
$259
$433
$656
$1,242
Public
Surface
3,300 to 10,000
$196
$198
$321
$564
$1,319
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-20
April 2024
-------
FINAL RULE
APRIL 2024
Table C-20: Distribution of Annualized Cost for Small CWSs, Option lc (PFOA and PFOS
MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$216
$281
$359
Private
Ground
100 to 500
$190
$190
$315
$352
$608
Private
Ground
500 to 1,000
$183
$184
$310
$487
$806
Private
Ground
1,000 to 3,300
$190
$206
$353
$609
$961
Private
Ground
3,300 to 10,000
$185
$281
$421
$744
$1,228
Private
Surface
Less than 100
$250
$250
$318
$420
$515
Private
Surface
100 to 500
$283
$283
$410
$472
$839
Private
Surface
500 to 1,000
$269
$269
$442
$498
$936
Private
Surface
1,000 to 3,300
$259
$259
$433
$540
$971
Private
Surface
3,300 to 10,000
$196
$227
$324
$582
$1,085
Public
Ground
Less than 100
$162
$162
$217
$281
$351
Public
Ground
100 to 500
$190
$190
$316
$352
$610
Public
Ground
500 to 1,000
$183
$184
$310
$456
$637
Public
Ground
1,000 to 3,300
$190
$206
$353
$604
$919
Public
Ground
3,300 to 10,000
$185
$288
$452
$755
$1,212
Public
Surface
Less than 100
$250
$250
$346
$420
$599
Public
Surface
100 to 500
$283
$283
$407
$472
$795
Public
Surface
500 to 1,000
$269
$269
$447
$497
$884
Public
Surface
1,000 to 3,300
$259
$259
$433
$503
$901
Public
Surface
3,300 to 10,000
$196
$198
$296
$518
$909
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-21
April 2024
-------
FINAL RULE
APRIL 2024
C.1.6 Distribution of Small Non-Community Non-Transient
Water System Costs
Table C-21: Distribution of Annualized Cost for Small NTNCWSs, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
(Commercial Cost of Capital, $2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$261
$341
$912
Private
Ground
100 to 500
$181
$190
$283
$412
$1,057
Private
Ground
500 to 1,000
$156
$183
$294
$480
$1,434
Private
Ground
1,000 to 3,300
$189
$190
$320
$657
$2,661
Private
Ground
3,300 to 10,000
$185
$186
$285
$656
$13,612
Private
Surface
Less than 100
$250
$250
$394
$498
$1,441
Private
Surface
100 to 500
$283
$283
$464
$688
$1,963
Private
Surface
500 to 1,000
$269
$278
$467
$893
$5,618
Private
Surface
1,000 to 3,300
$259
$259
$421
$958
$4,991
Private
Surface
3,300 to 10,000
$213
$321
$658
$1,852
$31,905
Public
Ground
Less than 100
$162
$162
$260
$337
$928
Public
Ground
100 to 500
$190
$190
$287
$373
$1,021
Public
Ground
500 to 1,000
$183
$183
$289
$389
$1,092
Public
Ground
1,000 to 3,300
$190
$190
$319
$560
$1,969
Public
Ground
3,300 to 10,000
$186
$248
$345
$684
$22,620
Public
Surface
Less than 100
$250
$250
$359
$472
$1,794
Public
Surface
100 to 500
$283
$283
$449
$698
$2,703
Public
Surface
500 to 1,000
$269
$270
$382
$577
$4,910
Public
Surface
1,000 to 3,300
$259
$279
$487
$1,066
$9,462
Public
Surface
3,300 to 10,000
$196
$214
$372
$1,197
$26,572
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-22
April 2024
-------
FINAL RULE
APRIL 2024
Table C-22: Distribution of Annualized Cost for Small NTNCWSs, Option la (PFOA and
PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$260
$339
$907
Private
Ground
100 to 500
$181
$190
$283
$410
$1,051
Private
Ground
500 to 1,000
$156
$183
$294
$478
$1,431
Private
Ground
1,000 to 3,300
$188
$190
$319
$654
$2,653
Private
Ground
3,300 to 10,000
$185
$186
$284
$652
$13,540
Private
Surface
Less than 100
$250
$250
$393
$498
$1,440
Private
Surface
100 to 500
$283
$283
$464
$684
$1,953
Private
Surface
500 to 1,000
$269
$278
$467
$890
$5,579
Private
Surface
1,000 to 3,300
$259
$259
$421
$955
$4,977
Private
Surface
3,300 to 10,000
$213
$320
$654
$1,836
$31,636
Public
Ground
Less than 100
$162
$162
$259
$336
$933
Public
Ground
100 to 500
$190
$190
$286
$372
$1,019
Public
Ground
500 to 1,000
$183
$183
$288
$387
$1,085
Public
Ground
1,000 to 3,300
$190
$190
$319
$557
$1,971
Public
Ground
3,300 to 10,000
$186
$247
$344
$680
$22,497
Public
Surface
Less than 100
$250
$250
$358
$471
$1,792
Public
Surface
100 to 500
$283
$283
$449
$694
$2,689
Public
Surface
500 to 1,000
$269
$270
$381
$576
$4,909
Public
Surface
1,000 to 3,300
$259
$279
$487
$1,061
$9,428
Public
Surface
3,300 to 10,000
$196
$214
$371
$1,191
$26,455
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-23
April 2024
-------
FINAL RULE
APRIL 2024
Table C-23: Distribution of Annualized Cost for Small NTNCWSs, Option lb (PFOA and
PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$259
$302
$838
Private
Ground
100 to 500
$181
$190
$279
$359
$957
Private
Ground
500 to 1,000
$156
$183
$290
$416
$980
Private
Ground
1,000 to 3,300
$189
$190
$319
$574
$1,169
Private
Ground
3,300 to 10,000
$185
$186
$284
$548
$4,888
Private
Surface
Less than 100
$250
$250
$393
$466
$999
Private
Surface
100 to 500
$283
$283
$464
$607
$1,248
Private
Surface
500 to 1,000
$269
$278
$457
$804
$2,755
Private
Surface
1,000 to 3,300
$259
$259
$421
$838
$2,270
Private
Surface
3,300 to 10,000
$213
$315
$615
$1,509
$14,494
Public
Ground
Less than 100
$162
$162
$259
$301
$825
Public
Ground
100 to 500
$190
$190
$285
$338
$925
Public
Ground
500 to 1,000
$183
$183
$286
$344
$924
Public
Ground
1,000 to 3,300
$190
$190
$319
$495
$1,083
Public
Ground
3,300 to 10,000
$186
$247
$335
$594
$8,372
Public
Surface
Less than 100
$250
$250
$357
$437
$1,049
Public
Surface
100 to 500
$283
$283
$449
$612
$1,399
Public
Surface
500 to 1,000
$269
$270
$380
$515
$2,462
Public
Surface
1,000 to 3,300
$259
$278
$473
$953
$4,249
Public
Surface
3,300 to 10,000
$196
$214
$345
$960
$11,166
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-24
April 2024
-------
FINAL RULE
APRIL 2024
Table C-24: Distribution of Annualized Cost for Small NTNCWSs, Option lc (PFOA and
PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$162
$162
$256
$281
$541
Private
Ground
100 to 500
$181
$190
$271
$320
$614
Private
Ground
500 to 1,000
$156
$183
$281
$344
$647
Private
Ground
1,000 to 3,300
$189
$190
$319
$486
$895
Private
Ground
3,300 to 10,000
$185
$186
$284
$428
$887
Private
Surface
Less than 100
$250
$250
$392
$423
$694
Private
Surface
100 to 500
$283
$283
$463
$525
$903
Private
Surface
500 to 1,000
$269
$277
$450
$717
$1,139
Private
Surface
1,000 to 3,300
$259
$259
$420
$625
$1,389
Private
Surface
3,300 to 10,000
$213
$304
$554
$1,234
$2,215
Public
Ground
Less than 100
$162
$162
$257
$281
$534
Public
Ground
100 to 500
$190
$190
$281
$320
$575
Public
Ground
500 to 1,000
$183
$183
$283
$311
$586
Public
Ground
1,000 to 3,300
$190
$190
$319
$388
$750
Public
Ground
3,300 to 10,000
$186
$247
$322
$516
$842
Public
Surface
Less than 100
$250
$250
$356
$420
$612
Public
Surface
100 to 500
$283
$283
$448
$511
$920
Public
Surface
500 to 1,000
$269
$270
$378
$455
$830
Public
Surface
1,000 to 3,300
$259
$277
$449
$795
$1,461
Public
Surface
3,300 to 10,000
$196
$214
$312
$732
$1,667
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-25
April 2024
-------
FINAL RULE
APRIL 2024
C.l. 7 Distribution of Small Community Water System Costs that
Treat or Change Water Source
Table C-25: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs
of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,750
$13,445
$14,523
$16,725
$30,195
Private
Ground
100 to 500
$17,003
$18,703
$22,738
$31,722
$49,790
Private
Ground
500 to 1,000
$24,843
$31,251
$41,268
$54,307
$83,580
Private
Ground
1,000 to 3,300
$35,629
$47,594
$65,988
$95,435
$144,530
Private
Ground
3,300 to 10,000
$58,719
$90,639
$165,080
$241,650
$341,780
Private
Surface
Less than 100
$13,451
$14,569
$15,794
$19,020
$32,602
Private
Surface
100 to 500
$17,809
$19,671
$24,002
$33,208
$53,699
Private
Surface
500 to 1,000
$25,064
$30,530
$39,135
$55,808
$84,746
Private
Surface
1,000 to 3,300
$37,748
$49,678
$67,880
$97,672
$150,360
Private
Surface
3,300 to 10,000
$57,404
$89,217
$151,960
$238,820
$351,490
Public
Ground
Less than 100
$13,272
$14,190
$15,621
$18,367
$32,161
Public
Ground
100 to 500
$19,193
$22,308
$28,496
$38,173
$56,711
Public
Ground
500 to 1,000
$28,965
$36,305
$47,597
$61,034
$93,788
Public
Ground
1,000 to 3,300
$42,284
$56,853
$78,948
$121,920
$189,720
Public
Ground
3,300 to 10,000
$67,894
$100,630
$194,660
$287,850
$420,440
Public
Surface
Less than 100
$13,983
$15,193
$16,803
$21,445
$34,256
Public
Surface
100 to 500
$20,177
$23,419
$29,974
$40,734
$64,081
Public
Surface
500 to 1,000
$30,679
$38,278
$48,596
$66,562
$106,290
Public
Surface
1,000 to 3,300
$47,855
$62,781
$85,383
$140,850
$210,670
Public
Surface
3,300 to 10,000
$83,040
$153,700
$227,760
$330,790
$488,940
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-26
April 2024
-------
FINAL RULE
APRIL 2024
Table C-26: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,750
$13,444
$14,523
$16,721
$30,175
Private
Ground
100 to 500
$17,000
$18,699
$22,730
$31,709
$49,758
Private
Ground
500 to 1,000
$24,829
$31,232
$41,240
$54,258
$83,465
Private
Ground
1,000 to 3,300
$35,604
$47,550
$65,923
$95,320
$144,350
Private
Ground
3,300 to 10,000
$58,663
$90,506
$164,840
$241,320
$341,230
Private
Surface
Less than 100
$13,451
$14,569
$15,792
$19,010
$32,603
Private
Surface
100 to 500
$17,805
$19,666
$23,995
$33,190
$53,660
Private
Surface
500 to 1,000
$25,058
$30,516
$39,103
$55,752
$84,657
Private
Surface
1,000 to 3,300
$37,739
$49,639
$67,831
$97,568
$150,260
Private
Surface
3,300 to 10,000
$57,294
$89,035
$151,700
$238,440
$350,960
Public
Ground
Less than 100
$13,272
$14,189
$15,620
$18,361
$32,132
Public
Ground
100 to 500
$19,191
$22,302
$28,489
$38,159
$56,689
Public
Ground
500 to 1,000
$28,950
$36,289
$47,574
$60,996
$93,686
Public
Ground
1,000 to 3,300
$42,251
$56,809
$78,883
$121,730
$189,640
Public
Ground
3,300 to 10,000
$67,835
$100,470
$194,460
$287,420
$419,990
Public
Surface
Less than 100
$13,983
$15,192
$16,800
$21,423
$34,249
Public
Surface
100 to 500
$20,173
$23,414
$29,965
$40,710
$64,052
Public
Surface
500 to 1,000
$30,669
$38,270
$48,580
$66,520
$106,180
Public
Surface
1,000 to 3,300
$47,829
$62,751
$85,354
$140,740
$210,600
Public
Surface
3,300 to 10,000
$83,018
$153,580
$227,660
$330,500
$488,610
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-27
April 2024
-------
FINAL RULE
APRIL 2024
Table C-27: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,744
$13,435
$14,509
$16,689
$30,245
Private
Ground
100 to 500
$16,962
$18,644
$22,623
$31,536
$49,241
Private
Ground
500 to 1,000
$24,492
$30,654
$40,663
$53,661
$81,797
Private
Ground
1,000 to 3,300
$34,843
$46,451
$64,668
$93,525
$140,540
Private
Ground
3,300 to 10,000
$56,508
$86,484
$155,640
$234,470
$327,280
Private
Surface
Less than 100
$13,322
$14,486
$15,717
$19,194
$30,953
Private
Surface
100 to 500
$17,725
$19,601
$23,791
$32,883
$51,940
Private
Surface
500 to 1,000
$25,152
$29,874
$38,097
$54,332
$79,085
Private
Surface
1,000 to 3,300
$37,209
$48,299
$65,996
$95,109
$142,570
Private
Surface
3,300 to 10,000
$58,036
$83,906
$141,760
$227,440
$329,830
Public
Ground
Less than 100
$13,248
$14,160
$15,579
$18,349
$31,377
Public
Ground
100 to 500
$19,130
$22,195
$28,291
$37,922
$56,127
Public
Ground
500 to 1,000
$28,629
$35,770
$46,975
$60,482
$92,174
Public
Ground
1,000 to 3,300
$41,636
$55,600
$77,506
$119,090
$187,280
Public
Ground
3,300 to 10,000
$66,000
$94,623
$187,120
$277,530
$408,740
Public
Surface
Less than 100
$13,758
$15,033
$16,692
$21,288
$33,078
Public
Surface
100 to 500
$20,024
$23,218
$29,664
$40,359
$62,450
Public
Surface
500 to 1,000
$30,155
$37,760
$48,043
$65,797
$103,350
Public
Surface
1,000 to 3,300
$46,963
$61,882
$84,056
$138,370
$208,570
Public
Surface
3,300 to 10,000
$79,968
$147,850
$223,690
$326,190
$483,870
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-28
April 2024
-------
FINAL RULE
APRIL 2024
Table C-28: Distribution of Annualized Cost for Small CWSs that Treat or Change Water
Source, Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022)
Annualized Cost Per CWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,726
$13,423
$14,474
$16,567
$29,647
Private
Ground
100 to 500
$16,852
$18,475
$22,290
$30,930
$47,110
Private
Ground
500 to 1,000
$23,340
$28,728
$38,086
$51,291
$73,348
Private
Ground
1,000 to 3,300
$32,489
$42,724
$59,347
$85,865
$124,860
Private
Ground
3,300 to 10,000
$63,179
$74,975
$120,970
$195,620
$266,600
Private
Surface
Less than 100
$14,263
$14,428
$15,247
$17,908
$24,626
Private
Surface
100 to 500
$17,999
$19,247
$22,685
$30,699
$42,930
Private
Surface
500 to 1,000
$30,887
$31,200
$34,564
$43,639
$62,927
Private
Surface
1,000 to 3,300
$48,078
$49,267
$57,787
$78,196
$117,010
Private
Surface
3,300 to 10,000
$91,326
$93,156
$112,970
$162,800
$255,790
Public
Ground
Less than 100
$13,157
$14,009
$15,354
$18,431
$28,042
Public
Ground
100 to 500
$18,884
$21,800
$27,477
$36,972
$53,864
Public
Ground
500 to 1,000
$27,398
$33,878
$44,461
$58,528
$85,588
Public
Ground
1,000 to 3,300
$39,739
$52,128
$73,006
$111,030
$176,390
Public
Ground
3,300 to 10,000
$60,139
$82,868
$160,830
$253,570
$375,840
Public
Surface
Less than 100
$14,956
$15,059
$15,876
$18,713
$26,060
Public
Surface
100 to 500
$19,711
$22,613
$28,278
$38,733
$55,114
Public
Surface
500 to 1,000
$28,949
$35,217
$45,324
$62,094
$87,753
Public
Surface
1,000 to 3,300
$42,915
$57,728
$79,786
$126,990
$193,340
Public
Surface
3,300 to 10,000
$75,406
$128,950
$208,950
$307,780
$464,170
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-29
April 2024
-------
FINAL RULE
APRIL 2024
C.1.8 Distribution of Small Non-Community Water Non-
Transient System Costs that Treat or Change Water
Source
Table C-29: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA
MCLs of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,451
$13,181
$14,416
$17,845
$38,382
Private
Ground
100 to 500
$17,039
$18,775
$22,782
$31,590
$50,907
Private
Ground
500 to 1,000
$26,001
$32,521
$41,022
$52,515
$79,603
Private
Ground
1,000 to 3,300
$34,937
$47,451
$63,747
$87,675
$129,160
Private
Ground
3,300 to 10,000
$108,430
$113,140
$147,680
$194,980
$271,070
Private
Surface
Less than 100
$12,596
$13,501
$14,949
$19,443
$29,428
Private
Surface
100 to 500
$17,916
$19,541
$23,718
$33,776
$47,778
Private
Surface
500 to 1,000
$30,616
$31,305
$37,024
$50,835
$73,636
Private
Surface
1,000 to 3,300
$44,409
$47,126
$60,542
$87,911
$130,210
Private
Surface
3,300 to 10,000
$89,611
$91,902
$124,350
$196,240
$306,580
Public
Ground
Less than 100
$12,920
$13,794
$15,400
$19,662
$36,481
Public
Ground
100 to 500
$18,966
$22,126
$28,266
$37,222
$54,701
Public
Ground
500 to 1,000
$29,992
$37,182
$45,574
$56,605
$86,674
Public
Ground
1,000 to 3,300
$42,391
$56,862
$72,128
$98,453
$159,680
Public
Ground
3,300 to 10,000
$133,780
$135,070
$157,540
$208,870
$317,210
Public
Surface
Less than 100
$15,324
$15,423
$16,416
$20,009
$28,258
Public
Surface
100 to 500
$21,224
$22,143
$26,479
$36,213
$49,068
Public
Surface
500 to 1,000
$38,033
$38,046
$38,993
$43,067
$59,764
Public
Surface
1,000 to 3,300
$63,245
$64,197
$75,362
$103,100
$154,320
Public
Surface
3,300 to 10,000
$142,410
$142,740
$157,880
$197,200
$296,500
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-30
April 2024
-------
FINAL RULE
APRIL 2024
Table C-30: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital,
$2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,450
$13,180
$14,413
$17,834
$38,373
Private
Ground
100 to 500
$17,038
$18,770
$22,774
$31,567
$50,857
Private
Ground
500 to 1,000
$25,991
$32,511
$41,004
$52,461
$79,515
Private
Ground
1,000 to 3,300
$34,920
$47,438
$63,728
$87,595
$129,030
Private
Ground
3,300 to 10,000
$108,520
$113,230
$147,790
$195,000
$270,800
Private
Surface
Less than 100
$12,595
$13,499
$14,950
$19,420
$29,411
Private
Surface
100 to 500
$17,914
$19,540
$23,712
$33,744
$47,744
Private
Surface
500 to 1,000
$30,623
$31,300
$36,993
$50,766
$73,525
Private
Surface
1,000 to 3,300
$44,475
$47,171
$60,500
$87,701
$130,060
Private
Surface
3,300 to 10,000
$89,650
$91,923
$123,960
$195,490
$305,670
Public
Ground
Less than 100
$12,919
$13,792
$15,397
$19,644
$36,471
Public
Ground
100 to 500
$18,962
$22,120
$28,257
$37,209
$54,651
Public
Ground
500 to 1,000
$29,984
$37,183
$45,561
$56,590
$86,580
Public
Ground
1,000 to 3,300
$42,375
$56,843
$72,105
$98,325
$159,530
Public
Ground
3,300 to 10,000
$133,820
$135,110
$157,390
$208,740
$316,590
Public
Surface
Less than 100
$15,310
$15,409
$16,399
$19,981
$28,247
Public
Surface
100 to 500
$21,210
$22,132
$26,475
$36,185
$48,997
Public
Surface
500 to 1,000
$37,995
$38,008
$38,954
$42,993
$59,642
Public
Surface
1,000 to 3,300
$63,193
$64,144
$75,370
$102,930
$154,010
Public
Surface
3,300 to 10,000
$141,960
$142,290
$157,490
$196,700
$295,600
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-31
April 2024
-------
FINAL RULE
APRIL 2024
Table C-31: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital,
$2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,441
$13,173
$14,395
$17,866
$37,796
Private
Ground
100 to 500
$16,994
$18,713
$22,658
$31,418
$50,164
Private
Ground
500 to 1,000
$25,367
$31,510
$40,057
$51,693
$76,955
Private
Ground
1,000 to 3,300
$33,885
$46,020
$62,235
$85,860
$124,840
Private
Ground
3,300 to 10,000
$116,740
$118,180
$137,910
$174,780
$254,240
Private
Surface
Less than 100
$12,934
$13,521
$14,884
$19,077
$26,928
Private
Surface
100 to 500
$18,569
$19,587
$23,329
$32,422
$44,212
Private
Surface
500 to 1,000
$32,811
$33,031
$36,224
$45,969
$68,309
Private
Surface
1,000 to 3,300
$48,775
$49,690
$58,778
$80,200
$125,140
Private
Surface
3,300 to 10,000
$99,823
$100,360
$117,680
$166,250
$273,720
Public
Ground
Less than 100
$12,907
$13,759
$15,342
$19,667
$35,400
Public
Ground
100 to 500
$18,891
$22,001
$28,049
$37,008
$54,283
Public
Ground
500 to 1,000
$29,480
$36,685
$45,184
$56,346
$84,445
Public
Ground
1,000 to 3,300
$41,109
$55,446
$70,910
$97,287
$153,630
Public
Ground
3,300 to 10,000
$136,940
$137,190
$148,180
$182,210
$277,450
Public
Surface
Less than 100
$15,374
$15,401
$15,954
$18,270
$25,901
Public
Surface
100 to 500
$22,637
$22,967
$25,984
$33,572
$46,835
Public
Surface
500 to 1,000
$34,784
$34,784
$35,146
$37,399
$49,010
Public
Surface
1,000 to 3,300
$66,752
$67,005
$72,937
$90,648
$138,740
Public
Surface
3,300 to 10,000
$140,200
$140,280
$146,950
$169,230
$246,610
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-32
April 2024
-------
FINAL RULE
APRIL 2024
Table C-32: Distribution of Annualized Cost for Small NTNCWSs that Treat or Change
Water Source, Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital,
$2022)
Annualized Cost Per NTNCWS
Ownership
Source
Population
10th
25th
50th
75th
90th
Water
Served Size
Percentile
Percentile
Percentile
Percentile
Percentile
Category
Private
Ground
Less than 100
$12,375
$13,131
$14,296
$17,751
$34,871
Private
Ground
100 to 500
$16,815
$18,491
$22,165
$30,419
$46,664
Private
Ground
500 to 1,000
$25,081
$28,923
$36,345
$47,778
$64,512
Private
Ground
1,000 to 3,300
$40,119
$42,603
$54,075
$73,768
$103,380
Private
Ground
3,300 to 10,000
$95,371
$95,371
$96,363
$100,690
$129,210
Private
Surface
Less than 100
$13,632
$13,637
$13,914
$15,226
$20,728
Private
Surface
100 to 500
$20,645
$20,662
$21,303
$24,048
$34,055
Private
Surface
500 to 1,000
$26,102
$26,105
$26,261
$27,217
$35,049
Private
Surface
1,000 to 3,300
$44,383
$44,383
$44,782
$47,239
$64,471
Private
Surface
3,300 to 10,000
$88,851
$88,851
$89,884
$95,680
$132,330
Public
Ground
Less than 100
$12,805
$13,612
$15,098
$19,358
$30,326
Public
Ground
100 to 500
$18,482
$21,440
$27,084
$35,855
$50,906
Public
Ground
500 to 1,000
$29,018
$34,032
$42,141
$53,541
$72,239
Public
Ground
1,000 to 3,300
$47,962
$51,574
$63,811
$86,975
$124,370
Public
Ground
3,300 to 10,000
$93,410
$93,410
$93,659
$96,190
$118,780
Public
Surface
Less than 100
$10,001
$10,001
$10,021
$10,173
$11,801
Public
Surface
100 to 500
$19,793
$19,794
$19,977
$20,935
$26,985
Public
Surface
500 to 1,000
$17,025
$17,025
$17,025
$17,117
$18,765
Public
Surface
1,000 to 3,300
$47,907
$47,907
$48,190
$49,706
$62,882
Public
Surface
3,300 to 10,000
$79,247
$79,247
$79,367
$81,463
$98,701
Abbreviations: NTNCWS - non-transient non-community water system.
Final PFAS Rule Economic Analysis
C-33
April 2024
-------
FINAL RULE
APRIL 2024
C.2 Household-Level Cost Details
Section C.2 provides estimates of household costs by primary source water, ownership, and
system size category. Costs are provided for all CWSs as well as for only CWSs that must treat
or change water source to comply with the regulatory option.
C.2.1 Household Costs for all Community Water Systems
Table C-33: Mean Annualized Cost per Household in CWSs, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI
of 1) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$42
$57
$74
Private
Ground
100 to 500
$24
$33
$43
Private
Ground
500 to 1,000
$9
$13
$18
Private
Ground
1,000 to 3,300
$6
$9
$13
Private
Ground
3,300 to 10,000
$4
$7
$11
Private
Ground
10,000 to 50,000
$14
$17
$21
Private
Ground
50,000 to 100,000
$9
$16
$23
Private
Ground
100,000 to 1,000,000
$5
$9
$14
Private
Surface
Less than 100
$36
$56
$80
Private
Surface
100 to 500
$18
$27
$37
Private
Surface
500 to 1,000
$6
$12
$18
Private
Surface
1,000 to 3,300
$4
$7
$11
Private
Surface
3,300 to 10,000
$3
$7
$11
Private
Surface
10,000 to 50,000
$9
$12
$15
Private
Surface
50,000 to 100,000
$11
$15
$19
Private
Surface
100,000 to 1,000,000
$13
$15
$18
Public
Ground
Less than 100
$49
$71
$95
Public
Ground
100 to 500
$22
$30
$40
Public
Ground
500 to 1,000
$7
$10
$14
Public
Ground
1,000 to 3,300
$5
$8
$10
Public
Ground
3,300 to 10,000
$13
$18
$24
Public
Ground
10,000 to 50,000
$15
$17
$18
Public
Ground
50,000 to 100,000
$11
$14
$17
Public
Ground
100,000 to 1,000,000
$12
$15
$19
Public
Surface
Less than 100
$53
$81
$115
Public
Surface
100 to 500
$19
$28
$37
Public
Surface
500 to 1,000
$7
$10
$13
Public
Surface
1,000 to 3,300
$5
$7
$9
Public
Surface
3,300 to 10,000
$12
$17
$23
Public
Surface
10,000 to 50,000
$13
$14
$16
Public
Surface
50,000 to 100,000
$11
$12
$14
Public
Surface
100,000 to 1,000,000
$11
$12
$14
Final PFAS Rule Economic Analysis
C-34
April 2024
-------
FINAL RULE
APRIL 2024
Table C-33: Mean Annualized Cost per Household in CWSs, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI
of 1) (Commercial Cost of Capital, $2022)
Ownership Source Population Served Size 5th Percentile Mean 95th Percentile
Water Category
Abbreviations: CWS - community water system.
Table C-34: Mean Annualized Cost per Household in CWSs, Option la (PFOA and
PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th
Percentile
Private
Ground
Less than 100
$42
$57
$74
Private
Ground
100 to 500
$24
$33
$43
Private
Ground
500 to 1,000
$9
$13
$18
Private
Ground
1,000 to 3,300
$6
$9
$13
Private
Ground
3,300 to 10,000
$4
$7
$11
Private
Ground
10,000 to 50,000
$14
$17
$21
Private
Ground
50,000 to 100,000
$9
$14
$22
Private
Ground
100,000 to 1,000,000
$5
$9
$14
Private
Surface
Less than 100
$36
$56
$80
Private
Surface
100 to 500
$18
$27
$37
Private
Surface
500 to 1,000
$6
$12
$18
Private
Surface
1,000 to 3,300
$4
$7
$11
Private
Surface
3,300 to 10,000
$3
$7
$11
Private
Surface
10,000 to 50,000
$9
$12
$15
Private
Surface
50,000 to 100,000
$10
$14
$19
Private
Surface
100,000 to 1,000,000
$12
$15
$17
Public
Ground
Less than 100
$49
$71
$96
Public
Ground
100 to 500
$22
$30
$41
Public
Ground
500 to 1,000
$7
$10
$14
Public
Ground
1,000 to 3,300
$5
$8
$10
Public
Ground
3,300 to 10,000
$13
$18
$24
Public
Ground
10,000 to 50,000
$15
$16
$18
Public
Ground
50,000 to 100,000
$11
$14
$17
Public
Ground
100,000 to 1,000,000
$12
$15
$18
Public
Surface
Less than 100
$53
$81
$115
Public
Surface
100 to 500
$19
$28
$38
Public
Surface
500 to 1,000
$7
$10
$13
Public
Surface
1,000 to 3,300
$5
$7
$9
Public
Surface
3,300 to 10,000
$12
$17
$24
Public
Surface
10,000 to 50,000
$13
$14
$16
Public
Surface
50,000 to 100,000
$11
$12
$14
Final PFAS Rule Economic Analysis
C-35
April 2024
-------
FINAL RULE
APRIL 2024
Table C-34: Mean Annualized Cost per Household in CWSs, Option la (PFOA and
PFOS MCLs of 4.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th
Percentile
Public
Surface
100,000 to 1,000,000
$11
$12
$14
Abbreviations:
CWS - community water system.
Table C-35: Mean Annualized Cost per Household in CWSs, Option lb (PFOA and
PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$32
$43
$57
Private
Ground
100 to 500
$18
$25
$33
Private
Ground
500 to 1,000
$6
$10
$14
Private
Ground
1,000 to 3,300
$4
$7
$10
Private
Ground
3,300 to 10,000
$3
$5
$8
Private
Ground
10,000 to 50,000
$10
$13
$16
Private
Ground
50,000 to 100,000
$6
$11
$17
Private
Ground
100,000 to 1,000,000
$3
$6
$11
Private
Surface
Less than 100
$27
$44
$64
Private
Surface
100 to 500
$14
$21
$29
Private
Surface
500 to 1,000
$4
$9
$14
Private
Surface
1,000 to 3,300
$3
$5
$8
Private
Surface
3,300 to 10,000
$2
$5
$8
Private
Surface
10,000 to 50,000
$7
$9
$12
Private
Surface
50,000 to 100,000
$8
$12
$16
Private
Surface
100,000 to 1,000,000
$10
$12
$14
Public
Ground
Less than 100
$38
$54
$73
Public
Ground
100 to 500
$16
$23
$31
Public
Ground
500 to 1,000
$5
$8
$11
Public
Ground
1,000 to 3,300
$4
$6
$8
Public
Ground
3,300 to 10,000
$9
$13
$18
Public
Ground
10,000 to 50,000
$12
$13
$14
Public
Ground
50,000 to 100,000
$8
$11
$13
Public
Ground
100,000 to 1,000,000
$9
$12
$15
Public
Surface
Less than 100
$40
$64
$93
Public
Surface
100 to 500
$15
$21
$29
Public
Surface
500 to 1,000
$5
$7
$10
Public
Surface
1,000 to 3,300
$3
$5
$7
Public
Surface
3,300 to 10,000
$9
$13
$17
Public
Surface
10,000 to 50,000
$10
$11
$12
Final PFAS Rule Economic Analysis
C-36
April 2024
-------
FINAL RULE
APRIL 2024
Table C-35: Mean Annualized Cost per Household in CWSs, Option lb (PFOA and
PFOS MCLs of 5.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Population Served Size 5th Percentile
Mean
95th Percentile
Water
Category
Public
Surface
50,000 to 100,000 $8
$
9 $11
Public
Surface
100,000 to 1,000,000 $8
$
9 $11
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-37
April 2024
-------
FINAL RULE
APRIL 2024
Table C-36: Mean Annualized Cost per Household in CWSs, Option lc (PFOA and
PFOS MCLs of 10.0 ppt) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$16
$20
$25
Private
Ground
100 to 500
$8
$11
$14
Private
Ground
500 to 1,000
$3
$4
$6
Private
Ground
1,000 to 3,300
$1
$2
$4
Private
Ground
3,300 to 10,000
$1
$2
$3
Private
Ground
10,000 to 50,000
$3
$5
$6
Private
Ground
50,000 to 100,000
$1
$3
$6
Private
Ground
100,000 to 1,000,000
$0
$1
$3
Private
Surface
Less than 100
$15
$22
$32
Private
Surface
100 to 500
$7
$10
$14
Private
Surface
500 to 1,000
$2
$4
$6
Private
Surface
1,000 to 3,300
$1
$2
$4
Private
Surface
3,300 to 10,000
$0
$2
$3
Private
Surface
10,000 to 50,000
$2
$3
$5
Private
Surface
50,000 to 100,000
$3
$5
$8
Private
Surface
100,000 to 1,000,000
$4
$5
$7
Public
Ground
Less than 100
$18
$25
$33
Public
Ground
100 to 500
$7
$10
$13
Public
Ground
500 to 1,000
$2
$3
$4
Public
Ground
1,000 to 3,300
$1
$2
$3
Public
Ground
3,300 to 10,000
$3
$4
$6
Public
Ground
10,000 to 50,000
$5
$5
$6
Public
Ground
50,000 to 100,000
$3
$4
$6
Public
Ground
100,000 to 1,000,000
$4
$6
$7
Public
Surface
Less than 100
$19
$30
$45
Public
Surface
100 to 500
$7
$9
$13
Public
Surface
500 to 1,000
$2
$3
$4
Public
Surface
1,000 to 3,300
$1
$2
$2
Public
Surface
3,300 to 10,000
$2
$4
$6
Public
Surface
10,000 to 50,000
$4
$4
$5
Public
Surface
50,000 to 100,000
$2
$3
$4
Public
Surface
100,000 to 1,000,000
$3
$4
$4
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-38
April 2024
-------
FINAL RULE
APRIL 2024
C.2.2 Household Costs for Community Water Systems that Treat
or Change Water Source
Table C-37: Mean Annualized Cost per Household in CWSs that Treat or Change
Water Source, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA,
HFPO-DA MCLs of 10 ppt each and HI of 1) (Commercial Cost of Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$755
$809
$899
Private
Ground
100 to 500
$413
$452
$513
Private
Ground
500 to 1,000
$151
$173
$198
Private
Ground
1,000 to 3,300
$99
$114
$130
Private
Ground
3,300 to 10,000
$67
$85
$104
Private
Ground
10,000 to 50,000
$45
$53
$62
Private
Ground
50,000 to 100,000
$32
$45
$62
Private
Ground
100,000 to 1,000,000
$12
$18
$27
Private
Surface
Less than 100
$567
$764
$1,004
Private
Surface
100 to 500
$310
$370
$444
Private
Surface
500 to 1,000
$120
$161
$206
Private
Surface
1,000 to 3,300
$76
$103
$132
Private
Surface
3,300 to 10,000
$56
$79
$106
Private
Surface
10,000 to 50,000
$40
$48
$57
Private
Surface
50,000 to 100,000
$35
$44
$55
Private
Surface
100,000 to 1,000,000
$26
$30
$34
Public
Ground
Less than 100
$884
$1,031
$1,190
Public
Ground
100 to 500
$388
$429
$486
Public
Ground
500 to 1,000
$129
$144
$161
Public
Ground
1,000 to 3,300
$89
$98
$108
Public
Ground
3,300 to 10,000
$176
$194
$215
Public
Ground
10,000 to 50,000
$51
$54
$59
Public
Ground
50,000 to 100,000
$36
$42
$48
Public
Ground
100,000 to 1,000,000
$33
$40
$47
Public
Surface
Less than 100
$803
$1,057
$1,373
Public
Surface
100 to 500
$344
$398
$461
Public
Surface
500 to 1,000
$121
$140
$162
Public
Surface
1,000 to 3,300
$90
$100
$110
Public
Surface
3,300 to 10,000
$199
$221
$245
Public
Surface
10,000 to 50,000
$54
$57
$61
Public
Surface
50,000 to 100,000
$41
$45
$49
Public
Surface
100,000 to 1,000,000
$34
$37
$40
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-39
April 2024
-------
FINAL RULE
APRIL 2024
Table C-38: Mean Annualized Cost per Household in CWSs that Treat or Change
Water Source, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Commercial Cost of
Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$755
$809
$899
Private
Ground
100 to 500
$411
$451
$511
Private
Ground
500 to 1,000
$152
$173
$198
Private
Ground
1,000 to 3,300
$99
$114
$130
Private
Ground
3,300 to 10,000
$67
$85
$104
Private
Ground
10,000 to 50,000
$44
$53
$61
Private
Ground
50,000 to 100,000
$29
$42
$57
Private
Ground
100,000 to 1,000,000
$12
$18
$26
Private
Surface
Less than 100
$568
$764
$1,004
Private
Surface
100 to 500
$305
$370
$442
Private
Surface
500 to 1,000
$118
$161
$206
Private
Surface
1,000 to 3,300
$76
$102
$132
Private
Surface
3,300 to 10,000
$55
$79
$106
Private
Surface
10,000 to 50,000
$40
$48
$57
Private
Surface
50,000 to 100,000
$34
$44
$55
Private
Surface
100,000 to 1,000,000
$25
$28
$32
Public
Ground
Less than 100
$884
$1,031
$1,199
Public
Ground
100 to 500
$388
$429
$486
Public
Ground
500 to 1,000
$129
$144
$161
Public
Ground
1,000 to 3,300
$89
$98
$108
Public
Ground
3,300 to 10,000
$176
$194
$215
Public
Ground
10,000 to 50,000
$50
$54
$58
Public
Ground
50,000 to 100,000
$35
$42
$48
Public
Ground
100,000 to 1,000,000
$33
$39
$47
Public
Surface
Less than 100
$803
$1,057
$1,392
Public
Surface
100 to 500
$344
$398
$461
Public
Surface
500 to 1,000
$120
$140
$164
Public
Surface
1,000 to 3,300
$89
$100
$110
Public
Surface
3,300 to 10,000
$199
$221
$243
Public
Surface
10,000 to 50,000
$54
$57
$60
Public
Surface
50,000 to 100,000
$40
$44
$48
Public
Surface
100,000 to 1,000,000
$34
$37
$40
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-40
April 2024
-------
FINAL RULE
APRIL 2024
Table C-39: Mean Annualized Cost per Household in CWSs that Treat or Change
Water Source, Option lb (PFOA and PFOS MCLs of 5.0 ppt) (Commercial Cost of
Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$743
$807
$895
Private
Ground
100 to 500
$408
$449
$511
Private
Ground
500 to 1,000
$148
$171
$197
Private
Ground
1,000 to 3,300
$95
$112
$131
Private
Ground
3,300 to 10,000
$64
$82
$106
Private
Ground
10,000 to 50,000
$40
$48
$57
Private
Ground
50,000 to 100,000
$25
$36
$51
Private
Ground
100,000 to 1,000,000
$9
$14
$23
Private
Surface
Less than 100
$527
$763
$1,044
Private
Surface
100 to 500
$296
$367
$450
Private
Surface
500 to 1,000
$111
$159
$216
Private
Surface
1,000 to 3,300
$72
$101
$135
Private
Surface
3,300 to 10,000
$50
$77
$109
Private
Surface
10,000 to 50,000
$37
$46
$56
Private
Surface
50,000 to 100,000
$33
$43
$55
Private
Surface
100,000 to 1,000,000
$22
$25
$29
Public
Ground
Less than 100
$864
$1,030
$1,234
Public
Ground
100 to 500
$383
$426
$485
Public
Ground
500 to 1,000
$127
$142
$160
Public
Ground
1,000 to 3,300
$88
$96
$107
Public
Ground
3,300 to 10,000
$168
$188
$210
Public
Ground
10,000 to 50,000
$47
$51
$55
Public
Ground
50,000 to 100,000
$33
$39
$46
Public
Ground
100,000 to 1,000,000
$31
$37
$45
Public
Surface
Less than 100
$775
$1,070
$1,474
Public
Surface
100 to 500
$333
$395
$465
Public
Surface
500 to 1,000
$117
$139
$164
Public
Surface
1,000 to 3,300
$87
$98
$110
Public
Surface
3,300 to 10,000
$192
$217
$242
Public
Surface
10,000 to 50,000
$51
$55
$59
Public
Surface
50,000 to 100,000
$38
$42
$46
Public
Surface
100,000 to 1,000,000
$31
$34
$38
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-41
April 2024
-------
FINAL RULE
APRIL 2024
Table C-40: Mean Annualized Cost per Household in CWSs that Treat or Change
Water Source, Option lc (PFOA and PFOS MCLs of 10.0 ppt) (Commercial Cost of
Capital, $2022)
Ownership
Source
Water
Population Served Size
Category
5th Percentile
Mean
95th Percentile
Private
Ground
Less than 100
$702
$795
$909
Private
Ground
100 to 500
$386
$441
$513
Private
Ground
500 to 1,000
$129
$165
$205
Private
Ground
1,000 to 3,300
$79
$106
$137
Private
Ground
3,300 to 10,000
$42
$76
$117
Private
Ground
10,000 to 50,000
$25
$34
$44
Private
Ground
50,000 to 100,000
$7
$17
$30
Private
Ground
100,000 to 1,000,000
$0
$8
$18
Private
Surface
Less than 100
$365
$732
$1,387
Private
Surface
100 to 500
$248
$363
$512
Private
Surface
500 to 1,000
$0
$148
$276
Private
Surface
1,000 to 3,300
$45
$96
$173
Private
Surface
3,300 to 10,000
$12
$68
$129
Private
Surface
10,000 to 50,000
$27
$36
$48
Private
Surface
50,000 to 100,000
$26
$41
$59
Private
Surface
100,000 to 1,000,000
$16
$20
$25
Public
Ground
Less than 100
$747
$1,020
$1,373
Public
Ground
100 to 500
$353
$418
$490
Public
Ground
500 to 1,000
$115
$137
$161
Public
Ground
1,000 to 3,300
$79
$92
$105
Public
Ground
3,300 to 10,000
$140
$172
$205
Public
Ground
10,000 to 50,000
$37
$41
$46
Public
Ground
50,000 to 100,000
$26
$34
$43
Public
Ground
100,000 to 1,000,000
$25
$35
$46
Public
Surface
Less than 100
$430
$997
$1,815
Public
Surface
100 to 500
$284
$389
$522
Public
Surface
500 to 1,000
$97
$135
$179
Public
Surface
1,000 to 3,300
$75
$95
$117
Public
Surface
3,300 to 10,000
$170
$207
$248
Public
Surface
10,000 to 50,000
$45
$49
$54
Public
Surface
50,000 to 100,000
$28
$33
$38
Public
Surface
100,000 to 1,000,000
$24
$28
$32
Abbreviations: CWS - community water system.
Final PFAS Rule Economic Analysis
C-42
April 2024
-------
FINAL RULE
APRIL 2024
Appendix D. PFOA and PFOS Serum Concentration-
Birth Weight Relationship
This appendix describes the methods used to estimate relationships between birth weight (BW)
and PFAS based on available studies. The EPA used these relationships to estimate incremental
changes in birth weight associated with reduced exposure to PFAS, namely PFOA and PFOS.
D.l Weight of Evidence of Birth Weight Effects
In the Health Effects Support Document (HESD) for PFOA (U.S. EPA, 2016b), the EPA
characterized the evidence for PFOA effects on birth weight as "plausible" based on human and
animal study data, and four of the five endpoints used for derivation of an RfD were lowered
fetal weights in rodents. The HESD for PFOS (U.S. EPA, 2016a) indicated that, despite
considerable uncertainty, the available human data "suggest an association of prenatal serum
PFOS with deficits in mean birth weight and with LBW [low birth weight]." The Agency for
Toxic Substances and Disease Registry (ATSDR, 2018) listed reduced birth weight as one of the
endpoints for which the available evidence "suggested" a relationship between human PFAS
exposure and effect. Negri et al. (2017), considering both toxicological and epidemiological
evidence, concluded that a causal relationship between PFOA and PFOS exposure and reduced
birth weight was "likely". The most recent syntheses of evidence, the EPA's Final Raman
Health Toxicity Assessments for PFOA and PFOS, found clear evidence of an association
between PFOA and PFOS and birth weight in both toxicological and epidemiological studies
(U.S. EPA, 2024b; U.S. EPA, 2024c). Based on these findings, the EPA's Office of Ground
Water and Drinking Water (OGWDW) derived exposure-response estimates for both
compounds.
D.2 Review of Available Meta-Analyses
The EPA's OGWDW reviewed literature identified in the EPA Office of Water, Office of
Science and Technology (OW/OST) literature reviews on the relationship between PFAS and
birth weight to identify previous estimates of serum PFAS-birth weight relationships. Many
epidemiological studies and several meta-analyses of existing studies have identified associations
between perfluorinated compound exposure and indices of fetal growth (primarily reduced birth
weight) (ATSDR, 2018; Johnson et al., 2014; Verner et al., 2015; Negri et al., 2017; Steenland et
al., 2018; Dzierlenga, Crawford, & Longnecker, 2020). Most studies of the relationship between
maternal serum PFOA and birth weight reported negative (i.e., inverse) relationships, while the
evidence for PFOS was more variable, as described below. Note that the EPA's review was
based primarily on secondary sources; OGWDW did not conduct a systematic literature search
or independent risk of bias (ROB) analyses for any identified systematic reviews and meta-
analyses. Rather, the EPA relied on previous authors who have analyzed the literature using
different protocols related to literature relevance, study quality, and ROB. However, OW/OST
has evaluated epidemiological literature for PFOA/PFOS as part of a systematic review to update
the 2016 HESDs for PFOS and PFOA.
The five studies considered by the U.S. Environmental Protection Agency Office of Science and
Technology (EPA/OST) for PFOA report the following slope estimates (in birth weight g per
ng/mL serum): -4.9 (Sagiv et al., 2018), -20.7 (Govarts et al., 2016), -41.0 (Wikstrom et al.,
Final PFAS Rule Economic Analysis
D-l
April 2024
-------
FINAL RULE
APRIL 2024
2019), -45.0 (Starling et al., 2017), and -45.2 (Chu et al., 2020). Compare these estimates to the
selected slope estimate from Negri et al. (2017) of -12.8 g per ng/mL. The four studies
considered by the EPA/OST for PFOS report the following slope estimates (in birth weight g per
ng/mL serum): -1.1 (Sagiv et al., 2018), -5.5 (Starling et al., 2017), -8.4 (Wikstrom et al., 2019),
and -11.0 (Chu et al., 2020). Compare these estimates to the selected exposure-response function
from Dzierlenga, Crawford, and Longnecker (2020) of -3.2 g per ng/mL.
The EPA reviewed six of the identified meta-analyses of PFAS-low birth weight relationships in
detail. One study, Monroy et al. (2008), presented regression results for body weight versus
maternal PFOA and PFOS concentrations, but the reported slope factors4 were not adjusted for
other covariates. Because of this it was not pursued further. Two of the analyses (Johnson et al.,
2014; Negri et al., 2017) used well-documented systematic review and ROB procedures to
identify relevant studies in the literature. The three other studies did not document ROB
protocols and study quality evaluation criteria (Verner et al., 2015; Dzierlenga, Crawford, &
Longnecker, 2020; Steenland et al., 2018). However, as discussed below, there was extensive
overlap in the data sets addressed in the various meta-analyses. Two of the meta-analyses
included exposure-response modeling for both PFOS and PFOA (Verner et al., 2015; Negri et
al., 2017), while one study addressed only PFOS (Dzierlenga, Crawford, & Longnecker, 2020)
and the remaining two addressed only PFOA (Johnson et al., 2014; Steenland et al., 2018).
There was relative conformity in the publications evaluated and ultimately selected for use in the
meta-analyses especially amongst the most recent ones, as later authors tended to include all the
studies evaluated in previous studies, adding newer results that had become available (Table
D-l):
• Johnson et al. (2014) conducted random effects meta-analysis based on data from nine
studies (including 4,149 births) published between 2007 and 2012. The authors requested
individual data on PFOA and covariates (variables other than PFAS exposure that may
predict study outcomes) from all authors of the primary studies used in their studies. In cases
where data were available, Johnson et al. (2014) used random effects methods to estimate
covariate-adjusted linear regression coefficients and used these values as inputs to their meta-
analysis. They found that including or excluding studies likely to have high ROB resulted in
only small effects on estimated slope factors for PFOA-birth weight relationships.
• Verner et al. (2015) included data from all the studies identified by Johnson et al. (2014),
with the exception of results from two studies: Fromme et al. (2010) and Kim et al. (2011).
Verner et al. (2015) excluded these studies because they were based on 50 or fewer
participants.
• Negri et al. (2017) included all the data sets identified by Johnson et al. (2014) plus five
newer data sets (Table D-l). Negri et al. (2017) also included data from an older study
(Monroy et al., 2008) that Johnson et al. (2014) omitted because "BW [birth weight] is not
the dependent model variable."
• Steenland et al. (2018) based their analyses of PFOA-birth weight effects on results from the
same studies in the Negri et al. (2017) meta-analysis (except for one study, Monroy et al.
4 When referring to a "slope factor" in this document, the EPA is discussing a measure of association between PFAS serum and
BW.
Final PFAS Rule Economic Analysis
D-2
April 2024
-------
FINAL RULE
APRIL 2024
(2008) plus 10 additional recent epidemiological studies (Table D-l). However, Steenland et
al. (2018) did not conduct a formal ROB evaluation to exclude these studies based on design
or analysis flaws, as was done in prior meta-analyses by Johnson et al. (2014) and Negri et al
(2017).5 Dzierlenga, Crawford, and Longnecker (2020) included PFOS-birth weight data
from all the studies identified by Verner et al. (2015), with the exception of results from Fei
et al. (2007), and an additional 22 studies, many of which overlap with studies evaluated in
Steenland et al. (2018). Although Dzierlenga, Crawford, and Longnecker (2020) did not
conduct formal ROB evaluations, the authors examined some study design aspects by
characterizing studies with respect to certain characteristics that might influence results and
evaluating those characteristics in meta-regression analyses.
5 Steenland et al. (2018) noted that ROB analyses have advantages in identifying biases, but stated that "using a quantitative score
of bias as a basis to exclude studies ultimately includes subjective components."
Final PFAS Rule Economic Analysis
April 2024
D-3
-------
FINAL RULE
APRIL 2024
Table D-l: Data Sources for PFOA/PFOS Meta-Analyses of Birth Weight Effects
Study
PFOA/PFOS-BW Relationship Studies Included in Meta-Analyses for Effects on BW
Johnson et al.
Verner et al.
Negri et al.
Steenland et al.
Dzierienga
EPA/OST Review
(2014)
(2015)
(2017)
(2018)
(2020)
(PFOA/PFOS) (2021)a
Apelberg et al. (2007)
X
X*
X*
X
X
X
Fei et al. (2007)
X
X*
X*
X
X
Hammet al. (2010)
X
X*
X*
X
X
X
Washino et al. (2009)
X
X*
X*
X
X
X
Fromme et al. (2010)
X
X
X
Kim et al. (2011)
X
X
X
Whitworth et al. (2012)
X
X*
X*
X
X
X
Maisonet et al. (2012)
X
X*
X*
X
X
X
Chen et al. (2012)
X
X*
X*
X
X
X
Darrow et al. (2013)
X
X
X
X
Bach et al. (2016)
X*
X
X
X
Lenters et al. (2016)
X*
X
X
X
Monroy et al. (2008)
X*
X
X
Robledo et al. (2015)m f
X*
X
X
X
Wu et al. (2012)
X
X
Savitz et al. (2012)
x**
X
Callan et al. (2016)
X
X
X
Govarts et al. (2016)
X
xd
Kwonet al. (2016)
X
X
Lee et al. (2016)
X
X
X
Wang et al. (2016)
X
X
X
Minatoya et al. (2017)
X
X
Shi et al. (2017)
X
X
X
Manzano-Salgado et al.
X
X
X
(2017)
Chen et al. (2017)
X
X
X
Starling et al. (2017)
X
X
xd
Sagiv et al. (2018)
X
X
xd
Asliley-Martin et al. (2017)
X
X
Lauritzen et al. (2017)'" '
X
X
M. Li et al. (2017)
X
X
Lind et al. (2017)m
-------
FINAL RULE
APRIL 2024
Cao et al. (2018)
X
X
Meng et al. (2018)
X
X
Marks et al. (2019)
X
X
Workman et al. (2019)
X
Xu et al. (2019)
X
Bell et al. (2018)
X
Louis et al. (2018)
X
Gao et al. (2019)
X
Chu et al. (2020)
xd
Hjermitslev et al. (2020)
X
Kashino et al. (2020)
X
Wikstrom et al. (2020)
xd
Abbreviations: BW - birth weight; the EPA/OST- U.S. Environmental Protection Agency Office of Science and Technology; PFOA - perfluorooctanoic acid; PFOS -
perfluorooctane sulfonic acid.
Notes:
aThe EPA/OST evaluation of study quality reflected in blue (high confidence), green (medium confidence) or pink (low confidence) cell shading. The EPA/OST literature
review focused on literature published between 2000 and 2020. Studies in this field reflect the studies the EPA reviewed to select those that were used for modeling.
* Indicates a data set used for PFOS, as well as PFOA meta-analysis.
** Indicates a data set included only in sensitivity analysis.
m< indicates results presented only stratified by sex or location [e.g., Lauritzen et al. (2017)].
indicates studies used by the EPA/OST for derivation of point of departures (PODs).
Final PFAS Rule Economic Analysis
D-5
April 2024
-------
FINAL RULE
APRIL 2024
The authors used different techniques to evaluate sources of variability in the meta-analyses. As
expected, random effects models generated results with lower heterogeneity (as measured by the
proportion of between-study variance in the data sets) than fixed effects models. Each of the
meta-analyses reported sensitivity analyses, stratified analyses, or leave-one-out results
(influence analyses) to explore the relative contributions of individual or groups of studies to the
quantitative pooled estimates of PFOA and PFOS effects on birth weight.
Johnson et al. (2014) reported a pooled beta across nine included studies of-18.9 g (95%CI: -
29.8, -7.9) for PFOA per each 1 ng/mL. Johnson et al. (2014) used well-documented meta-
analytical methods: random effects models with inverse variance weighting. In addition, Johnson
et al. (2014) conducted analyses omitting several small studies with relatively high ROB, as well
as one that included a large study (Savitz et al., 2012) that modeled maternal serum levels based
on historical exposures, rather than measured exposures. Johnson et al. (2014) found that
inclusion or exclusion of high-ROB studies and studies based on modeled serum levels resulted
in only a small effect on the estimated slope factor for PFOA-birth weight relationships (Johnson
et al. (2014) reported a pooled beta across nine included studies of-18.9 g (95%CI: -29.8, -7.9)
for PFOA per each 1 ng/mL. Johnson et al. (2014) used well-documented meta-analytical
methods: random effects models with inverse variance weighting. In addition, Johnson et al.
(2014) conducted analyses omitting several small studies with relatively high ROB, as well as
one that included a large study (Savitz et al., 2012) that modeled maternal serum levels based on
historical exposures, rather than measured exposures. Johnson et al. (2014) found that inclusion
or exclusion of high-ROB studies and studies based on modeled serum levels resulted in only a
small effect on the estimated slope factor for PFOA-birth weight relationships (Figure D-l).6
Verner et al. (2015) reported a pooled beta across seven included studies of-5.00 g (95% CI: -
8.92, -1.09) for PFOS and -14.72 g (95% CI: -21.66, -7.78) for PFOA each per each 1 ng/mL. In
addition, Verner et al. (2015) also investigated the potential impact of changing glomerular
filtration rate (GFR), an index of kidney function, on PFAS-birth weight relationships. They
based their analysis on the fact that maternal GFR and blood volume are known to change across
the three trimesters of pregnancy in such a way that the assumed independent effect of GFR on
birth weight, coupled with changes in PFAS excretion rates, could account for part of the birth
weight reduction found in the epidemiological studies of PFAS exposure. In addition to a
standard meta-analysis, they simulated PFOA/PFOS levels in a hypothetical population, using a
pharmacologically based pharmacokinetic (PBPK) model, and evaluated the impact of changes
in GFR on PFAS-associated changes in birth weight across trimesters. The results of the
conventional meta-analysis for the overall effects of PFAS on birth weight were similar to those
derived by Johnson et al. (2014) (Figure D-l). Verner et al. (2015) concluded, however, that a
portion of the observed association may be attributable to confounding by GFR, with the effect
of GFR increasing across trimesters. This suggested that studies which have not controlled for
GFR might overestimate the impact of prenatal exposure to PFAS on fetal growth.
All of the simulations employed different assumptions related to variability in PFOA/PFOS
levels and the strength of GFR impacts on birth weight. The simulated estimated relationships
6 Note that this finding may not apply to all meta-analyses, especially if they did not use the exact studies and same
ROB methods as those employed in Johnson et al. (2014).
Final PFAS Rule Economic Analysis D-6 April 2024
-------
FINAL RULE
APRIL 2024
between PFOA/PFOS and birth weight remained negative for all sample collection times, except
for the initial sampling time (at conception).
Negri et al. (2017) reported a pooled beta across eight included studies of -0.92 g (95% CI: -3.4,
1.6) for PFOS and twelve included studies of-12.8 g (95% CI: -23.2, -2.4) for PFOA each per
each 1 ng/mL. Negri et al. (2017) conducted random effects meta-analyses based on 14 studies.
In addition to the main analysis, Negri et al. (2017) conducted a sensitivity analysis related to
model form (fixed versus random effects), degree of adjustment (full, defined as adjustment for
infant sex, gestational age, maternal age, pre-pregnancy body mass index, education, parity,
and smoking, versus partial, which includes only some of these covariates), and location of
populations (America, Asia, and Europe). They also ran separate analyses for studies in which
the time of blood sampling varied (1st and 2nd trimester, 3rd trimester, and cord blood), to
further investigate the potential impacts of time of blood sampling as a proxy for changes in
GFR. Negri et al. (2017) found that the degree of adjustment had relatively little effect on the
magnitude of estimated slopes for PFOA and PFOS. The pooled PFOA/PFOS effect estimates
(i.e., beta coefficients) for studies in which sampling occurred late in pregnancy reported birth
weight decreases larger magnitude than for those where sampling occurred in the first two
trimesters, but the results were quite uncertain due to the small numbers of studies with late-term
sampling.
Steenland et al. (2018) reported a pooled beta across twenty-four included studies of-10.5 g
(95% CI: -16.7, -4.4) for PFOA per each 1 ng/mL. Steenland et al. (2018) conducted a random
effects meta-analysis based on 24 studies. In addition, they estimated PFOA slope factors
separately for studies of maternal and cord blood and for studies where PFOA serum levels were
measured in the first trimester versus any time later in pregnancy (Figure D-l). The slope factor
from the main analysis was significantly negative and similar in magnitude to that derived by
Negri et al. (2017). Coefficients for maternal blood were slightly smaller in magnitude than in
studies where cord blood was sampled, but still negative. The coefficient for the nine data sets
where blood PFOA was measured during the first trimester was small in magnitude
(-3.3 g per ng/mL), but not significant.
The most recent meta-analysis from Dzierlenga, Crawford, and Longnecker (2020) reported a
pooled beta across thirty-two included studies of -3.2 g (95% confidence interval: -5.1, -1.3) for
PFOS per each 1 ng/mL. The study conducted a random effects meta-analysis based on 32
results from 29 studies. The authors of the analysis estimated a slope of-3.2 g birth weight per
ng PFOS/mL (95% confidence interval: -5.1, -1.3) with significant moderate heterogeneity (I2 =
58%>). Sensitivity analyses suggested that the results are sensitive to timing of blood samples.
Among those with blood measurements before or early in pregnancy, however, PFOS was
inversely associated with birth weight (-1.35, 95% confidence interval: -2.33, -0.37), and for
the later pregnancy group, the association was -7.17 (95% confidence interval: -10.93, -3.41).
Final PFAS Rule Economic Analysis
D-7
April 2024
-------
FINAL RULE
APRIL 2024
Study
Johnson et al. (2014)
Verner et al. (2015)
Negri et al. (2017)
Steenland et al. (2018)
Verner et al. (2015)
Negri et al. (2017)
Fieurc D-l: Results and Confidence Limits from PFOA. PFOS Meta-Analyses:
Estimate
Mean (g per ng/mL)
Lower CI Upper CI Number of Studies
Heterogeneity 12
PFOA - Main
PFOA - High ROB study included
PFOA - Main
PFOA - Adjusted for GFR
PFOA - Main
PFOA - First/second trimester
PFOA - Third trimester
PFOA - Cord Blood
PFOA - Main
PFOA - First Trimester
PFOA - Second/Third trimester
PFOA - Include Savitz (2012)
PFOS - Main
PFOS - Adjusted for GFR
PFOS - Main
PFOS - First/second trimester
PFOS - Third trimester
PFOS - Cord Blood
Dzierlenga et al. (2020)
PFOS - Main
PFOS - Before or early in pregnancy
PFOS - Later pregnancy
-18.9
-15.4
-29.8
-26.5
-7.9
-4.3
-14.7
-7.9
-21.7
-9.4
-7.8
-6.4
-12.8
-10.5
-20
-35.3
-23.2
-23.6
-52.1
-101
-2.4
2.6
12.1
30.7
-10.5
-3.3
-17.8
-1
-16.7
-9.6
-25
-2.4
-4.4
3
-10.6
0.4
-5
-1.5
-8.9
-1.8
-1.09
-1.1
-0.92
0.6
-4
-11.3
-3.4
-1.4
-16.3
-17.4
1.6
2.5
8.2
-5.2
-3.2
-1.35
-7.17
-5.1
-2.33
-10.39
-1.3
-0.37
-3.41
9
10
12
24
7
17
25
32
10
22
38
72
52.9
63
68
29
74.3
58
5
55
p-value
0.12
0
>0.05
0.016
<0.0001
<0.0001
0.13
<0.05
<0.001
0
0.4
0.001
i—i—i—i—i—i—i—i—i—i
-50-40-30-20-10 0 10 20 30 40
Final PFAS Rule Economic Analysis
D-8
April 2024
-------
FINAL RULE
APRIL 2024
D.3 Exposure-Response Functions Based on Epidemiological
Studies
The EPA selected the exposure-response result for PFOA from the main analysis reported by
Steenland et al. (2018) for use in the risk assessment from exposure to PFOA and benefits
analysis of reducing PFOA in PWS even though this study did not use a systematic ROB
analysis of the studies included in the meta-analysis. Although Negri et al. (2017) employed a
systematic ROB analysis for the studies included in the meta-analysis and showed moderate
heterogeneity among studies (I2 = 38%)7, the EPA did not select it because the study is less
recent and includes fewer studies than Steenland et al. (2018). The agency selected the main
(random effects) analysis from Steenland et al. (2018) because it is the most recent meta-analysis
on PFOA-birth weight and included the largest number of studies. The pooled beta estimate for
PFOA effects on birth weight in Steenland et al. (2018) is -10.5 g (95% confidence interval: -
16.7; -4.4) birth weight per ng serum PFOA/mL based on 24. The agency also uses the 95%
confidence limits of -16.7 and -4.4 g birth weight per ng PFOA/mL as lower and upper bound
slope estimates for a sensitivity analysis. The pooled mean estimate (g birth weight per ng
PFOA/mL) for all studies is in the midrange of the results for the early, middle, and late blood
sampling results (Figure D-l).
The EPA selected the exposure-response result for PFOS from the most recent meta-analysis of
32 observations from 29 publications reported by Dzierlenga, Crawford, and Longnecker (2020)
for use in the risk assessment from exposure to PFOS and benefits of reducing PFOS in PWS.8
The agency chose the main analysis from Dzierlenga, Crawford, and Longnecker (2020)
because it considered the largest number of recent studies, the heterogeneity among studies was
moderate (I2 = 58%), and sensitivity analyses suggested an inverse relationship with birth
weight. Additionally, sensitivity analyses suggested that the results were not particularly
sensitive to timing of blood samples, consistent with the early pregnancy subgroup analysis
result. Dzierlenga, Crawford, and Longnecker (2020) also examined study quality aspects by
characterizing studies with respect to certain characteristics9 that might influence results and
examining those in meta-regression analyses.
712 represents the proportion of total variance in the estimated model due to inter-study variation; a value of 38 percent is
considered "moderate", suggesting that the studies are not seriously inhomogeneous and that a pooled model (meta-analysis)
is appropriate.
8 Although Negri et al. (2017) also estimated an exposure-response slope for PFOS effects on BW based on eight studies,
the analysis includes a slope factor derived from the Maisonet et al. (2012) study that was given as (positive) 5.77 (95%
confidence limits = 2.01, 9.53). However, in the original Maisonet et al. (2012) study, the relationship between maternal PFOS
and female infant BW was reported as being negative; it appears that there was a transcription error in the Negri et al. (2017)
analysis.8 An sensitivity analysis from Negri et al. (2017) that excluded the Maisonet et al. (2012) study resulted in a pooled
estimate of -2.0 g BW per ng/mL PFOS, which is similar in magnitude to the estimate reported by Dzierlenga, Crawford, and
Longnecker (2020). Also, although the estimated slope factor for PFOS effects from Verner et al. (2015), based on seven studies,
included the slope factor from Maisonet et al. (2012) as (negative) -5.77 g BW per ng PFOS/mL (95% confidence limits -9.53, -
2.01), Dzierlenga, Crawford, and Longnecker (2020) includes a larger number of studies, many of which were published more
recently than those considered in Negri et al. (2017) and Verner et al. (2015) (32 results from 29 studies conducted from 2007 to
2019, compared to seven and eight studies considered in Negri et al., 2017 and Verner et al., 2015, respectively, that were
conducted from 2007
to 2016).
9 For example, the quality of evidence was characterized as low for the BW-PFOS associations when the timing of blood draw
was before or early in pregnancy.
Final PFAS Rule Economic Analysis
D-9
April 2024
-------
FINAL RULE
APRIL 2024
The EPA reanalyzed the pooled estimate from this study after determining that the original
Dzierlenga, Crawford, and Longnecker (2020) pooled estimate included a duplicated estimate
from Chen et al. (2017). The EPA reran the analysis excluding the duplicated estimate to obtain a
slope of-3.0 g birth weight per ng PFOS/mL with the same heterogeneity (I2 = 58%) as the prior
estimate (p-value for heterogeneity <0.001).
Final PFAS Rule Economic Analysis
D-10
April 2024
-------
FINAL RULE
APRIL 2024
Appendix E. Effects of Reduced Birth Weight on
Infant Mortality
This appendix summarizes the EPA's analysis of the relationship between infant mortality and
birth weight. This relationship is fundamental in estimating benefits from changes in birth weight
among infants whose mothers were exposed to PFOA or PFOS during or prior to pregnancy. The
EPA developed a cross-sectional model to quantify this relationship based on recent 2016/17 and
2017/18 Centers for Disease Control and Prevention (CDC) Period Cohort Linked Birth-Infant
Death Data files.
E.l Birth Weight-Mortality Relationship
Low birth weight (LBW), defined as weight at birth <2,500 grams, is recognized as a significant
predictor of infant mortality (McCormick, 1985; World Health Organization, 2014).
The majority of infants born with LBW are premature, but other gestational factors such as
maternal hypertensive disorders and anemia can result in full-term infants who are born at LBW
(Joyce et al., 2012). Many of the top 10 causes of infant mortality are factors associated with
preterm birth, including LBW (Jacob, 2016). Advances in U.S. prenatal and neonatal care and
successes in public health initiatives, such as those designed to decrease maternal smoking, have
increased LBW survival rates and reduced the prevalence of LBW infants (Callaghan et al.,
2017; Singh & Stella, 2019). To quantify potential mortality impacts from changes in infant birth
weight resulting from changes in maternal PFOA and PFOS exposure via drinking water, robust
data supporting a relationship between incremental changes in infant birth weight and mortality
risk are needed.
A number of epidemiological studies in the U.S. have reported relationships between birth
weight and mortality. However, most of these studies evaluate relationships between infant
mortality and birth weight above or below various birth weight thresholds (e.g., Mclntire et al.,
1999; Lau et al., 2013). The EPA identified only two studies that show statistically significant
relationships between incremental changes in birth weight and infant mortality that can be
leveraged for PFOS/PFOA health impact modeling: Ma et al. (2010) and Almond et al. (2005).
Ma and Finch (2010) used 2001 National Center for Health Statistics/National Vital Statistics
System (NCHS/NVSS) linked birth/infant death data for singleton and multiple birth infants
among subpopulations defined by sex and race/ethnicity to estimate a regression model assessing
the associations between 14 key birth outcome measures, including birth weight, and infant
mortality. They found notable variation in the relationship between birth weight and mortality
across race/ethnicity subpopulations, with odds ratios for best-fit birth weight-mortality models
ranging from 0.8-1 per 100 gram (g) birth weight change. Almond et al. (2005) used 1989-1991
NCHS linked birth/infant death data for multiple birth infants to analyze relationships between
birth weight and infant mortality within birth weight ranges. For their preferred model, they
reported coefficients in deaths per 1,000 births per 1 g increase in birth weight that range from -
0.420 to -0.002.
However, the data used in these studies (Almond et al., 2005 and Ma & Finch, 2010) are old
(1989-1991 and 2001, respectively). Given the significant decline in infant mortality over the last
30 years (discussed in Section E.2 below), and changes in other maternal and birth characteristics
Final PFAS Rule Economic Analysis
E-l
April 2024
-------
FINAL RULE
APRIL 2024
that are likely to influence infant mortality (e.g., average maternal age and rates of maternal
smoking), the birth weight-mortality relationship estimates from Almond et al. (2005) and Ma
and Finch (2010) are likely to overestimate benefits of birth weight changes. Moreover, Almond
et al. (2005) focused on multiple birth infants to analyze relationships between birth weight and
infant mortality.
LBW is determined by two main processes: duration of gestation and rate of fetal growth
(Institute of Medicine, 1985; Quah, 2016). Thus, infants can be LBW because they are born
preterm or are born small for gestational age, which is a proxy for intrauterine growth
retardation. Researchers have found that birth weight and gestational age are closely associated
but not perfectly correlated (e.g., Kiely et al., 1994; Mathews, 2013). A study by Almond et al.
(2005) found that gestational age is an important determinant of birth weight as it explains over
half of the overall variance in birth weight among a pooled sample of twins. Moreover, multiple
studies suggest that, when available, both birth weight and gestational age should be included
when predicting infant mortality odds (Almond et al., 2005; Ma & Finch, 2010; Ray et al., 2017).
Cole et al. (2010) developed a logistic regression model showing that gestational age and birth
weight z-score10 were the strongest predictors of survival among very preterm infants. Ma and
Finch (2010) predicted infant mortality by combining birth weight and gestational age variables
to distinguish between the two major causes of LBW. Ray et al. (2017) used modified Poisson
regression to show that singleton infants who are born preterm and small for gestational age have
a higher risk of neonatal death than infants born preterm alone.
The CDC indicated that the mortality rate among multiples is very high for reasons that are often
unrelated to birth weight and recommended that a model based on singletons may provide a
more representative relationship between birth weight and infant mortality (Communication with
Horon, 2020). Studies of birth weight-specific infant mortality among singletons and multiples
suggest that, due to differences in intrauterine growth restriction, prematurity rates, and zygosity,
analyses that examine perinatal outcomes should be stratified by plurality (Russell et al., 2003;
Cooke, 2010). Furthermore, singleton infants represent the majority of U.S. births (96% of
infants born in 2016 and 2017). Following CDC's recommendations, the EPA developed cross-
sectional models to estimate a relationship between birth weight at four distinct gestational age
categories and infant mortality based on the most recently available 2016-2018 NCHS/NVSS
data and focusing on singleton infants. To identify variation in the birth weight-mortality
relationship across race/ethnicity subpopulations, the EPA estimated separate relationships for
non-Hispanic Black, non-Hispanic White, and Hispanic subpopulations.
In developing the singleton models, the EPA used similar variables and partitioning techniques
as detailed in Ma et al. (2010). Specifically, the EPA developed separate models for different
race/ethnicity categories and interacted birth weight with gestational age. Ma et al. (2010) found
that key predictors of infant mortality include birth weight, Apgar score,11 and gestational age.
Ma et al. (2010) developed multivariate logistic regression models for gender- and race-specific
subpopulations12 to assess associations of various combinations of birth weight, gestational age,
10 Z-scores describe how far from the mean a given data point is.
11 Apgar score refers to a metric indicating the health of a newborn. The score, which ranges from 0 to 10, is based on skin color,
heart rate, reflexes, muscle tone, and breathing rate/effort.
12 Separate models were fit for non-Hispanic white girls, non-Hispanic white boys, non-Hispanic black girls, non-Hispanic black
boys, Mexican girls, and Mexican boys.
Final PFAS Rule Economic Analysis
E-2
April 2024
-------
FINAL RULE
APRIL 2024
fetal growth rate, and Apgar scores with four mortality outcomes (infant mortality, early
neonatal, late neonatal, and post-neonatal mortality). In addition to these covariates, Ma et al.
(2010) automatically selected covariates such as parental characteristics (e.g., maternal age and
education), maternal risk factors (e.g., smoking), and child characteristics (e.g., birth order)
based on predictive power. Ma et al. (2010) showed that the baseline rates of each birth outcome
differ by both race/ethnicity and postnatal period. Model results indicated that birth weight is a
stronger predictor of infant mortality among the non-Hispanic Black subpopulation compared to
the non-Hispanic White and Hispanic subpopulations.
E.2 Basis for Updated Birth Weight-Mortality Relationship
There has been a notable decline in U.S. infant mortality rates during the two decades since
analyses reported in Ma et al. (2010) and Almond et al. (2005). In the last 30 years, overall infant
mortality rates have declined steadily (ICF, 2020).13 The infant mortality rate in 2018 was
5.67 per 1,000 live births, while the infant mortality rate in 1991 was 8.6 per 1,000 live births.
Except for infants born with birth weight lower than 500 grams, for whom mortality rates have
not changed considerably, mortality rates for infants with birth weight greater than 500 grams are
decreasing and converging on a low rate.14
Given a decline in infant mortality in the birth weight categories lower than 1,500 g, a unit
change in birth weight is likely to produce less of an impact on the probability of mortality in
2016-2018 compared to 1989-1991 (the years evaluated in Almond et al., 2005) or 2001 (the
year evaluated in Ma & Finch, 2010). Despite recent declines in U.S. infant mortality, disparities
in infant mortality experience continue to exist across race/ethnicity subpopulations (Osterman et
al., 2015). Recent research indicates that infant mortality is consistently highest among Black
infants (both Hispanic and non-Hispanic), while non-Hispanic White and Hispanic White infants
have the lowest mortality rates (Rice et al., 2017; Rowley & Hogan, 2012; Collins Jr & David,
2009).
In addition to the decline in infant mortality in LBW categories, other maternal and birth
characteristics that are likely to influence infant mortality have evolved over time. Almond et al.
(2005) provided sample means for birth and maternal characteristics for singletons based on the
1989 NCHS/NVSS Linked Natality-Mortality Detail file. The EPA provides similar statistics for
singletons from the 2016-2018 NCHS/NVSS Period/Cohort Linked Birth-Infant Death Data
Files15 that demonstrate how birth and mortality characteristics have changed over time.
Table E-l shows a subset of the 1989 sample means among singletons born to non-Hispanic
Black and non-Hispanic White mothers from Almond et al. (2005) Table II and the same
statistics derived from the 2016-2018 data. The comparison shows that teen pregnancy rates,
pregnancy among mothers with less than a high school education, and maternal smoking during
13 CDC publishes National Vital Statistics Reports that summarize mortality trends over time (e.g., Kochanek et al., 2019) and
provides detailed tables of infant mortality trends by race and age at death in annual Health, United States reports (National
Center for Health Statistics, 2019).
14 The EPA assembled summary statistics on infant mortality by BW category provided in the documentation for 1983-2018
Linked Infant Birth-Death Detail Files. These files are published on the online data portal by NCHS/NVSS:
https://www.cdc.gov/nchs/data_access/vitalstatsonline.htm
15 The 2016-2018 NCHS/NVSS Period/Cohort Linked Birth-Infant Death Data Files represent two separate datasets.
The 2016/2017 data includes infants bom in 2016 and follows their mortality experience for one year (through the end of 2017).
The 2017/2018 data includes infants bom in 2017 and follows their mortality experience through the end of 2018.
Final PFAS Rule Economic Analysis
E-3
April 2024
-------
FINAL RULE
APRIL 2024
pregnancy have decreased since 1989. While mean and median birth weight has decreased
slightly for singleton infants, the 1-year infant mortality rate has decreased by about 42%.
Possible explanations for this trend may include advancements in prenatal and postnatal care
(e.g., advances in infection control practices and the use of intubation to reduce infant lung
injury; Callaghan et al., 2017) as well as positive effects of public health education (e.g., reduced
smoking during pregnancy; Singh & Stella, 2019).
In addition to a decreasing 1-year mortality rate, Table E-l shows a decrease in the fraction of
infants with congenital anomalies and a decrease in median gestational age. The decrease in
gestational age is supported by analysis from Donahue et al. (2010), who found that gestational
age among full-term singletons in the United States decreased by more than two days from 1990-
2005.
Table E-l: Comparison of Sample Means for Singletons between the 1989
Natality-Mortality Detail File and the Combined 2016-2018 Period/Cohort Linked
Birth-Infant Death Data Files
Variable
Sample Meansab 0
1989
2016-2018 (% Change)
Sample size
2,655,977
4,212,764
Infant deaths (per 1000 live births)
Within 1 year of birth (infant mortality)
8.46
4.94 (-42%)
Within 28 days (neonatal)
4.99
2.94 (-41%)
28 days to 1 year (postneonatal)
3.49
2.00 (-43%)
Fraction of dead with birth weight < 2500 g
Infant mortality
0.570
0.592 (+4%)
Within 24-hour mortality
0.890
0.285 (-68%)
Neonatal mortality
0.760
0.463 (-39%)
Postneonatal mortality
0.300
0.129 (-57%)
Infant birth weight (g)
Mean
3,369
3,313 (-2%)
Median
3,402
3,345 (-2%)
5th percentile
2,410
2,390 (-1%)
Fraction LB W (<2500 g)
0.061
0.065 (+7%)
Gestational age (in weeks)
Mean
39
39 (0%)
Median
40
39 (-3%)
5th percentile
35
35 (0%)
Characteristics of birth
5-minute Apgar score (0-10)
8.97
8.79 (-2%)
Fraction male
0.512
0.512(0%)
Fraction congenital anomalyd
0.019
0.001 (-93%)
Final PFAS Rule Economic Analysis
E-4
April 2024
-------
FINAL RULE
APRIL 2024
Table E-l: Comparison of Sample Means for Singletons between the 1989
Natality-Mortality Detail File and the Combined 2016-2018 Period/Cohort Linked
Birth-Infant Death Data Files
Variable
Sample Means3
1989
2016-2018 (% Change)
Mother's demographic characteristics
Fraction Black
0.195
0.193 (-1%)
Fraction high school dropout
0.184
0.085 (-54%)
Fraction college graduate
0.187
0.451 (+141%)
Age
26.3
28.6 (+9%)
Fraction teenager
0.129
0.049 (-62%)
Fraction 30+
0.289
0.444 (+54%)
Fraction married
0.736
0.595 (-19%)
Mother's risk factors
Number of prenatal visits
11.2
11.5 (+3%)
Fraction smoke during pregnancy
0.212
0.100 (-53%)
Abbreviations BW - birth weight; LBW - low birth weight.
Notes:
aThe data are restricted to non-Hispanic Black and White mothers bom in the United States, as reported in Almond et al.
(2005) Table II.
bThe 1989 data summary in Almond et al. (2005) included anemia of mother, assisted ventilation (<30 minutes) and assisted
ventilation (>= 30 minutes), which are not included in the 2016-2018 NCHS/NVSS dataset. The 2016-2018 NCHS/NVSS
dataset does include assisted ventilation and assisted ventilation (6 hours), but these variables are not necessarily comparable
to the assisted ventilation variables included in the 1989 NCHS/NVSS dataset. Similarly, 1989 data summary in Almond et al.
(2005) included "pregnancy-associated hypertension" which is further split up into "gestational hypertension" and
"hypertension eclampsia" in the 2016-2018 NCHS/NVSS dataset. Due to differences in variable definitions among the data,
the EPA excludes hypertension.
cRecords with "Unknown" or "Not Stated" values not included in the 2016-2018 summary.
''Congenital anomalies among the 1989 and 2016-2018 data are not directly comparable due to differences in the congenital
anomalies included in this metric between the datasets. The 1989 dataset includes the following congenital anomalies:
Anencephalus, spina bifida/meningocele, hydrocephalus, other central nervous system anomalies, heart malformations, other
circulatory/respiratory anomalies, rectal atresia/stenosis, trachea-esophageal fistula/esophageal atresia,
omphalocele/gastroschisis, other gastrointestinal anomalies, malformed genitalia, renal agenesis, other urogenital anomalies,
cleft lip/palate, Polydactyly, club foot, diaphragmatic hernia, other musculoskeletal/integumental anomalies, down's
syndrome, other chromosomal anomalies, and other congenital anomalies. The 2016-2018 dataset includes the following
congenital anomalies: anencephaly, meningomyelocele/spina bifida, cyanotic congenital heart disease, congenital
diaphragmatic hernia, omphalocele, gastroschisis.
The remainder of this appendix summarizes the development of regression models implemented
using newer data.
E.3 Development of the Analytical Dataset
E.3.1 Data Sources
This analysis relies on Period/Cohort Linked Birth-Infant Death Data Files published by
NCHS/NVSS from the 2017 period/2016 cohort and the 2018 period/2017 cohort.16 Each dataset
10 https://www.cdc.gov/nchs/data_access/vitalstatsonline.htm
Final PFAS Rule Economic Analysis
E-5
April 2024
-------
FINAL RULE
APRIL 2024
includes files linking all infant deaths during the period and cohort years to information from
corresponding birth certificates and separate files consisting of all births occurring during the
period. The data include all infants under 1 year of age in the U.S. or its territories (Centers for
Disease Control and Prevention, 2017f). This analysis excludes multiple birth infants. In addition
to infant birth and mortality information, the data include details on maternal characteristics
(e.g., mother's education, marital status, and age category), maternal risk factors (e.g., smoking
status), and pregnancy and birth characteristics (e.g., gestational age, infant birth weight,
presence of congenital anomalies, and birth order).
E.3.2 Dataset Development
The EPA combined the infant birth and death files using the SAS code examples from the user
guides accompanying the datasets to create user-created cohort files, which follow the birth
cohorts for an entire year to ascertain their mortality experience (Centers for Disease Control and
Prevention, 2017f, 2018). At this stage, the EPA also selected variables of interest for the
regression analysis. These variables include maternal demographic and socioeconomic
characteristics, maternal risk and risk mitigation factors, and infant birth characteristics. The
EPA included several variables used in Ma et al. (2010) as well as additional variables to
augment the set of covariates included in the regression analyses. Variable selection was
informed by literature on the leading causes of infant mortality (e.g., Ahrens et al., 2017; Mishra
et al., 2018; Centers for Disease Control and Prevention, 2020a, 2020b; Ely & Driscoll, 2020).
E.3.3 Identification of Infant Mortality Risk Factors
To identify infant mortality risk factors for inclusion in the regression analyses, the EPA relied
on multiple data sources, including key risk factors identified by the CDC and prior studies of
the relationship between infant mortality and various maternal and birth characteristics. Although
risks to infant mortality include conditions related to infant and maternal health, demographic
and socioeconomic characteristics also contribute to infant mortality outcomes. Based on the
studies the EPA reviewed, infant mortality risk factors generally fall within three general
categories described below:
• Birth Characteristics:
o Birth Weight and Gestational Age: The CDC identifies preterm birth and LBW as
leading causes of infant death in the United States (Ely & Driscoll, 2020). The
majority of infant deaths in 2018 occurred among infants born preterm
(gestational age <37 weeks; Ely & Driscoll, 2020). Previous studies of the
relationship between birth weight and infant mortality identify birth weight and
gestational age as important predictors of infant mortality (e.g., Almond et al.,
2005; Ma & Finch, 2010).
o Other Infant Birth Characteristics: Studies of leading causes of infant mortality
suggest that birth order plays a significant role in infant mortality outcomes.
Higher birth order is linked to risk of injury and may be indicative of other
socioeconomic factors (Ahrens et al., 2017; Mishra et al., 2018). Another
substantive predictor of infant mortality is five-minute Apgar score (Almond et
al., 2005; Ma & Finch, 2010). Birth defects, such as the presence of congenital
anomalies, also contribute to infant mortality (Ely & Driscoll, 2020).
Final PFAS Rule Economic Analysis
E-6
April 2024
-------
FINAL RULE
APRIL 2024
• Maternal Risk and Risk Mitigation Factors: Many causes of infant death are exacerbated
by tobacco use, substance use, and stress (Centers for Disease Control and Prevention,
2020a). CDC guidance suggests that regular prenatal care visits17 lead to detection of infant
mortality risk factors (e.g., hypertension).
• Maternal Demographic and Socioeconomic Characteristics: Infant birth outcomes are
influenced by demographic and socioeconomic factors such as maternal race/ethnicity, age,
education, and marital status (Ma & Finch, 2010). Infant mortality rates vary for mothers of
different ages, with the lowest mortality rates among mothers age 30-34 and highest
mortality rates among teen mothers and mothers over 40 in 2018 (Ely & Driscoll, 2020).
Singh et al. (2019) found that the risk of 1-year mortality in 2016 was 3.7 times greater for
mothers with less than 12 years of education than for mothers with 16 or more years of
education. Marital status also influences the risk of infant mortality—studies show that the
risk of infant mortality increases when one parent is absent (Ngui et al., 2015; Alio et al.,
2011). In 2018, the non-Hispanic Black subpopulation had the highest infant mortality rate at
10.8 deaths per 1,000 live births, while Hispanic and non-Hispanic White subpopulations
experienced much lower rates of infant mortality (4.9 and 4.6 deaths per 1,000 births,
respectively; Ely & Driscoll, 2020).
While maternal risk variables such as hypertension, diabetes, and infection lead to premature
birth, LBW, and reduced motor function, birth-related factors such as Apgar score, birth weight,
and gestational age likely account for these risks (Backes et al., 2011; Centers for Disease
Control and Prevention, 2016c; M. Li et al., 2017). Given that birth weight impacts on infant
mortality are the focus of our analysis, selected covariates do not include maternal risk factors,
such as maternal hypertension, diabetes, and infection, whose mortality influence pathway is
primarily through birth weight, gestational age, and Apgar score.18
E.4 Development of Variables
The dependent variable (BIRTHMORT) is a binary variable indicating whether the infant died
within one year of birth. Covariates included in the regression analyses fall under three
categories:
• Birth characteristics (denoted with BIRTH prefix)
• Maternal risk and risk mitigation factors (denoted with MRF prefix)
• Maternal demographic and socioeconomic characteristics (denoted with MDEM prefix)
Table E-2 provides a detailed description of all variables included in the singleton regression
analysis and the corresponding variables from the NCHS/NVSS data used to develop the
variables. The EPA estimated different regression models for three race/ethnicity
subpopulations: Non-Hispanic Black, non-Hispanic White, and Hispanic. Infants whose mothers
fall into these race/ethnicity subpopulations are identified using the MRACEHISP variable from
the NCHS/NVSS data.
17 While prenatal care visits fall under the maternal risk and risk mitigation factors category, it could also be considered a
maternal demographic and socioeconomic characteristic indicative of access to care.
18 Pearson correlation tests indicated significant relationships between these variables (p-values < 5%).
Final PFAS Rule Economic Analysis
E-7
April 2024
-------
FINAL RULE
APRIL 2024
The focus of the EPA's analysis is the relationship between birth weight and infant mortality.
However, Ma and Finch (2010) noted that the practice of specifying regression models that
assume that every 1-gram increase in birth weight has the same effect on infant mortality
outcome (regardless of gestational age or LBW status of the infant) has been challenged.19
Following researchers who emphasize the importance of examining birth outcomes from the
perspective of combined birth weight and gestational age variables (Solis et al., 2000; Powers et
al., 2006), Ma and Finch (2010) found that models with birth weight-gestational age interaction
variables had higher predictive power than models that only used birth weight and gestational
age separately. Following best practices from the health economic literature (e.g., Solis et al.,
2000; Powers et al., 2006; Ma & Finch, 2010), the EPA interacted continuous birth weight with
four gestational age category indicator variables (extremely pre-term, very pre-term, moderately
pre-term, term as defined by the World Health Organization, 2018) to account for the
heterogeneity in birth weight impact with respect to the gestational age of the infant. The EPA
expected that birth weight effects would be highest for extremely pre-term infants and lowest for
full-term infants.
In addition to the set of birth weight-gestational age category interaction variables, the EPA
added variables for other infant birth characteristics (birth order, birth year, sex, Apgar score,
congenital anomaly indicator), maternal risk and risk mitigation factors (smoker status,
categorized number of prenatal care visits), and maternal demographic and socioeconomic
characteristics (education, age, marital status). These variables control for additional factors
beyond birth weight and gestational age that contribute to the probability of infant mortality.20
The EPA included categorized Apgar score variables based on analysis from Ma and Finch
(2010), who found that Apgar scores, separated into low (0-3), medium (4-6), and high (7-10)
categories, were the strongest predictor of infant mortality among race/ethnicity-specific models.
Further, the 2016-2018 NCHS/NVSS data show that Apgar scores are significantly higher for
non-Hispanic White infants than for non-Hispanic Black infants. Ma and Finch (2010) also
found that the inclusion Apgar scores in models predicting infant mortality significantly
improved goodness of fit. The EPA also included a variable indicating whether the infant was
born in 2016 or 2017 (BIRTH YR 2016) as a control to determine whether there are any
significant differences between the 2016 and 2017 NCHS/NVSS datasets that are not readily
captured by other covariates.
Table E-2: Variables Used in Singleton Mortality Regression Analysis
Variable
Variable
Type
Variable Definition
Basis for Variable in
NCHS/NVSS Dataset
Dependent Variable
BIRTH MORT
Binary
Binary variable indicating whether
the infant died within one year of
birth
DODYY
19 Ma and Finch (2010) indicate that birth weight effects vary according to the position on the distribution of birth weight
(they characterize the birth weight-mortality distribution as a reverse J-shaped distribution).
20 The EPA also explored adding additional maternal risk factor variables, including maternal hypertension, diabetes, and
infection, based on CDC's identified infant mortality risk factors (see Section E.3.1.2). However, the inclusion of these variables
in our models produced counterintuitive results and they were eliminated from the covariate set.
Final PFAS Rule Economic Analysis
E-8
April 2024
-------
FINAL RULE APRIL 2024
Table E-2: Variables Used in Singleton Mortality Regression Analysis
Variable
Variable
Type
Variable Definition
Basis for Variable in
NCHS/NVSS Dataset
Covariates
Birth Weight and GA
BIRTH BW I EXT PRETER
M
Discrete/
Continuous
Continuous BW (in grams) if
gestational age is <=28 weeks
(extremely preterm), 0 if otherwise
BRTHWGT,
COMBGEST
BIRTH BW I VER PRETER
M
Discrete/
Continuous
Continuous BW (in grams) if
gestational age is >28 weeks and
<=32 weeks (very preterm), 0 if
otherwise
BRTHWGT,
COMBGEST
BIRTH BW I MOD PRETER
M
Discrete/
Continuous
BW (in grams) if gestational age is
>32 weeks and <=37 weeks
BRTHWGT,
COMBGEST
(moderately preterm), 0 if otherwise
BIRTHBWITERM
Discrete/
Continuous
Continuous BW (in grams) if
gestational age is >37 weeks (term),
0 if otherwise
BRTHWGT,
COMBGEST
Other Infant Birth Characteristics3
BIRTH MALE
Binary
Binary variable indicating that the
infant is male
SEX
BIRTHCONANOM
Binary
Binary variable indicating that the
infant experienced one or more of the
following congenital anomalies:
anencephaly,
meningomyelocele/spina bifida,
cyanotic congenital heart disease,
congenital diaphragmatic hernia,
omphalocele, gastroschisis
CA ANEN,
CA MNSB,
CA CCHD,
CA CDH,
CA OMPH,
CAGAST
BIRTH_APGAR_0_3
Binary
Binary variable indicating that the
five-minute Apgar score is between 0
and 3. Five-minute Apgar score
indicates the health of a newborn
based on skin color, heart rate,
reflexes, muscle tone, and breathing
rate/effort.
APGAR5
BIRTH_APGAR_4_6
Binary
Binary variable indicating that the
five-minute Apgar score is between 4
and 6. Five-minute Apgar score
indicates the health of a newborn
based on skin color, heart rate,
reflexes, muscle tone, and breathing
rate/effort.
APGAR5
BIRTHYR2016
Binary
Binary variable indicating whether
the infant was born in 2016. If 0, the
infant was born in 2017.
N/A; based on CDC
dataset
BIRTHBOCatl
Binary
Binary variable indicating that the
infant has one sibling (second-born)
LBOREC
BIRTH_BOCat2
Binary
Binary variable indicating that the
infant has two or more siblings
(third- or later-born)
LBOREC
Final PFAS Rule Economic Analysis
E-9
April 2024
-------
FINAL RULE APRIL 2024
Table E-2: Variables Used in Singleton Mortality Regression Analysis
Variable
Variable
Type
Variable Definition
Basis for Variable in
NCHS/NVSS Dataset
Maternal Risk and Risk Mitigation Factors'3 d
MRFNOPRECARE
Binary
Binary variable indicating that the
mother had no prenatal care visits
PREVIS
MRF19PRECARE
Binary
Binary variable indicating that the
mother had 1 to 9 prenatal care visits
PREVIS
MRF 16 ORMORE PRECAR
E
Binary
Binary variable indicating that the
mother had 16 or more prenatal care
visits
PREVIS
MRFSMOKE
Binary
Binary variable indicating that, if
maternal smoking status is known,
the mother was a smoker
CIGREC
Maternal Demographic and Socioeconomic Characteristics0 d
MDEMINOHS
Binary
Binary variable indicating that the
mother's education is known and that
the mother did not graduate high
school or obtain a GED
MEDUC
MDEMICOLLEGEPLU S
Binary
Binary variable indicating that the
mother's education is known and that
the mother attended college or higher
education
MEDUC
MDEMAGETEEN
Binary
Binary variable indicating that the
mother's age is <20
MAGER
MDEMAGEAD V_3 5 40
Binary
Binary variable indicating that the
mother's age is >34 and <= 40
MAGER
MDEM_AGE_ADV_40plus
Binary
Binary variable indicating that the
mother's age is >40
MAGER
MDEM I MARRIED
Binary
Binary variable indicating that the
mother's marital status is known and
that the mother is married
DMAR
Abbreviations: BW - birth weight; GA - gestational age; NCHS - National Center for Health Statistics; NVSS - National
Vital Statistics System.
Notes:
Reference categories for binary variables in the other infant birth characteristics category include female infants, infants who
did not experience a congenital anomaly, infants with Apgar scores from 7 to 10, infants born in 2017, and infants who have
no siblings.
bReference categories for binary variables in the maternal risk and risk mitigation factors category include mothers who had
10 to 15 prenatal care visits and mothers who do not smoke.
Reference categories for binary variables in the maternal demographic and socioeconomic characteristics category include
mothers who went to high school but who did not attend any college, mothers aged 25 to 34, and mothers whose marital status
is unknown or single.
dThe maternal age (MDEM_AGE) variables are split into three categories to show effects associated with teen mothers,
mother's aged 35 to 40, and mothers over the age of 40 with respect to the reference case of mother's aged 20 to 34. This is to
reflect differences in infant mortality rates associated with different maternal age groups. In 2018, the CDC indicated that total
mortality rates were highest for infants of mothers under age 20, while infants of mother's age 30-34 had the lowest mortality
rates (Ely & Driscoll, 2020). Infant mortality rates increased among infants born to older mothers, especially those over age 40
(Ely & Driscoll, 2020).
Of the available singleton data, 0.8% had no race information. These records are excluded from
consideration. For regression modeling, records with incomplete or missing data (specified as
"Unknown" or "Not Stated" in the raw NCHS/NVSS data) for any of the covariates listed in
Final PFAS Rule Economic Analysis
E-10
April 2024
-------
FINAL RULE
APRIL 2024
Table E-2 were excluded from the analytical dataset. Records with incomplete or missing
covariate information account for 8.5% of the non-Hispanic Black records, 6.5% of the non-
Hispanic White records, and 7.0% of the Hispanic records (for a combined total of 7.0% of all
records). The EPA did not attempt to fill in these data gaps using imputations or assumptions,
because records with missing data constituted less than 10% of all records. The resulting sample
sizes are: 981,212 for the non-Hispanic Black subpopulation, 3,644,499 for the non-Hispanic
White subpopulation, 1,646,713 for the Hispanic subpopulation.
E.5 Summary Statistics
Table E-3 presents maternal and infant characteristics of the study population, including number
and proportion of the sample associated with different age ranges, gestation weeks, races and
ethnicities, educational attainment, marital status, number of prenatal care visits, and whether or
not the mother smoked during pregnancy. Sample statistics indicate that the majority of mothers
are between ages 20 and 33, have full-term pregnancies, are non-Hispanic White, graduated high
school, had more than ten prenatal care visits, and did not smoke during pregnancy.
Final PFAS Rule Economic Analysis
E-ll
April 2024
-------
FINAL RULE APRIL 2024
Table E-3: Maternal and Infant Characteristics of the Study Population
Description
N
Proportion (%)
Age
<20 years
343,784
5.48
20-33 years
4,606,124
73.43
34-39 years
1,138,646
18.15
40+ years
183,870
2.93
Gestation Week
<=28
43,654
0.70
>28 and <=32
80,408
1.28
>32 and <=37
106,8585
17.04
>37
5,079,777
80.99
Race/Ethnicity
Non-Hispanic White
3,644,499
58.10
Non-Hispanic Black
981,212
15.64
Hispanic
1,646,713
26.25
Education
No high school or GED
871,274
13.89
Graduated high school
2,963,900
47.25
Attended college3
2,437,250
38.86
Marital Status
Married
3,504,095
55.87
Unmarried
2,768,329
44.13
Number of Prenatal Care Visits'3
None
100,231
1.60
1-9
1,519,825
24.23
10-15
4,066,046
64.82
16+
586,322
9.35
Smoking During Pregnancy
Yes
455,758
7.27
No
5,816,666
92.73
Apgar Score
Apgar score between 0 and 3
32,518
0.52
Apgar score between 4 and 6
82,762
1.32
Apgar score between 7 and 10
6,157,144
98.16
Notes:
aRefers to mothers who obtained an associate's degree or more. Mothers who obtained some college credit but not a degree are
categorized in the "Graduated high school" field.
bNumber of prenatal care visits in the study population range from 0 to 98.
Final PFAS Rule Economic Analysis
E-12
April 2024
-------
FINAL RULE
APRIL 2024
E.6 Estimation Methods
The EPA fit the logistic regression model using Stata 15.1 (StataCorp, 2013a). The model is fit
to three different race/ethnicity singleton subpopulations (non-Hispanic Black, non-Hispanic
White, and Hispanic)21 as there are known disparities in the prevalence of LBW by race and
ethnicity (Collins Jr & David, 2009; Rice et al., 2017; Rowley & Hogan, 2012; Ratnasiri et al.,
2018). Coefficients of non-linear regression models with a binary outcome indicate direction of
the effect that covariates have on outcome probability. That is, negative coefficients indicate that
the probability of mortality decreases as the covariate increases, while positive coefficients
indicate that the probability of mortality increases as the covariate increases.
In this analysis, the EPA reported the results of regression modeling using both odds ratios22 and
marginal effects. While the odds ratio is an effect metric commonly reported in epidemiological
research, the impact of a marginal change in the covariate on the probability of the outcome (i.e.,
the marginal effect) is easier to interpret. The magnitude of this marginal effect depends on all
estimated coefficients of the model as well as specific values of all the covariates included in the
model. When estimating marginal effects, the EPA used actual observed values for the covariates
rather than using covariate means.23 For non-birth weight-gestational age variables, the EPA
estimated marginal effects based on covariate values from all observations included in the
models. For birth weight-gestational age variables, the EPA estimated marginal effects based on
covariate values from the subset of observations falling within each gestational age category (see
N columns for sample size used for each marginal effect calculation).24
Section E.5 presents the EPA's preferred models. These models had the best fit and offered most
intuitive results, in terms of variable sign and significance. The EPA estimated additional model
specifications prior to the final models, including models with the infant birth weight categories
used in Almond et al. (2005) and a separate continuous gestational age variable, models with
different specifications for maternal age, and models with different combinations of maternal
risk factors. The EPA does not believe that exclusion of maternal risk factor variables creates
omitted variable bias, given that their effects are accounted for using more direct newborn health
state variables such as Apgar score. The additional model specifications that the EPA tested prior
to determining the final model form resulted in marginal effects estimates that were inconsistent
with scientifically expected directionality of their effects.
21 The EPA did not develop a model for other race subpopulations because doing so for each individual race/ethnicity or
combinations of all "other" races would suffer from effects of low sample size (i.e., odds ratios and marginal effects that lack
significance).
22 The natural exponent of the logistic regression coefficient is a ratio of odds of the outcome when the value of the predictor
variable is changed by a certain amount relative to the odds of the outcome using the baseline value of the predictor variable.
The odds are the ratio of the probability that the outcome of interest occurs to the probability that the outcome of interest does not
occur.
23 The EPA calculated marginal effects using the "margins, dydx(*)" command in Stata (StataCorp, 2013b). Hie EPA used the
default as observed option.
24 The EPA estimated BW-gestational age category-specific marginal effects using subsets of data that contain infants with BW
in the corresponding gestational age category to account for correlations between gestational age and other variables included in
the model. For example, infants in the preterm gestational age categories have lower Apgar score on average.
Final PFAS Rule Economic Analysis
E-13
April 2024
-------
FINAL RULE
APRIL 2024
E.7 Results and Discussion
E.7.1 Mortality Regression Models
Overall, the sign and significance of covariates in the regression models align with expectations
based on previous literature. Table E-4 presents odds ratios and marginal effects (in terms of
deaths per 1,000 births) for the non-Hispanic Black, non-Hispanic White, and Hispanic models.25
A marginal effect estimate represents the effect of a 1-unit change in a given covariate on the
infant mortality rate per 1,000 births. Pseudo R2 values are approximately 40%, which is in line
with previous literature.26 The agency notes that the estimated models are potentially subject to
omitted variable bias from other sources, such as income level, but the EPA does not have
adequate information to evaluate the impacts of this bias on the marginal birth weight-mortality
relationship. The following subsections discuss the effects of regression model covariates on the
probability of infant mortality.
E.7.1.1 Birth Characteristics
The results for the birth weight-gestational age variables match literature-based expectations. In
all three models, the coefficients and marginal effects for birth weight among different
gestational age categories are negative and statistically significant (p<0.01). Negative marginal
effect values for the birth weight- gestational age categories indicate that a 1-gram birth weight
increase is associated with decreases in the infant mortality rate per 1,000 births, ranging from -
0.20 (extremely preterm) to -0.005 (term) for the non-Hispanic Black population, from -0.12 to -
0.002 for the non-Hispanic White population, and from -0.15 to -0.002 for the Hispanic
population. The magnitude of birth weight marginal effect is lower in gestational age categories
corresponding to longer gestation, indicating that the probability of mortality decreases as both
gestational age and birth weight increase.
Determining the magnitude of the mortality probability decrease is straightforward using
marginal effects. For example, using marginal effects from the non-Hispanic Black model, for
extremely preterm infants a 100 g birth weight increase would translate to 20 fewer infant deaths
per 1000 births in this gestational age category or a 2% decrease in the probability of mortality
within one year of birth.27 The same birth weight increase at a higher gestational age would still
decrease mortality risk but to a lesser extent. A 100 g birth weight increase for a non-Hispanic
Black infant in the moderately pre-term category would translate to only 1 fewer infant death per
1000 births or a 0.1% decrease in the probability of mortality within one year of birth.
Figure E-l shows variability of marginal effects for birth weight among different gestational age
categories across race/ethnicity subpopulations, with larger magnitudes estimated for the non-
Hispanic Black subpopulation compared to those estimated for the non-Hispanic White
subpopulation or Hispanic subpopulation, indicating that LBW increases the probability of
25 The EPA reports the results of regression modeling using both odds ratios and marginal effects, which are more informative
than reporting estimated coefficients. Because estimated coefficients are in log-odds units, they are difficult to interpret and are
therefore often converted into odds ratios in epidemiological literature by taking the exponent of each regression coefficient. The
EPA reported odds ratios via the "logit" command in Stata (StataCorp, 2013a).
20 Ma and Finch (2010) reported a Pseudo R2 value of approximately 27%.
27 The implied decrease in probability of death is calculated as (100 g)*(marginal effect in terms of deaths per 1,000 births
per g)/( 1,000 births) and multiplied by 100 to obtain a percentage: [(100 g)*(-0.19440/1000)] *(100) = -1.94%.
Final PFAS Rule Economic Analysis
E-14
April 2024
-------
FINAL RULE
APRIL 2024
mortality within the first year more so among non-Hispanic Black infants than among non-
Hispanic White and Hispanic infants. This pattern is more pronounced for the extremely preterm
infants and very preterm infants.
Non-Hispanic Black Non-Hispanic White —Hispanic
o
«¦»—
to
_C
4-*
15
CD
>
o
o
CD
T—1
a>
5
Q.
CD
JZ
C
^-»
"
03
CD
CD
to
"O
ru
¦M
CD
v_
C
u
45
c
_c
CuO
4—
O
T—1
03
a;
u
c
QJ
T3
'u
_C
c
CD
CuO
c
CD
-C
U
o.oo
-0.05
-0.15
-0.20
-0.25
/
/ /
/' / /
/ '
/ , /
/
/ /
/
/
/
•! A
•' *
/
/
/
/
/
/
d
Extremely Preterm Very Preterm Moderately Preterm Term
Gestational Age Category
Figure E-l: Comparison of Change in Incidence of Infant Death per 1 g Increase
in Birth Weight by Gestational Age Category and Race/Ethnicity (Deaths per 1,000 Births)
Notes: Gestational age categories defined as extremely preterm (<=28 weeks), very preterm (>28 weeks and <=32 weeks),
moderately preterm (>32 weeks and <=37 weeks), and term (>37 weeks). Related covariates in the regression model include
BIRTH_BW_I_EXT_PRETERM, BIRTH_BW_I_VER_PRETERM, BIRTH_BW_I_MOD_PRETERM, BIRTH_BW_I_TERM.
Data based on the 2016/17 and 2017/18 CDC Period Cohort Linked Birth-Infant Death Data Files obtained from NCHS/NVSS.
For the birth order variables (BIRTHBOCatl, BIRTH_BOCat2), the reference category is first-
born children. Across all three models, odds ratios and marginal effects for these variables are
large and significant (p<0.01). Effects for BIRTH_BOCat2 are larger than for BIRTH BOCatl,
which is consistent with research indicating that second- or later-born infants have increasingly
higher probabilities of mortality compared to first-borns (Mishra et al., 2018; Ahrens et al.,
2017). Coefficients and marginal effects for variables indicating male infants (BIRTH MALE)
and infants with congenital anomalies (BIRTH CONANOM) indicate that the probability of
mortality increases when the infants are male and when infants experience at least one congenital
anomaly. The effect of calendar birth year was not statistically different from zero at a 5%
significance level.
Final PFAS Rule Economic Analysis
E-15
April 2024
-------
FINAL RULE
APRIL 2024
Marginal effects for the birth characteristics variables also vary by race/ethnicity. For example,
the marginal effects for the BIRTHBOC atl variables indicate that, relative to first-born infants,
the infant mortality rate per 1,000 births increases by 1.13, 0.90, and 0.59 for second-born non-
Hispanic Black, non-Hispanic White, and Hispanic infants, respectively.28 Compared to the non-
Hispanic White and Hispanic subpopulations, 5-minute Apgar score has a stronger association
with infant mortality among the non-Hispanic Black subpopulations. The marginal effects for the
BIRTHCONANOM variables indicate that, relative to infants without any congenital
anomalies, the infant mortality rate per 1,000 births increases by 18.82, 8.99, and 9.66 for non-
Hispanic Black, non-Hispanic White, and Hispanic infants with congenital anomalies,
respectively.
E.7.1.2 Maternal Risk and Risk Mitigation Factors
The probability of infant mortality varies among certain maternal risk or risk mitigation factors.
The probability of infant mortality increases for mothers who smoke or mothers without a high
school diploma. Maternal smoking increases the infant mortality rate per 1,000 births by 1.34,
0.47, and 0.57 for non-Hispanic Black, non-Hispanic White, and Hispanic infants, respectively.
The probability of infant mortality decreases for mothers with a college education or higher.
Relative to mothers with a high school education, the infant mortality rate per 1,000 births
decreases by 1.29, 0.82, and 0.27 for non-Hispanic Black, non-Hispanic White, and Hispanic
infants born to mothers with a college education or higher, respectively. Relative to the 10 to 15
prenatal care visit category, which is most common in the data (See Table E-3), the probability
of infant mortality increases with zero visits, 1 to 9 visits, and 16 or higher visits. Marginal
effects indicate that having no prenatal care visits increases the infant mortality rate per 1,000
births by 3.03, 0.95, and 0.91 for non-Hispanic Black, non-Hispanic White, and Hispanic infants,
respectively.
E.7.1.3 Maternal Demographic and Socioeconomic Characteristics
Results for the maternal demographic and socioeconomic characteristic variables vary by
race/ethnicity and largely match the EPA's expectations. The education variables serve as
proxies for socioeconomic status, and results among all three models indicate that, relative to
mothers with a high school diploma, the probability of infant mortality increases for mothers
without a high school diploma and decreases for mothers with a college education or higher.
Maternal education effects on infant mortality probability vary by race/ethnicity. For example,
relative to mothers with a high school education, the infant mortality rate per 1,000 births
decreases by 1.29, 0.82, and 0.27 for non-Hispanic Black, non-Hispanic White, and Hispanic
infants born to mothers with a college education or higher, respectively.
The maternal age variables align with available infant mortality statistics showing the highest
infant mortality rates when mothers are under age 20 and elevated rates when mothers are over
40 (Ely & Driscoll, 2020). Compared to mothers aged 20 to 34 years, probability of infant
mortality is higher for mothers younger than 20 years, lower for mothers aged 35 to 40 years,
and higher for mothers older than 40 years. Relative to infants born to mothers aged 20 to 34
28 The implied decrease in probability of death is calculated as (marginal effect in terms of deaths per 1,000 births)/(1,000 births)
and multiplied by 100 to obtain a percentage. Example calculation using the marginal effects for BIRTH_BOCatl from the non-
Hispanic Black model: (1.19100/1000)*( 100) = 0.119%.
Final PFAS Rule Economic Analysis
E-16
April 2024
-------
FINAL RULE
APRIL 2024
years, infants born to mothers younger than 20 years' experience 0.79, 0.61, and 0.68 additional
infant deaths per 1,000 births in non-Hispanic Black, non-Hispanic White, and Hispanic
subpopulations, respectively. The decreased death probability for mothers aged 35 to 40 might
be capturing effects of the financial stability of mothers in this age group.
Negative and significant coefficients and marginal effects among all models for the mother's
marital status variable, MDEM I MARRIED, indicate that the risk of infant mortality decreases
among infants with two parents, consistent with studies indicating that paternal involvement
reduces the probability of infant mortality (Ngui et al., 2015; Alio et al., 2011). Compared to
infants born to mothers who are not married or mothers whose marital status is unknown, infants
born to married mothers experience 0.35, 0.51, and 0.30 fewer deaths per 1,000 births for non-
Hispanic Black, non-Hispanic White, and Hispanic subpopulations, respectively.
Final PFAS Rule Economic Analysis
E-17
April 2024
-------
FINAL RULE
APRIL 2024
Table E-4: Odds Ratios and Marginal Effects for the Non-Hispanic Black, Non-Hispanic White, and Hispanic Mortality
Regression Models
Variable
Odds Ratios (95% CI)ab
Marginal Effects (Deaths per 1,000 Births (95% CI)a c
Black
White
Hispanic
Black
White
Hispanic
BIRTHBWIEXTPRETERM
0.99817
(0.99802,
0.99832)
0.99866
(0.99855,
0.99878)
0.99835
(0.99817,
0.99853)
-0.20400
(-0.21910, -
0.18890)
-0.12160
(-0.13080, -
0.11240)
-0.15260
(-0.1677, -
0.13750)
BIRTHBWIVERPRETERM
0.99816
(0.99804,
0.99827)
0.9985
(0.99842,
0.99858)
0.99846
(0.99835,
0.99858)
-0.04580
(-0.04820, -
0.04340)
-0.03290
(-0.03430, -
0.03140)
-0.03290
(-0.0351, -
0.03070)
BIRTHBWIMODPRETERM
0.99852
(0.99846,
0.99857)
0.99867
(0.99863,
0.99872)
0.99856
(0.99849,
0.99862)
-0.01030
(-0.01080, -
0.00985)
-0.00677
(-0.00702, -
0.00652)
-0.00626
(-0.00659, -
0.00592)
BIRTHBWITERM
0.99856
(0.99851,
0.99860)
0.99865
(0.99861,
0.99868)
0.99849
(0.99844,
0.99855)
-0.00453
(-0.00472, -
0.00434)
-0.00228
(-0.00236, -
0.00221)
-0.00219
(-0.00229, -
0.00208)
BIRTHBOCatl
1.20078
(1.12406,
1.28272)
1.37498
(1.30875,
1.44458)
1.23256
(1.14005,
1.33256)
1.13170
(0.72263,
1.54080)
0.90320
(0.76267,
1.04370)
0.59091
(0.37013,
0.81170)
BIRTH_BOCat2
1.43158
(1.34271,
1.52634)
1.66176
(1.57927,
1.74859)
1.36704
(1.26426,
1.47818)
2.21920
(1.81950,
2.61890)
1.44050
(1.29450,
1.58650)
0.88360
(0.66192,
1.10530)
BIRTHAPGAR03
19.89802
(18.35772,
21.56734)
43.36705
(40.67038,
46.24253)
45.87636
(41.39996,
50.83677)
18.49800
(17.92800,
19.06800)
10.69200
(10.46100,
10.92300)
10.81300
(10.466,
11.15900)
BIRTHAPGAR46
3.8631
(3.54196,
4.21336)
5.92239
(5.54208,
6.32880)
6.86084
(6.16310,
7.63750)
8.35950
(7.79370,
8.92530)
5.04500
(4.83850,
5.25150)
5.44270
(5.1129,5.7726)
BIRTH MALE
1.28589
(1.22265,
1.35240)
1.29367
(1.24351,
1.34583)
1.19405
(1.12581,
1.26643)
1.55530
(1.24280,
1.86790)
0.73028
(0.61753,
0.84304)
0.50123
(0.33447,
0.66798)
BIRTHCONANOM
20.95317
(16.73647,
26.23226)
23.81106
(21.33609,
26.57338)
30.45195
(25.31381,
36.63302)
18.81800
(17.39300,
20.24300)
8.99150
(8.65400,
9.32900)
9.65470
(9.096, 10.21300)
BIRTHYR2016
1.04910
(0.99784,
1.10298)
1.01725
(0.97816,
1.05791)
0.97538
(0.91965,
1.03449)
0.29646
(-0.01339,
0.60632)
0.04852
(-0.06265,
0.15968)
-0.07045
(-0.23671,
0.09582)
Final PFAS Rule Economic Analysis
E-18
April 2024
-------
FINAL RULE
APRIL 2024
Table E-4: Odds Ratios and Marginal Effects for the Non-Hispanic Black, Non-Hispanic White, and Hispanic Mortality
Regression Models
Variable
Odds Ratios (95% CI)ab
Marginal Effects (Deaths per 1,000 Births (95% CI)a c
Black
White
Hispanic
Black
White
Hispanic
MRFNOPRECARE
1.63300
(1.46647,
1.81844)
1.39979
(1.24374, 1.5754)
1.37859
(1.19240,
1.59383)
3.03350
(2.36630,
3.70070)
0.95389
(0.61828,
1.28950)
0.90736
(0.49675,
1.31800)
MRF19PRECARE
1.37775
(1.29674,
1.46382)
1.34652
(1.28399,
1.41209)
1.17236
(1.09445,
1.25582)
1.98210
(1.60560,
2.35870)
0.84385
(0.70831,
0.97940)
0.44942
(0.25471,
0.64414)
MRF_ 160RM0REPREC ARE
1.12520
(1.00220,
1.26329)
1.12394
(1.04280,
1.21139)
1.35485
(1.20490,
1.52345)
0.72964
(0.013350,
1.44590)
0.33139
(0.11875,
0.54403)
0.85827
(0.52611,
1.19040)
MRFSMOKE
1.24139
(1.13425,
1.35866)
1.17977
(1.11549,
1.24776)
1.22117
(1.02459,
1.45549)
1.33750
(0.77763,
1.89740)
0.46889
(0.30933,
0.62846)
0.56471
(0.06794,
1.06150)
MDEMINOHS
1.05467
(0.97987,
1.13519)
1.10367
(1.03289,
1.17930)
1.02742
(0.95914,
1.10056)
0.32924
(-0.12598,
0.78447)
0.27977
(0.09167,
0.46788)
0.07644
(-0.11791,
0.27079)
MDEMICOLLEGEPLU S
0.81232
(0.75874,
0.86969)
0.7478
(0.71366,
0.78357)
0.90822
(0.83434,
0.98863)
-1.28570
(-1.70930, -
0.86211)
-0.82429
(-0.95807, -
0.6905)
-0.27208
(-0.51214, -
0.03202)
MDEMAGETEEN
1.13705
(1.02800,
1.25767)
1.24116
(1.13208,
1.36077)
1.27144
(1.13883,
1.41948)
0.79446
(0.17048,
1.41840)
0.61279
(0.35157,
0.87402)
0.67869
(0.36668,
0.99071)
MDEMAGEAD V_3 5 40
0.90639
(0.83721,
0.98130)
0.85079
(0.80231,
0.90220)
0.95193
(0.87380,
1.03704)
-0.60792
(-1.0992, -
0.11665)
-0.45831
(-0.62493, -
0.29170)
-0.13923
(-0.38131,
0.10286)
MDEM_AGE_ADV_40plus
1.37377
(1.17433,
1.60708)
0.96251
(0.83754,
1.10613)
1.2633
(1.07379,
1.48624)
1.96430
(0.99358,
2.93490)
-0.10838
(-0.50285,
0.28609)
0.66055
(0.20117,
1.11990)
mdem_i MARRIED
0.94432
(0.88719,
1.00513)
0.83555
(0.79827,
0.87458)
0.89883
(0.84382,
0.95743)
-0.35439
(-0.74074,
0.03196)
-0.50957
(-0.63965, -
0.37949)
-0.30144
(-0.48028, -
0.12260)
# Model Observations
981,212
3,644,499
1,646,713
Pseudo R2
0.389
0.357
0.416
Final PFAS Rule Economic Analysis
E-19
April 2024
-------
FINAL RULE
APRIL 2024
Table E-4: Odds Ratios and Marginal Effects for the Non-Hispanic Black, Non-Hispanic White, and Hispanic Mortality
Regression Models
Variable
Odds Ratios (95% CI)ab
Marginal Effects (Deaths per 1,000 Births (95% CI)a c
Black
White
Hispanic
Black
White
Hispanic
Abbreviations: CI - confidence intervals.
Notes:
Confidence intervals and significance testing do not include adjustments for multiple comparisons.
bLogistic regression models and ORs estimated using the "logit" likelihood function in Stata 15.1.
cMarginal effects estimated using the "margins, dydx(*)" command in Stata 15.1 with the default observed option. For non-BW-GA variables, the EPA estimated marginal
effects based on covariate values from all observations in the models. For BW-GA variables, the EPA estimated marginal effects based on covariate values from the subset of
observations falling within each GA category (see Supplementary Table 3).
Final PFAS Rule Economic Analysis
E-20
April 2024
-------
FINAL RULE
APRIL 2024
E.7.2 Comparison to Prior Studies
The EPA's evaluation of the relationship between birth weight and infant mortality differs from
those used in prior literature in terms of included covariates, model specification, and sample
characteristics. In terms of modeling approach, our analysis is closest to the one used by Ma et
al. (2010), who also find that birth weight and GA are important predictors of infant mortality
risk and that the effects of birth weight on infant mortality vary by race/ethnicity. However,
methodological differences between Ma et al. (2010) and our work, summarized in Table E-5,
prevent us from making direct comparisons of birth weight-infant mortality effect magnitudes.
Even in the absence of methodological differences, the EPA expects that results would differ
from those reported by older studies due to changes in infant mortality, maternal and birth
characteristics, and maternal demographic over the past 30 years (see Table E-l).
Table E-5: Comparison of Ma et al. (2010) and the EPA Analysis
Analysis Component
Ma et al. (2010)
EPA
Year(s) of NCHS/NVSS
Data
2001
2016-2018
Data Sample
Singletons and multiples
Singletons only
Race/Ethnicity Models
Non-Hispanic Black, non-Hispanic
White, Mexican
Non-Hispanic Black, non-Hispanic
White, Hispanic
Birth Weight-Gestation
Specification3
Birth weight (100 g increment),
gestational age (weeks), and birth weight
x gestational age (continuous product of
birth weight and gestational age)
Birth weight interacted with four
gestational age categories (extremely
preterm, very preterm, moderately
preterm, and term)
Other Covariatesb
Categorized APGAR score (low: 0-3 and
medium: 4-6, with high: 7-10 as
reference category), maternal age,
maternal education, marital status,
whether mother was born in U. S.,
whether father was unreported on birth
certificates, prenatal care,
tobacco/alcohol use during pregnancy,
and birth order
Categorized Apgar score (low: 0-3
and medium: 4-6, with high: 7-10 as
reference category), categorized
number of prenatal care visits (None,
1-9,16+, with ,10-15 as reference
category), maternal education, maternal
age, marital status, smoker status, sex,
presence of congenital anomalies, birth
year, birth order (see Table E-2)
Abbreviations: NCHS - National Center for Health Statistics; NVSS - National Vital Statistics System.
Notes:
aAlthough Ma et al. (2010) tested several different models, the EPA focuses on one of their highest-performing model forms,
Model 12, in which the interaction term between gestational age and birth weight is almost always significant.
bThe EPA notes that Ma et al. (2010) did not report coefficients for a number of maternal and birth characteristics (i.e.,
maternal age, maternal education, marital status, whether mother was bom in U.S., whether father was unreported on birth
certificates, prenatal care, tobacco/alcohol use during pregnancy, and birth order) or discussed these variables in detail.
E.8 Limitations and Uncertainties
Table E-6 summarizes limitations and sources of uncertainty associated with the estimated
relationship between infant birth weight and mortality.
Final PFAS Rule Economic Analysis
E-21
April 2024
-------
FINAL RULE
APRIL 2024
Table E-6: Limitations and Uncertainties in the Analysis of the Birth Weight-Mortality
Relationship
Uncertainty/Assumption
Notes
Transcription errors may be present in the
NCHS/NVSS dataset.
Infant birth and death records are compiled based on hand-
written forms and tabulated for use in the NCHS/NVSS
dataset.
The models do not directly account for maternal
socioeconomic status and other potentially
important factors that contribute to LBW and
infant mortality.
Though review of the infant mortality literature suggests that
socioeconomic status is an indicator of infant mortality (Ma
& Finch, 2010; Ely & Driscoll, 2020), the NCHS/NVSS
does not have a variable that would account for individual
socioeconomic status of the mother (e.g., household income)
or even community-level socioeconomic status (e.g., median
income at the county- or state-level). The EPA tested a
variable for hospital payment source for delivery that
specifies those who use Medicaid, but model results that
included this variable did not match expectations (variable
coefficient was not significant for all race/ethnicity
subpopulations, mixture of negative and positive coefficients
depending on race/ethnicity subpopulation). Thus, the
variable was excluded from our models. The maternal
education, maternal age, and marital status variables serve as
rough proxies for socioeconomic status in our models. Other
factors, such as indicators of parental support networks (e.g.,
access to paid care or grandparents that live nearby) may
contribute to the relationship between birth weight and
infant mortality, but such information is not publicly
available at the individual infant scale.
The analysis relies only on singleton data to
develop relationships between birth weight and
infant mortality.
Because singletons represent the majority of U.S. births
(96% of infants born in 2016 and 2017), the EPA does not
expect this to be a significant limitation. In order to address
this limitation, a separate model would be required because
multiples are often born at smaller birth weight than
singleton infants, the mortality rate among multiples is often
higher than singletons for reasons often unrelated to birth
weight (Horon, 2020), and the sample size of multiples in
the 2016-2018 NCHS/NVSS data are likely not adequate to
represent the relationship between birth weight and
mortality.
The EPA does not model birth weight-mortality
impacts for infants who fall into race categories
other than non-Hispanic White, non-Hispanic
Black, and Hispanic.
While the NCHS/NVSS data specifies additional race
categories, developing models for each individual race or
even a combination of all "other" races would suffer from
effects of low sample size, including coefficient and
marginal effects that lack significance. All combined, the
"other" race/ethnicity subpopulation would have a sample
size that is at least 30 percent smaller than any one of the
non-Hispanic White, non-Hispanic Black, and Hispanic
race/ethnicity models.
Abbreviations: LBW - low birth weight; NCHS - National Center for Health Statistics; NVSS - National
Vital Statistics System.
Final PFAS Rule Economic Analysis
E-22
April 2024
-------
FINAL RULE
APRIL 2024
Appendix F. Serum Cholesterol Dose Response
Functions
This appendix describes the EPA's literature review to identify studies to estimate relationships
between cholesterol levels and serum PFAS for inclusion in a meta-analysis of these
relationships. This approach has been peer reviewed by the EPA's Science Advisory Board
(SAB); input provided by that organization has been considered in finalizing this analysis (U.S.
EPA, 2022). Statistical analyses that combine the results of multiple studies, such as meta-
analyses, are widely applied to investigate the dose-specific relationship between contaminant
levels and associated health effects. Such analyses are suitable for economic assessments
because they can improve precision and statistical power (Engels et al., 2000; Deeks, 2002;
Riicker et al., 2009). This appendix also provides details on the meta-data development, results
of the meta-analysis, and limitations and uncertainties associated with the estimated
relationships. The EPA used the estimated relationships to estimate cardiovascular disease
(CVD) risk reduction associated with exposure to PFAS mediated by changes in serum
cholesterol markers.
F.l Data Sources
The EPA relied on two literature review efforts to identify potential sources of exposure-
response information for the effect of PFAS on serum cholesterol, lipids, and lipoproteins: A
literature review built on the one conducted by the Agency for Toxic Substances and Disease
Registry (ATSDR) in the development of their Toxicological Review Public Comment Draft
(ATSDR, 2018), which included literature through mid-2017.
The most recent systematic review of the newly published epidemiological literature for PFAS
performed by the EPA included literature from 2013 to 2020 (U.S. EPA, 2024b; U.S. EPA,
2024c). The relationships between exposure to PFAS and serum total cholesterol (TC) and high-
density lipoprotein cholesterol (HDLC) identified based on these literature reviews allowed the
EPA to generate inputs for the Pooled Cohort Atherosclerotic Cardiovascular Disease (ASCVD)
risk model (Gofif et al., 2014).29 30
F.l.l Literature Review and Studies Identification for the
Meta-Analysis
Two reviewers independently screened references retrieved from the literature search by title and
abstract, and then reviewed relevant studies in full text. The EPA evaluated studies identified
during the search according to the following criteria prior to inclusion in the meta-analysis to
ensure validity, consistency, and applicability. Briefly, of interest were studies conducted on
adults in the general population, evaluating the outcomes of TC and HDLC, and evaluating the
29 The ASCVD model relies on the following inputs: demographic information, smoking and diabetes status,
serum TC, and HDLC.
30 Note that the EPA evaluated HDLC effects as part of a sensitivity analysis (see Appendix K). The EPA did not model the
effects of PFOA/PFOS changes on 1TDLC levels in the overall benefits analysis because evidence of an association between
PFOA/PFOS and HDLC effects is uncertain (U.S. EPA, 2024b; U.S. EPA, 2024c).
Final PFAS Rule Economic Analysis
F-l
April 2024
-------
FINAL RULE
APRIL 2024
exposures of PFOA and PFOS. Because the EPA evaluates CVD risk among a general
population of adults aged 40 to 89, studies performed on specific population subsets, such as
occupational populations, were not considered for inclusion in the meta-analysis due to the
potential for greater levels of exposure to PFOA and PFOS in these populations compared to the
general population.
Applicability: The EPA evaluated each study to determine whether it estimated the association
between exposure to PFOA or PFOS (measured in serum or plasma) and a quantitative measure
of TC or HDLC in general populations (age 20 and older). Of the 39 studies identified as part of
the ATSDR-based literature review that provided information on the relationship between
exposure to PFAS and TC and HDLC levels, 9 were general population studies. Of the 41 studies
identified as part of the EPA/OST literature review that provided information on the relationship
between exposure to PFAS and TC and HDLC levels, 14 were general population studies.
These studies31 were further evaluated for inclusion in the meta-analysis.
Research methods and study details: The EPA evaluated each study to determine whether it
reported numbers of participants, quantitative effect estimates (beta coefficients), measures of
effect estimate variance (95% confidence intervals [CIs], standard errors [SEs], or standard
deviations [SDs]). The EPA retained studies with missing measures of effect estimate variance
but with reported p-values for differences. For such studies, the EPA used the approach in the
Cochrane Handbook for Systematic Reviews (Higgins et al., 2019) to calculate SDs or SEs.
Briefly, the approach estimates the SEs using the correspondence between the p-value and the t-
statistic, with degrees of freedom equal to the difference between the sample size and the number
of parameters in the model that provided the effect estimate. Then the SE is obtained by dividing
the effect estimate by the t-statistic.
Additional exclusion criteria: The EPA also excluded studies that reported data only for pregnant
women, infants, or children. Although there is some evidence that PFAS exposure is associated
with cardiometabolic impairment in children and younger adults (Rappazzo et al., 2017), the
EPA did not extract data from these studies because lipid levels are known to change during
pregnancy from pre-pregnancy levels, and the relationships between lipid profiles at early life
stages are not as well defined as they are at later life stages. Another frequent reason for study
exclusion was the reporting of only relative risks or odds ratios for hypercholesteremia or
hyperlipidemia; results in this form could not be used to estimate continuous exposure-response
relationships.
F.1.2 Assessment of Study Applicability to the Meta-Analysis
Figure F-l presents a flow diagram of the studies reviewed as part of the ATSDR-based and the
EPA/OST-based literature reviews and the selection of studies retained for inclusion in the
meta-analysis. Using the study inclusion criteria described in Section F.l.l, the EPA retained
14 studies for use in the meta-analysis. Of these, five were identified as part of the ATSDR
literature review (Chateau-Degat et al., 2010; Fisher et al., 2013; Fu et al., 2014; Nelson et al.,
2010; Steenland et al., 2009), seven were identified from the EPA systematic review (Dong et
al., 2019; Fan et al., 2020; Jain & Ducatman, 2019; Y. Li et al., 2020; C. Y. Lin et al., 2020; P.-I.
31 Of the general population studies identified as part of the EPA/OST literature review, five overlapped with studies identified as
part of the ATSDR-based literature review.
Final PFAS Rule Economic Analysis
F-2
April 2024
-------
FINAL RULE
APRIL 2024
D. Lin et al., 2019; Yang et al., 2018), and two were identified in both literature reviews (He et
aL 2018; Liu et al., 2018).
Legend:
Literature
Review Basis
Retained Studies
Excluded Studies j
Final Retained
Studies
N otes:
ATSDR = Agency forToxic Substances and Disease Registry, EPA = Environmental Protection Agency, OST = Office of Science and
Technology, PFAS = per- and polyfiuoroalkyl substances, TC = Total Cholesterol, HDLC = high-density lipoprotein cholesterol
""Included literature through mid-2017,
hncluded literature published from 2016 to 2020.
"For example, studies based on occupational data or data only for pregnant women, infants, or children.
Some studies did not include the estimates required for meta-analysis calculations. For example, certain studies did not report effect
estimates or interquartile ranges.
"Of these studies. Sane based on datafrom the United States and 6are based on data outside of the United States.
Figure F-l: Diagram of Literature Retained for Use in the Meta-Analysis and Data
Sources.
Table F-l summarizes the 14 studies that were identified in the ATSDR-based and the EPA
literature review that the EPA used to derive slope estimates for PFOA and PFOS associations
with serum TC and HDLC levels.52 Six of the studies that the EPA retained for use in the meta-
analysis were based on PFAS and serum lipid measurements from the U.S. general population
(National Health and Nutrition Examination Survey [NHANES]) (Dong et al., 2019; Fan et al.,
2020; He et al., 2018; Jain & Ducatman, 2019; Liu et al., 2018; Nelson et al., 2010); there were
also general population studies from Canada (Fisher et al., 2013), Sweden (Y. Li et al., 2020),
32 For this effort, the EPA focused on PFOA and PFOS, since these are by far the most well-studied perfluorinated compounds.
Final PFAS Rule Economic Analysis
F-3
April 2024
-------
FINAL RULE
APRIL 2024
Taiwan (Yang et al., 2018; C. Y. Lin et al., 2020), and Henan Province, China (Fu et al., 2014).
Chateau-Degat et al. (2010) reported on the relationship between PFOS and serum lipids in a
Canadian Inuit population. The EPA also retained the results from a study of a highly exposed
population in the United States (the C8 Health Project cohort) (Steenland et al., 2009) and from a
study using participants in a U.S. diabetes prevention program (P.-I. D. Lin et al., 2019).
The EPA excluded two general population studies identified in the ATSDR-based literature
review (Eriksen et al., 2013; Seo et al., 2018) and two general population studies identified based
on the agency's systematic review (Convertino et al., 2018; Huang et al., 2018) that were
inadequate for use in the meta-analysis because they did not include the estimates required for
meta-analysis calculations. For example, the EPA excluded the studies identified in the ATSDR
literature review from the meta-analysis because the authors did not report either the effect
estimates (Seo et al., 2018) or interquartile ranges (Eriksen et al., 2013) needed for
calculations.33 Similarly, the EPA excluded the studies identified as part of the agency's
systematic review because they involved a Phase 1 controlled trial with modeled exposures in
cancer patients dosed with ammonium perfluorooctanoate (Convertino et al., 2018) or reported
effect estimates (Spearman correlation coefficients) that were not suitable for use in the meta-
analysis (Huang et al., 2018). The EPA also considered the longitudinal study by Fitz-Simon et
al. (2013) of adults participating in the C8 Health Project who were not taking cholesterol-
lowering medication and who were examined twice, with an average of 4.4 years between
examinations. In subjects whose serum PFOA levels halved between examinations, there was a
decrease of an average of 1.65% (95% confidence interval: 0.32%, 2.97%) for TC and 1.33%
(-0.21%), 2.85%) for HDLC. In subjects whose serum PFOS levels halved between
examinations, there were similar decreases, although larger in magnitude and variability: a
decrease of an average of 3.20%> (95% confidence interval: 1.63%, 4.76%) for TC and 1.28%
(-0.59%), 3.12%)) for HDLC. However, given the nature of the results, the effect estimates from
this study were inadequate for inclusion in the meta-analysis.
33 Efforts to contact the study authors for the missing data were unsuccessful at the time of this report.
Final PFAS Rule Economic Analysis
F-4
April 2024
-------
FINAL RULE
APRIL 2024
Table F-l: Studies Selected for Inclusion in the Meta-Analyses
Author and Year
Title
Cholesterol and PFAS Relationship
Evaluated
TC HDLC
PFOA PFOS PFOA PFOS
Medications
Association of Perfluorooctanoic Acid and
Participants using lipid-lowering
Steenland et al., 2009'"1 Perfluorooctane Sulfonate With Serum Lipids Among X
X
X
X
medications were excluded
Adults Living Near a Chemical Plant
Chateau-Degat et al..
2010a'd
Nelson et al.. 2010a
Effects of Perfluorooctanesulfonate Exposure
on Plasma Lipid Levels in the Inuit Population
of Nunavik (Northern Quebec)
Exposure to Polyfluoroalkyl Chemicals and
Cholesterol, Body Weight, and Insulin Resistance in the X
General U.S. Population
X
X
X
X
X
Use of lipid-lowering
medication considered in
statistical analysis
Participants using lipid-lowering
medications were excluded
Fisher et al., 2013a
Fu et al., 2014a
He et al., 2018°
Liu et al., 2018°
Do Perfluoroalkyl Substances Affect Metabolic
Function and Plasma Lipids?—Analysis of the 2007-
2009, Canadian Health Measures Survey (CHMS)
Cycle 1
Associations Between Serum Concentrations of
Perfluoroalkyl Acids and Serum Lipid Levels in a
Chinese Population
PFOA is Associated with Diabetes and Metabolic
Alteration in US Men: National Health and Nutrition
Examination Survey 2003-2012
Association Among Total Serum Isomers of
Perfluorinated Chemicals, Glucose Homeostasis, Lipid
Profiles, Serum Protein and Metabolic Syndrome in
Adults: NHANES, 2013-2014
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Participants using lipid-lowering
medications were excluded
Not taken into consideration
Not taken into consideration
Use of lipid-lowering
medication considered in
statistical analysis
Final PFAS Rule Economic Analysis
F-5
April 2024
-------
FINAL RULE
APRIL 2024
Table F-l: Studies Selected for Inclusion in the Meta-Analyses
Author and Year
Title
Cholesterol and PFAS Relationship
Evaluated
TC HDLC Medications
PFOA PFOS PFOA PFOS
Yang et al., 2018b
Association of Serum Levels of Perfluoroalkyl
Substances (PFASs) With the Metabolic Syndrome
(MetS) in Chinese Male Adults: A Cross-Sectional
Study
Not taken into consideration
X X
Dong et al., 2019b
Using 2003-2014 U.S. NHANES Data to Determine
the Associations Between Per- and Polyfluoroalkyl
Substances and Cholesterol: Trend and Implications
Roles of Gender and Obesity in Defining Correlations
Jain & Ducatman, 2019b Between Perfluoroalkyl Substances and
Lipid/Lipoproteins
X
X
X
X
X
X
X
Participants using lipid-lowering
medications were excluded
Use of lipid-lowering
medication considered in
statistical analysis
Per- and Polyfluoroalkyl Substances and Blood Lipid
P.-I. D. Lin et al.. 2019b Levels in Pre-Diabetic Adults—Longitudinal Analysis
of the Diabetes Prevention Program Outcomes Study
Fanet al., 2020b
X
Serum Albumin Mediates the Effect of Multiple Per-
and Polyfluoroalkyl Substances on Serum Lipid Levels
X
X
X
X X X X
Participants using lipid-lowering
medications were excluded
Not taken into consideration
Y. Li et al.. 2020b
Associations Between Perfluoroalkyl Substances and
Serum Lipids in a Swedish Adult Population With
Contaminated Drinking Water
X X X X
Not taken into consideration
C. Y. Lin et al., 2020b
The Association Between Total Serum Isomers of Per-
and Polyfluoroalkyl Substances, Lipid Profiles, and the
DNA Oxidative/Nitrative Stress Biomarkers in Middle-
Aged Taiwanese Adults
X
X
Not taken into consideration
Final PFAS Rule Economic Analysis
F-6
April 2024
-------
FINAL RULE
APRIL 2024
Table F-l: Studies Selected for Inclusion in the Meta-Analyses
Cholesterol and PFAS Relationship
Evaluated
Author and Year
Title
TC
HDLC
Medications
PFOA PFOS PFOA PFOS
Abbreviations: PFAS - per-and polyfluoroalkyl substances; PFOA - perfluorooctanoic acid; PFOS - perfluorooctanesulfonic acid; TC - total cholesterol; HDLC - high-density
lipoprotein cholesterol.
Notes: Study quality reflected in green (medium confidence) or pink (low confidence) cell shading.
aStudies identified based on ATSDR literature review.
bStudies identified based on the EPA literature review.
cStudies available in both assessments.
dStudies available in PFOA and/or PFOS health effects support documents (U.S. EPA, 2016a, 2016b).
Final PFAS Rule Economic Analysis
F-7
April 2024
-------
FINAL RULE
APRIL 2024
F.2 Meta-Analysis
Based on the study inclusion criteria discussed in Section F.l.l, the EPA included 14 studies in
the meta-analysis. Of these 14 studies, 11 were used to develop exposure-response relationships
for serum PFOA and TC, 13 were used to develop exposure-response relationships for serum
PFOA and HDLC, 12 studies were used to develop exposure-response relationships for serum
PFOS and TC, and 13 studies were used to develop exposure-response relationships for serum
PFOS and HDLC (Table F-l). The EPA conducted four separate meta-analyses: one analysis for
each combination of chemical (PFOA or PFOS) and health outcome (TC or HDLC).
All studies were evaluated for risk of bias, selective reporting, and sensitivity as applied in
developing the EPA's Final Human Health Toxicity Assessments for PFOA and PFOS (U. S.
EPA, 2024b; U.S. EPA, 2024c). Briefly, the main considerations specific to evaluating the
quality of studies on serum lipids included use of medications, fasting, and potential for reverse
causality. Because lipid-lowering medications strongly affect serum lipid levels, studies that did
not account for the use of lipid-lowering medications by restriction, stratification, or adjustment
were rated as deficient in the participant selection domain. For TC and HDLC measurements,
fasting is not likely to introduce measurement error because the serum levels of the lipids
considered change minimally after a meal (Mora, 2016). Measuring PFOS and serum lipids
concurrently was considered adequate in terms of exposure assessment timing. Given the long
half-life of PFOA and PFOS (Ying Li et al., 2018), current blood concentrations are expected to
correlate well with past exposures. Furthermore, although reverse causation due to
hypothyroidism (Dzierlenga, Allen, et al., 2020) or enterohepatic cycling of bile acids (Fragki et
al., 2021) has been suggested, there is not yet clear evidence to support these reverse causal
pathways.
Based on these considerations, of the 14 studies, ten were medium confidence in ROB
evaluations, with only four deemed low confidence (Fu et al., 2014; He et al., 2018; Yang et al.,
2018; Y. Li et al., 2020). These low confidence studies had deficiencies in participant selection,
outcome assessment, or confounding domains. None of these studies considered use of lipid-
lowering medications in the selection process or in the statistical analyses. Additional details on
the ROB evaluations are available in ICF (2021).
F.3 Extraction of Slope Values for TC and HDLC
If studies reported linear slope relationships (change in serum TC or HDLC in mg/dL per ng/mL
change in serum PFOA/PFOS), the EPA extracted these values, along with their confidence
limits, directly as reported by the study authors. If results from multiple models with different
adjustments for confounders were reported within a single study, either the most adjusted results
or the main model results as presented by the study authors were selected. When studies
provided results for both untransformed and log-transformed PFOA/PFOS, the EPA used
untransformed PFOA/PFOS to reduce bias due to back-transformations of effect estimates. For
studies that provided results only for log-transformed PFOA/PFOS (five studies) or log-
transformed outcomes (two studies), or log-transformed both PFOA/PFOS and outcomes (two
studies), the EPA approximated the results for an untransformed analysis using the approach
outlined by Rodriguez-Barranco et al. (2017) and Dzierlenga, Crawford, and Longnecker (2020).
When not reported, the EPA assumed that the natural logarithm was the basis of the
Final PFAS Rule Economic Analysis
F-8
April 2024
-------
FINAL RULE
APRIL 2024
transformation. An independent reviewer of the EPA evaluated the extracted slope values for
quality assurance.
F.4 Methods and Key Assumptions
The summary measure of association was a beta coefficient relating changes in TC or HDLC in
mg/dL to increases in serum or plasma34 PFOA or PFOS in ng/mL. The EPA conducted random-
effects meta-analyses using the DerSimonian and Laird (1986) approach, which uses weights
based on the inverse of the variance of the coefficient of each study plus the addition of an extra
component of variance between studies. When studies reported beta coefficients by quartiles
(e.g., He et al., 2018), the EPA estimated a linear coefficient using a weighted linear regression
of the midpoints of the quartiles and the reported beta coefficients, using the inverse of standard
errors as the regression weights.
The EPA assessed between-study heterogeneity using Cochran's Q test (Cochran, 1954) and the
I2 statistic (Higgins et al., 2003). The EPA developed forest plots to display the results. The EPA
developed funnel plots and performed an Egger regression on the estimates of effect size to
assess potential publication bias (Begg & Mazumdar, 1994; Egger et al., 1997; Egger et al.,
2008). Because back-transformations of effect estimates with log-transformed outcomes or
exposures could introduce bias and could be a source of heterogeneity, the EPA also conducted
sub-analyses by type of model that provided the study-specific effect estimate (e.g., only
including studies that reported linear associations [six studies] or linear-log associations [five
studies]).
If publication bias was observed, the EPA conduced sensitivity analyses using trim-and-fill
methods (Duval & Tweedie, 2000a, 2000b) to estimate the number of missing studies and predict
the impact of the hypothetical "missing" studies on the pooled effect estimate. To investigate
sources of heterogeneity, the EPA conducted several sensitivity analyses:
The EPA evaluated the impact of using other estimation methods for the between-study variance
(tau2) besides the DerSimonian and Laird (1986) approach, such as restricted maximum
likelihood (Raudenbush, 2009) or Sidik and Jonkman (2005).
• To assess potential impact of a single study on the overall effect estimate, the EPA conducted
leave-one-out meta-analyses.
• To assess potential impact of study quality on the overall effect estimate, the EPA conducted
sensitivity analyses excluding the four studies considered to have higher ROB.
• To assess the impact of using multiple regression coefficients from the same study (which are
correlated), the EPA excluded a study that contributed four effect estimates (gender- and
obesity-specific) for each analysis, which also accounted for most of the weight in the overall
pooled beta coefficient (Jain & Ducatman, 2019). The EPA also conducted a sensitivity
analysis using a single pooled estimate from the four study-specific estimates.
• The EPA also assessed the impact of non-U. S or Canadian general population studies in
sensitivity analyses excluding studies conducted in China (Fu et al., 2014), Taiwan (Yang et
al., 2018; C. Y. Lin et al., 2020), or Sweden (Y. Li et al., 2020), the Canadian Inuit
34 PFOA or PFOS concentrations is serum or plasma were treated interchangeably.
Final PFAS Rule Economic Analysis
F-9
April 2024
-------
FINAL RULE
APRIL 2024
population study (Chateau-Degat et al., 2010), and the U.S. high-exposure community study
(Steenland et al., 2009).
Six studies that the EPA retained for use in the meta-analysis were based on PFAS and serum
lipid measurements using data from overlapping NHANES cycles: Dong et al. (2019) used data
from 2003-2014, while He et al. (2018) used 2003-2012 data; Jain and Ducatman (2019) used
2005-2014 data; Fan et al. (2020) used 2011-2014 data; Liu et al. (2018) used 2013-2014; and
Nelson et al. (2010) used data from 2003-2004. Although the datasets and models were not
exactly the same in all NHANES-based studies, to avoid estimate dependency issues due to
overlapping populations in the meta-analysis, the EPA also performed a sensitivity analysis
including only the data from the study covering the broadest range of NHANES cycles (2003-
2014) (Dong etal., 2019).
The EPA performed statistical analyses using the software STATA, version 16.1 (StataCorp,
2019), with the combine, meta esize, meta set, meta summarize, metainf, meta funnel, meta bias,
and meta trimfill packages (Palmer & Sterne, 2016). Results of the meta-analyses are presented
in Table F-2 and Table F-3. Overall, there is a high degree of heterogeneity when all studies are
combined. Excluding Jain and Ducatman (2019) did not significantly reduce the heterogeneity;
however restricting analyses to studies reporting linear or linear-log associations did reduce
heterogeneity in most cases.
F.4.1 Slope Estimation for PFOA
When including the six studies reporting linear associations, there was a statistically significant
positive increase in TC of 1.57 (95% confidence interval: 0.02, 3.13) mg/dL per ng/mL serum
PFOA (p-value = 0.048,12 = 87%). The association for HDLC and PFOA was positive (0.11;
95% CI: -0.22, 0.43) but not statistically significant (Table F-2, Figure F-2). Adjusting for
possible publication bias through funnel plots and trim-and-fill analysis suggested the imputation
of two additional studies for HDLC and PFOA with a smaller effect (-0.01, 95% confidence
interval: -0.42, 0.41). For TC and PFOA, the pooled associations did not change when adjusting
for possible publication bias (Figure F-3). However, methods to assess heterogeneity and
publication bias have limitations in small sample-size meta-analyses, thus these results should be
interpreted cautiously (von Hippel, 2015).
Final PFAS Rule Economic Analysis
F-10
April 2024
-------
FINAL RULE APRIL 2024
Table F-2: Results for PFOA Meta-Analyses
Group
Outcome
Number of
Studies/
Number of
Estimates
Beta
(mg/dL per ng/mL)
95% CIs
p-value
Qa
p-value
for Q
I2
Tau2
TC
11/14
0.003
-0.001
0.006
0.177
123.68
<0.001
89.49
0
All Studies
HDLC
13/17
0.001
-0.001
0.004
0.291
54.74
<0.001
70.77
0
TC
4
1.574
0.018
3.130
0.048
23.43
< 0.001
87.19
1.910
Linear Models Only
HDLC
5
0.105
-0.219
0.428
0.526
14.01
0.007
71.45
0.069
Sensitivity Analyses
All lower risk of bias
TC
8/11
0.003
-0.003
0.008
0.321
88.86
<0.001
88.75
0
studies
HDLC
9/13
0.002
-0.002
0.005
0.290
28.34
0.005
57.65
0
Exclude Jain and Ducatman
TC
10
0.004
-0.002
0.010
0.179
82.04
<0.001
89.03
0
(2019)
HDLC
12/13
0.001
-0.003
0.006
0.500
50.18
<0.001
76.09
0
Exclude non-US/Canada
TC
8/11
0.002
-0.003
0.006
0.496
55.65
<0.001
82.03
0
and high exposure studies
HDLC
8/11
0.001
-0.003
0.005
0.647
26.17
0.004
61.79
0
All studies, pooled Jain and
TC
11
0.003
-0.002
0.008
0.183
91.42
<0.001
89.06
0
Ducatman (2019)
HDLC
13/14
0.001
-0.002
0.004
0.412
53.07
<0.001
75.51
0
All studies, no NHANES
TC
6
0.017
-0.033
0.067
0.505
21.56
0.001
76.9
0.001
overlap
HDLC
8/9
0.0030
0.0029
0.0031
<0.001
4.12
0.844
0
0
Linear models only, no
TC
lb
1.480
0.180
2.780
0.026
0.00
NA
NA
NA
NHANES overlap
HDLC
2
0.185
-0.897
1.249
0.773
1.29
0.26
22.61
0.29
TC
3/6
0.002
-0.004
0.007
0.594
31.56
<0.001
84.16
0
Linear-log models only
HDLC
5/9
0.001
-0.003
0.006
0.490
13.56
0.094
41.01
0
TC
1
1.632
-0.841
2.422
>0.05
0.00
NA
NA
NA
P.-I. D. Lin et al. (2019)
HDLC
1
-0.131
-0.370
0.107
>0.05
0.00
NA
NA
NA
TC
1
1.480
0.180
2.780
0.026
0.00
NA
NA
NA
Dong et al. (2019)
NA
NA
NA
HDLC
1
-0.025
-0.443
0.393
>0.05
0.00
Abbreviations: CI - confidence interval; HDLC - high-density lipoprotein cholesterol; TC - total cholesterol; PFOA- Perfluorooctanoic Acid.
Notes:
aQ statistics for heterogeneity. Tau2 is the between-studies variance. I2 represents the proportion of total variance in the estimated model due to inter-study variation.
bData from Dong et al. (2019) Statistics for heterogeneity do not apply when only one study is used.
Final PFAS Rule Economic Analysis
F-ll
April 2024
-------
FINAL RULE
APRIL 2024
Study
TC
Effect estimate
95% CI
Weight
<%)
Nelson et al. (2010)
He et al. (2018)
Dong et al. (2019)
Fan et al. (2020)
Overall
Study
1.2200 [ 0.0400, 2.4000] 27.74
0.0016 [ 0.0001, 0.0030] 33.01
1.4800[ 0.1800, 2.7800] 26.83
6.7400[ 3.2551, 10.2249] 12.43
1.5737 [ 0.0177, 3.1297]
0 ~~5 10
Beta (95% confidence interval)
HDLC
Effect estimate
Weight
95% CI
<%)
-0.1200 [
-0.4050,
0.1650]
25.55
1.5855 [
-1.1600,
4.3310]
1.40
-0.0007 [
-0.0022,
0.0008]
31.46
-0.3000 [
-0.7100,
0.1100]
21.27
0.6500 [
0.0150,
1.2850]
14.63
2.2300 [
0.9700,
3.4900]
5.69
0.1495 [
-0.1826,
0.4817]
Nelson et al. (2010)
Fu et al. (2014)
Heet al. (2018)
Dong et al. (2019)
Dong et al. (2019)
Fan et al. (2020)
Overall
-2 0 2 4
Beta (95% confidence interval)
Figure F-2: Forest Plots Showing the Beta Coefficients
Relating PFOA Concentrations to TC and HDLC in Each Study Reporting
Linear Associations, and Pooled Estimates After Random-Effects Meta-Analysis.
Final PFAS Rule Economic Analysis
F-12
April 2024
-------
FINAL RULE
APRIL 2024
-------
FINAL RULE
APRIL 2024
Study
Steenland (2009)
Nelson (2010)
Fisher (2013)
Fu (2014)
He (2018)
Liu (2018)
Dong (2019)
Jain (2019) f no
Jain (2019) fo
Jain (2019) m no
Jain (2019) m o
Lin (2019)
Fan (2020)
Li (2020)
Overall
(a) TC and PFOA Serum Relationship
o.o
2.5 5.0 7.5
Beta (95% confidence interval)
10.0
Study
Steenland (2009)
Nelson (2010)
Fisher (2013)
Fu (2014)
He (2018)
Liu (2018)
Yang (2018)
Dong (2019)
Jain (2019) f no
Jain (2019) fo
Jain (2019) m no
Jain (2019) m o
Lin (2019)
Fan (2020)
Li (2020)
Lin (2020) br PFOA
Lin (2020) lin PFOA
Overall
-2.5 0.0 2.5 5.0
Beta (95% confidence interval)
Effect Estimate
[95% CI]
Weight
0.0060
[0.0059,
0.0060]
18.5895
1.2200
[0.0400,
2.4000]
0.0009
0.0504
[-0.2375,
0.3383]
0.0157
0.2447
[-0.5493,
1.0386]
0.0021
0.0016
[0.0001,
0.0030]
18.0390
2.4008
[0.6889,
4.1126]
0.0004
1.4800
[0.1800,
2.7800]
0.0008
0.0019
[-0.0022,
0.0059]
15.0655
-0.0004
[-0.0047,
0.0039]
14.7342
-0.0011
[-0.0036,
0.0015]
17.0060
0.0058
[0.0014,
0.0102]
14.5987
1.6317
[0.8413,
2.4221]
0.0021
6.7400
[3.2551,
10.2249]
0.0001
0.0064
[-0.0181,
0.0309]
1.9450
0.0025
[-0.0011,
0.0061]
Serum Relationship
Effect Estimate
[95% Cll
Weiqht
0.0030
[0.0029,
0.0031]
22.7126
-0.1200
[-0.4050,
0.1650]
0.0082
0.0004
[-0.0708,
0.0716]
0.1309
1.5855
[-1.1600,
4.3310]
0.0001
-0.0007
[-0.0022,
0.0008]
21.1506
0.8304
[0.2907,
1.3701]
0.0023
2.2240
[-2.4463,
6.8943]
0.0000
-0.0253
[-0.4438,
0.3933]
0.0038
0.0027
[-0.0021,
0.0075]
12.7084
0.0029
[-0.0045,
0.0104]
7.8535
-0.0006
[-0.0041,
0.0028]
16.0751
0.0016
[-0.0031,
0.0062]
13.0741
-0.1313
[-0.3697,
0.1072]
0.0117
2.2300
[0.9700,
3.4900]
0.0004
0.0027
[-0.0061,
0.0115]
6.2645
-0.5205
[-3.8363,
2.7952]
0.0001
0.1765
[-0.2432,
0.5961]
0.0038
0.0014
[-0.0012,
0.0040]
Figure F-4: Forest Plots Showing the Beta Coefficients Relating TC and HDLC to PFOA
Concentrations in Each Study, and Pooled Estimates After Random-Effects Meta-Analvsis.
Abbreviations: f - females; m - males; o - obese; no - non-obese
Final PFAS Rule Economic Analysis
F~14
April 2024
-------
FINAL RULE
APRIL 2024
-------
FINAL RULE
APRIL 2024
F.4.2 Slope Estimation for PFOS
When including the five studies reporting linear associations, there was a positive increase in TC
of 0.08 (95% CI: -0.01, 0.16) mg/dL per ng/mL serum PFOS (p-value = 0.064,I2 = 84%) that
was significant at the 0.10 level. The association for PFOS and HDLC was positive but not
statistically significant (Table F-3, Figure F-6). Adjusting for possible publication bias through
funnel plots and trim-and-fill analysis suggested the imputation of additional studies; however,
the magnitude or significance of the pooled associations did not change significantly (Figure
F-7).
When all studies were combined (12 studies, 15 results), the EPA observed a borderline
statistically significant positive increase in TC of 0.066 (95% CI: -0.001, 0.132) mg/dL per
ng/mL serum PFOS (p-value = 0.055,12 = 100%) (Table F-3, Figure F-8). Adjusting for possible
publication bias through funnel plots and trim-and-fill analysis suggested the imputation of three
additional studies for TC and five for HDLC; however, the pooled effect estimates did not
change significantly (Figure F-9). The EPA observed similar results in leave-one-out analyses,
sensitivity analyses restricted to U.S. or Canadian general population studies, and analyses
excluding Jain and Ducatman (2019), estimates. Similar results were observed when the analysis
excluded the overlapping NHANES studies. When the analysis excluded the higher ROB studies,
the association was significantly positive with an increase in in TC of 0.09 (95% CI: 0.01, 0.17)
mg/dL per ng/mL serum PFOS (p-value = 0.047).
The pooled estimate based on the studies reporting linear associations was 0.08 (95% CI: -0.01,
0.16) and significant at the 0.10 level (p-value = 0.064) and there is evidence supporting a
positive and significant relationship between PFOS and TC: the EPA/OST's review of 41 recent
epidemiological studies showed positive associations between PFOS and TC in the general
population and the meta-analysis performed with all studies combined showed a positive
increase in TC per ng/mL serum PFOS that was significant at the 0.10 level. Given this weight of
evidence, the large degree of heterogeneity in the pooled associations when all data were
included, and the likelihood of bias that back-transformation of effect estimates with log-
transformed outcomes or exposures could introduce (and difficulty with estimating the
directionality of this bias towards or away from the null), the EPA relied on the results from
analyses restricted to studies reporting similar models, favoring the pooled slope (from the six
studies reporting linear associations) of 0.08 mg/dL TC and 0.05 mg/dL HDLC per ng/mL serum
PFOS for interpretability and use in the CVD risk reduction analysis.35
35 The EPA characterizes uncertainty surrounding this estimate as described in Appendix L.
Final PFAS Rule Economic Analysis
F-16
April 2024
-------
FINAL RULE APRIL 2024
Table F-3: Results for PFOS Meta-Analyses
Group
Outcome
N Studies/
Number of
Estimates
Beta
(mg/dL per ng/mL)
95% CIs
p-value
Qa
p-value
for Q
I2
Tau2
All Studies
TC
12/15
0.066
-0.001
0.132
0.055
630000
<0.001
100
0.012
HDLC
14/19
0.0003
-0.001
0.001
0.631
158.85
<0.001
88.67
0
Linear Models Only
TC
HDLC
5
6/7
0.079
0.050
-0.005
-0.005
0.162
0.105
0.064
0.074
25.84
31.69
< 0.001
< 0.001
84.52
81.06
0.004
0.003
Sensitivity Analyses
All lower risk of bias
TC
9/12
0.086
0.001
0.170
0.047
450000
<0.001
100
0.016
studies
HDLC
10/15
0.001
-0.001
0.002
0.606
84.54
<0.001
83.44
0
Exclude Jain and Ducatman
TC
11
0.114
0.012
0.217
0.028
510000
<0.001
100
0.019
(2019)
HDLC
13/15
-0.002
-0.002
0.001
0.778
126.90
<0.001
88.97
0
Exclude non-US/Canada
TC
8/11
0.001
-0.0004
0.001
0.301
34.71
<0.001
71.20
0
and high exposure studies
HDLC
8/11
0.001
-0.0002
0.001
0.165
13.12
<0.001
23.76
0
All studies, pooled Jain and
TC
12
0.094
0.010
0.179
0.029
590000
<0.001
100
0.015
Ducatman (2019)
HDLC
14/16
-0.0001
-0.0014
0.0013
0.943
157.53
<0.001
90.48
0
All studies, no NHANES
TC
7
0.109
-0.016
0.234
0.088
120000
<0.001
100
0.022
overlap
HDLC
9/11
-0.001
-0.002
0.002
0.642
94.82
<0.001
89.45
0
Linear models only, no
TC
2b
0.192
-0.162
0.546
0.288
6.88
0.009
85.46
0.057
NHANES overlap
HDLC
3/4
0.078
0.001
0.155
0.048
7.32
0.062
59.03
0.003
Linear-log models only
TC
HDLC
3/6
5/9
0.0003
0.001
-0.0003
-0.001
0.001
0.002
0.342
0.270
8.33
15.74
0.139
0.046
39.99
49.18
0
0
P.-I. D. Lin et al. (2019)
TC
HDLC
1
1
0.132
-0.021
-0.005
-0.062
0.269
0.020
>0.05
>0.05
0.00
0.00
NA
NA
NA
NA
NA
NA
Dong et al. (2019)
TC
HDLC
1
1
0.40
0.014
0.13
-0.084
0.67
0.110
<0.01
>0.05
0.00
0.00
NA
NA
NA
NA
NA
NA
Abbreviations: CI - confidence interval; HDLC- high-density lipoprotein cholesterol; TC- total cholesterol; NHANES - National Health and Nutrition Examination; PFOS-
perfluorooctanesulfonic acid.
Notes:
aQ statistics for heterogeneity. Tau2 is the between-studies variance. I2 represents the proportion of total variance in the estimated model due to inter-study variation.
bData from Dong et al. (2019) and Chateau-Degat et al. (2010).
Final PFAS Rule Economic Analysis
F-17
April 2024
-------
FINAL RULE
APRIL 2024
Study
TC
Effect estimate
95% CI
Weight
(%)
Chateau-Degat (2010)
•
0.0348 [
-0.0049,
0.0745]
38.68
Nelson (2010)
0.2700 [
0.0550,
0.4850]
11.05
He (2018)
¦
0.0008 [
0.0003,
0.0012]
42.42
Dong (2019)
—
0.4000 [
0.1300,
0.6700]
7.74
Fan (2020)
3.8500 [
1.2750,
6.4250]
0.10
Overall
0.0786 [
-0.0045,
0.1617]
Study
0 2 4
Beta (95% confidence interval)
HDLC
Chateau-Degat (2010) f
Chateau-Degat (2010) m
Nelson (2010)
Fu (2014)
He (2018)
Dong (2019)
Fan (2020)
Overall
0.1624 [
0.0664,
0.2584]
14.78
0.0619 [
0.0254,
0.0984]
24.24
0.0200 [
-0.0500,
0.0900]
18.80
2.5909 [
-0.6767,
5.8584]
0.03
0.0002 [
-0.0004,
0.0008]
27.19
0.0135 [
-0.0836,
0.1107]
14.62
1.2400 [
0.3200,
2.1600]
0.35
0.0498 [
-0.0048,
0.1045]
0 2 4 6
Beta (95% confidence interval)
Figure F-6: Forest Plots Showing the Beta Coefficients Relating TC
and HDLC to PFOS Concentrations in Each Study Reporting Linear Associations,
and Pooled Estimates After Random-Effects Meta-Analysis.
Abbreviations: f - females; m - males; o - obese; no - non-obese
Final PFAS Rule Economic Analysis
F-18
April 2024
-------
FINAL RULE
APRIL 2024
c
-------
FINAL RULE
APRIL 2024
Study
Steenland (2009)
Chateau-Degat (2010)
Nelson (2010)
Fisher (2013)
Fu (2014)
He (2018)
Liu (2018)
Dong (2019)
Jain (2019) f no
Jain (2019) f o
Jain (2019) m no
Jain (2019) m o
Lin (2019)
Fan (2020)
Li (2020)
Overall
Study
Steenland (2009)
Chateau-Degat (2010) f
Chateau-Degat (2010) m
Nelson (2010)
Fisher (2013)
Fu (2014)
He (2018)
Liu (2018)
Yang (2018)
Dong (2019)
Jain (2019) f no
Jain (2019) fo
Jain (2019) m no
Jain (2019) m o
Lin (2019)
Fan (2020)
Li (2020)
Lin (2020) br PFOS
Lin (2020) lin PFOS
Overall
(a) TC and PFOS Serum Relationship
0 2 4 6
Beta (95% confidence interval)
Effect Estimate
[95% CI]
Weight
0.2091
[0.2086,
0.2095]
9.3960
0.0348
[-0.0049,
0.0745]
9.0953
0.2700
[0.0550,
0.4850]
4.7663
0.0077
[-0.0669,
0.0824]
8.4110
0.1252
[-0.8085,
1.0590]
0.4863
0.0008
[0.0003,
0.0012]
9.3960
0.2116
[-0.4377,
0.8609]
0.9530
0.4000
[0.1300,
0.6700]
3.7111
0.0004
[-0.0006,
0.0014]
9.3959
0.0009
[-0.0003,
0.0021]
9.3958
-0.0003
[-0.0009,
0.0004]
9.3960
0.0006
[-0.0004,
0.0016]
9.3959
0.1318
[-0.0052,
0.2688]
6.7386
3.8500
[1.2750,
6.4250]
0.0670
0.0003
[-0.0008,
0.0014]
9.3958
0.0655
[-0.0014,
0.1324]
(b) HDLC and PFOS Serum Relationship
-2 0 2 4
Beta (95% confidence interval)
Effect Estimate
[95% CI]
Weight
-0.0015
[-0.0016,
-0.0014]
16.4599
0.1624
[0.0664,
0.2584]
0.0122
0.0619
[0.0254,
0.0984]
0.0844
0.0200
[-0.0500,
0.0900]
0.0230
-0.0030
[-0.0293,
0.0233]
0.1613
2.5909
[-0.6767,
5.8584]
0.0000
0.0002
[-0.0004,
0.0008]
15.6444
0.1578
[-0.0801,
0.3958]
0.0020
0.1855
[-1.5301,
1.9010]
0.0000
0.0135
[-0.0836,
0.1107]
0.0120
0.0010
[-0.0004,
0.0025]
12.5475
0.0015
[-0.0005,
0.0034]
10.5739
0.0000
[-0.0008,
0.0008]
14.9834
0.0007
[-0.0006,
0.0019]
13.2682
-0.0208
[-0.0620,
0.0203]
0.0664
1.2400
[0.3200,
2.1600]
0.0001
0.0001
[-0.0003,
0.0005]
16.1448
-2.0986
[-3.9343,
-0.2628]
0.0000
0.0977
[0.0144,
0.1809]
0.0163
0.0003
[-0.0008,
0.0013]
Figure F-8: Forest Plots Showing the Beta Coefficients Relating PFOS Concentrations to
TC and HDLC in Each Study, and Pooled Estimates After Random-Effects Meta-Analysis.
Abbreviations: f - females; m - males; o - obese; no - non-obese.
Final PFAS Rule Economic Analysis
F-20
April 2024
-------
FINAL RULE
APRIL 2024
CD
"O
CT3
"O
C
03
CO '
Funnel plot
V
TC
1
0
Effect size
Pseudo 95% CI • Observed studies
Estimated 0Dl • Imputed studies
a)
"O
CT3 1
"O
£=
3
CO
Funnel plot
i
b HDLC
.
•
/
•
\
•
•
/
/
•
•
o
Effect size
Pseudo 95% CI • Observed studies
Estimated 0Dl • Imputed studies
Figure F-9: Filled-in Funnel Plots to Evaluate Publication Bias of the PFOS and TC (Left) or HDLC (Right) Association.
Note: The funnel plot shows individual studies included in the analysis according to random-effect beta estimates (x-axis) and the standard error of each study-specific beta (y-
axis). The red vertical line indicates the pooled estimate for all studies combined and the gray lines indicate pseudo 95% confidence limits around the pooled estimate. Number of
observed studies: 12 (TC) and 14 (HDLC).
Final PFAS Rule Economic Analysis
F-21
April 2024
-------
FINAL RULE
APRIL 2024
F.4.3 Sensitivity Analyses
The EPA considered two studies for use in single-study sensitivity analyses to understand the
impact of using the estimates from the meta-analyses in the CVD risk reduction modeling output.
These analyses are described in greater detail in Appendix K.
Using data from NHANES (2003-2014) on 8,948 adults, Dong et al. (2019) reported significant
increases in TC: 1.48 (95% CI: 0.18, 2.78) mg/dL per ng/mL serum PFOA and 0.40 (95% CI:
0.13, 0.67) mg/dL per ng/mL PFOS (Table F-2). For HDLC the associations were of -0.03 (95%
CI: -0.44, 0.39) mg/dL per ng/mL PFOA and 0.01 (95% CI: -0.08, 0.11) mg/dL per ng/mL
PFOS. The results were adjusted for age, gender, race, family income index, body mass index,
waist circumference, physical activities, diabetes status, smoking status, and number of alcoholic
drinks per day. Participants using lipid-lowering medications were excluded. As part of
developing the EPA's Final Human Health Toxicity Assessments for PFOA and PFOS, the EPA
considered this medium quality study for estimating point of departure for potential use in
toxicity value derivation (U.S. EPA, 2024b; U.S. EPA, 2024c).
The P.-I. D. Lin et al. (2019) study included participants in a clinical trial of the effect of lifestyle
modifications on pre-diabetes. This study included 888 pre-diabetic adults who were recruited
from 27 medical centers in the US during 1996-1999. The study considered both cross-sectional
(baseline) and prospective assessments, with the results showing evidence of an association
between PFOA and increased TC and hypertriglyceridemia. Each doubling of plasma PFOA
concentration at baseline was associated with 6.1 mg/dL (95% CI: 3.1, 9.0) increase in TC. The
results were adjusted for age, sex, race and ethnicity, marital status, educational attainment,
drinking, smoking, percent of daily calorie from fat intake, daily fiber intake, physical activity
level, and waist circumference at baseline. Participants using lipid-lowering medications were
excluded. The results from the longitudinal analysis were not considered because they were not
presented in a format amenable for dose-response analyses. The study provides another line of
evidence to support associations with TC among adults with pre-diabetes and comparable plasma
PFAS concentrations to the U.S. general population.
F.4.4 Limitations and Uncertainties
Table F-4 summarizes limitations and sources of uncertainty associated with the estimated serum
cholesterol dose-response functions. The effects of these limitations and sources of uncertainty
on estimates of risk reduction and benefits evaluated in the PFAS National Primary Drinking
Water Regulation (NPDWR) are uncertain.
Final PFAS Rule Economic Analysis
F-22
April 2024
-------
FINAL RULE
APRIL 2024
Table F-4: Limitations and Uncertainties in the Analysis of the Serum Cholesterol Dose
Response Functions
Uncertainty/Assumption
Notes
All of the studies included in the meta-analysis, except
one (P.-I. D. Lin et al., 2019), are cross-sectional
designs with various design or methodologic
limitations. The cross-sectional nature of designs could
raise concerns about reverse causality.
Measuring PFOA or PFOS and serum lipids
concurrently, as was the case in cross-sectional
designs, was considered adequate in terms of exposure
assessment timing. Given the long half-lives of PFOA
and PFOS (with median half-lives of 2.7 and 3.5
years, respectively; Ying Li et al., 2018), current blood
serum concentrations are expected to correlate well
with past exposures. Furthermore, although reverse
causality due to reverse causation due to
hypothyroidism (Dzierlenga, Allen, et al., 2020) or
enterohepatic cycling of bile acids (Fragki et al., 2021)
has been suggested, there is not yet clear evidence to
support these reverse causal pathways. Regarding
methodology, several NHANES-based studies (Dong
et al., 2019; He et al., 2018) did not clearly report
whether sampling weights were used in the analyses to
account for the complex sampling design (as is the
norm in such survey-based studies).
Some NHANES-based studies used data from
overlapping NHANES cycles.
Using study results with overlapping years of data
could result in double counting certain data and may
introduce uncertainty in the meta-analysis estimates.
Dong et al. (2019) used data from 2003-2014, while
He et al. (2018) used data from 2003-2012; Jain and
Ducatman (2019) used data from 2005-2014; Fan et
al. (2020) used data from 2011-2014; Liu et al. (2018)
used data from 2013-2014; and Nelson et al. (2010)
used data from 2003-2004. A sensitivity analysis
excluding the overlapping NHANES studies supported
the main findings.
Studies used a variety of statistical models for
estimating the associations of interest (including
NHANES-based studies).
Most studies provided measurements of PFOA and
PFOS in serum, except in three studies that used
measurements in plasma (Chateau-Degat et al., 2010;
Fisher et al., 2013; P.-I. D. Lin et al., 2019).
Distribution of PFAS to plasma is chain-length
dependent, and within human blood fractions, PFOS
and PFOA accumulate to the highest levels in plasma,
followed by whole blood and serum. Typically, the
study-specific estimated associations are rescaled
when the study-specific measurements are in whole
blood, but in common practice serum and plasma-
based associations are not rescaled.
Including these studies in meta-analyses introduces
uncertainty in the estimates.
Final PFAS Rule Economic Analysis
F-23
April 2024
-------
FINAL RULE
APRIL 2024
Table F-4: Limitations and Uncertainties in the Analysis of the Serum Cholesterol Dose
Response Functions
Uncertainty/Assumption
Notes
Existing approaches are limited in their ability to
evaluate statistical heterogeneity and the potential for
publication bias.
The EPA performed statistical evaluations to assess
sources of heterogeneity in effect estimates, and to
evaluate potential for publication bias. However, the
approaches for evaluating heterogeneity and
publication bias are sometimes limited in their ability
to do so. Evaluating statistical heterogeneity in meta-
analyses with a small number of studies is limited by
the potential that the I2 statistic can be imprecise and
biased, and thus results should be interpreted
cautiously (von Hippel, 2015).a In evaluating
publication bias, the funnel plot asymmetry is a
subjective assessment and is recommended only when
at least 10 studies are included in the meta-analysis
(Higgins et al., 2021). Furthermore, the Egger
regression test and Begg's rank tests for publication
bias (Begg & Mazumdar, 1994; Egger et al.,
1997; Egger et al., 2008) may suffer from inflated type
I error and limited power in certain situations,
especially when there is a high degree of heterogeneity
(L. Lin & Chu, 2018). Finally, the small number of
studies reporting slopes from similar models limits the
power of the meta-analysis.
Abbreviations: NHANES-The National Health and Nutrition Examination Survey; PFOA- perfluorooctanoic acid;
PFOS- perfluorooctanesulfonic acid.
Note:
aI2 represents the percentage of variation across studies that is due to heterogeneity rather than chance.
Final PFAS Rule Economic Analysis
F-24
April 2024
-------
FINAL RULE
APRIL 2024
Appendix G. CVD Benefits Model Details and Input
Data
This appendix provides details of the CVD model linking changes in TC, HDLC, and systolic
blood pressure (BP) to changes in incidence of first hard CVD events in populations exposed to
PFOA/ PFOS through drinking water. These approaches have been peer reviewed by the EPA's
SAB; input provided by that organization has been considered in finalizing this analysis (U.S.
EPA, 2022). As discussed in the SAB in-person meetings and the final report (U.S. EPA, 2022),
SAB members and the formal report considered the approaches taken in this document, including
using the life table approach and ASCVD model, to be reasonable and valid approaches for
estimating reduced CVD cases associated with reduced PFOA and PFOS.
TC and HDLC were linked to serum PFOA and serum PFOS, as described in Appendix F.
However, evidence of an association between PFOA and PFOS and HDLC effects was
inconclusive (U.S. EPA, 2024b; U.S. EPA, 2024c); therefore, the EPA modeled HDLC effects
only as part of a sensitivity analysis (see Appendix K). The relationship between BP and serum
PFOS among those not using hypertensive medications is discussed in Section 6.5 of the
economic analysis (EA). First hard CVD events included in the model include non-fatal
myocardial infarction (MI), non-fatal ischemic stroke (IS), and coronary heart disease (CHD)
deaths. The model also captures post-acute CVD mortality experienced by the first non-fatal MI
or IS survivors within 6 years of the initial event.
G.l Model Overview and Notation
The CVD model is designed to estimate a time series of hard CVD event incidence for a
population cohort characterized by sex, race/ethnicity, birth year, and age at the beginning of the
evaluation period (i.e., 2023), and birth year-, age- and sex-specific TC, HDLC, and BP level
time series estimated upstream. The first hard CVD event incidence estimates are generated
using the Pooled Cohort ASCVD model (Goff et al., 2014), whose predictors include age,
cholesterol levels, blood pressure, smoking status, and diabetes status. For those ages 40-80, the
ASCVD model predicts the 10-year probability of a hard CVD event—non-fatal MI, fatal and
non-fatal IS, or CHD death—to be experienced by a person without a prior history of MI, IS,
congestive heart failure, percutaneous coronary intervention, coronary bypass surgery, or atrial
fibrillation. The EPA models post-acute CVD mortality for survivors of the first MI or IS at ages
45-65 using race/ethnicity- and sex-specific estimates at 1-year and 5-year follow-up from Thom
et al. (2001). For survivors of the first MI or IS at age 66 or older, the EPA models post-acute
CVD mortality using estimates at 1- to 6-year follow-ups from S. Li et al. (2019).
The CVD model integrates the ASCVD model predictions and post-acute CVD mortality
estimates in the series of recurrent calculations that produce a life table estimate for the
population cohort of interest (e.g., non-Hispanic White females aged 70 years at the beginning of
the evaluation period). For each PWS, the EPA evaluates population cohorts defined by a
combination of birth year and age in or after 2023 (i.e., pairs of (2023,0), (2022,1), (2021,2),
(1938,85+) and pairs of (2024,0), (2025,0), ... , (2065,0)), sex (males and females), and
race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, Other). In addition to the
standard life table components, such as the annual number of all-cause survivors and deaths for
all ages, for ages 40+, the CVD model estimates the number of surviving persons with and
Final PFAS Rule Economic Analysis
G-l
April 2024
-------
FINAL RULE
APRIL 2024
without a history of hard CVD events, the number of persons experiencing hard CVD events at a
given age, and deaths from CVD and non-CVD causes at a given age.
Figure G-l summarizes the main types of CVD model calculations for a population cohort age 0
at the start of the evaluation period.36 The CVD model calculations are identical across the
race/ethnicity and sex demographic subgroups but use subgroup-specific coefficients.37 For
cohorts born prior to or in 2023, the CVD model is initialized using the PWS-, age-,
race/ethnicity-, and sex-specific number of persons estimated to be alive in 2021. For cohorts
born after 2023, the CVD model is initialized using the PWS-, race/ethnicity-, and sex-specific
number of persons aged 0 estimated to be alive in 2021. PWS- and sex, race/ethnicity-, and age-
specific population details are included in Appendix B. Once the model is initialized, the
following types of calculations occur for each year within the simulation period:
• Recurrent standard life table calculations that rely on the all-cause age-specific annual
mortality rates to evaluate the number of deaths among persons of a specific integer age and
the number of survivors to the beginning of the next integer age. These calculations are
executed whenever the current cohort age is in the 0-39 range. They are represented by the
green segments of the timeline shown in Figure G-l.
• Recurrent life table calculations that separately track subpopulations with and without a
history of hard CVD events, including estimation of the number of annual CVD and non-
CVD deaths (in either subpopulation), as well as the number of annual post-acute CVD
deaths experienced by survivors of the first hard CVD events that occurred, at most, 5 years
ago. These calculations are executed whenever the current cohort age is over age 40.38 These
calculations are represented by the red segment of the timeline in Figure G-l. Figure G-2
further illustrates the year-specific calculations required for explicit tracking of
subpopulations with and without a hard CVD event history.
36 This initial population cohort age is chosen because it allows for the illustration of the full set of calculation types used in the
CVD model.
37 There are different ASC VD model coefficients for non-Hispanic White and non-Hispanic Black males and females. The figure
shows the generalized approach of the CVD model.
38 People 85 years or older are treated as a single cohort in the model. The mortality rates for this cohort are assumed to be the
average mortality rate for those aged 85-100 years. The EPA also relied on serum PFOA/PFOS values at age 85 for the 85+
cohort.
Final PFAS Rule Economic Analysis
G-2
April 2024
-------
FINAL RULE
APRIL 2024
Model initialization
with estimated
location- and
race/ethnicity-specific
male/female age 0
population Recurrent standard life table
calculations using age-,
race/ethnicity-, and
sex-specific all-cause
annual mortality rates
k A
Modeled population is
split into CVD and non-CVD
subpopulations using
age-, race/ethnicity-, and sex-specific
CVD prevalence data
~\
r
Recurrent life table calculations
with explicit treatment of CVD population,
using age- and sex-specific CVD prevalence,
cause-specific annual mortality rates,
ASCVD model-based CVD incidence,
and post-acute CVD event mortality
Figure G-l: Overview of Life Table Calculations in the CVD Model.
Note: The figure illustrates the model for population cohort age 0 at the beginning of the evaluation period (i.e., calendar year 2023). Hie model is initialized using an age 0
PWS-specific population (see Appendix B for PWS population details).
Final PFAS Rule Economic Analysis
G-3
April 2024
-------
FINAL RULE
APRIL 2024
Figure G-2 provides additional information on the post-acute CVD mortality estimation. Each
person included in the surviving current age-specific incident CVD subpopulation39
(corresponding to the group F result in Figure G-2) is tracked for 5 additional years to estimate
the number of CVD deaths occurring in that timeframe. The recurrent estimates rely on age-
specific non-CVD mortality, estimated based on CDC life table data and age- and sex-specific
annual CVD mortality rates, and age- and post-acute CVD mortality, estimated based on Thom
et al. (2001) and S. Li et al. (2019).
39 For example, persons who experienced their first non-fatal MI or IS at age 70 and survived through the first post-event year.
Final PFAS Rule Economic Analysis
April 2024
G-4
-------
FINAL RULE
APRIL 2024
Non-CVD population eligible for
the first hard CVD event in the
following year (A-B-C-D).
First hard CVD event deaths (C)
First hard CVD event survivors (D)
CVD event survivor population (D) is adjusted for post-acute
excess mortality in year 0 since the initial event (E)*
First hard CVD event
survivors at the end of
first post-event year (F)
Post-acute excess deaths
among CVD survivors in
first post-event year (E)
Living subpopulation without prior history
of CVD events.
Note:
* Estimated number of CVD events is an input to
the monetization step.
Deaths occurring at the current integer age
Living subpopulation that experienced first
hard CVD at the current integer age
Current-year calculations
Calculations occurring in years 1-5
following the first hard CVD event
Figure G-2: CVD Model Calculations Tracking CVD
and Non-CVD Subpopulations for a Specific Current Age of Cohort.
Final PFAS Rule Economic Analysis
G-5
April 2024
-------
FINAL RULE
APRIL 2024
Table G-l summarizes the data elements and notation of the CVD model.40 The CVD model
elements fall into four categories: indices, data, quantities computed upstream, and internally
computed quantities. Information sources and computational notes for the model elements
identified as "data" are fully described in Section G.5. Changes in the modeled biomarker levels
(Att>,a,s,t) are a birth year, age, sex, and calendar year-specific quantities computed upstream for
the regulatory alternatives as described in Section 6.5 of the economic analysis.41 Section G.2
describes the estimation of first hard CVD event incidence and post-acute CVD mortality, which
are internally computed quantities. Derivation of the remaining internally computed quantities
for the baseline life table is given in Section G.3.1 and Section G.3.2, while derivation of those
quantities for the regulatory alternative life table is given in Section G.3.3.
Table G-l: CVD Life Table Model Elements and Notation Summary
Model Element
Element Type
Definition
b
s
r
f
V
b,a,s,r,max (0,b)
Ib,a,s,r,t
db,a,s,r,t
b,a,s,r,t,p
Index
Index
Index
Index
Index
Index
Index
Index
Index
Data
Internally computed
quantity
Internally computed
quantity
Data
Internally computed
quantity
Current integer age, A = {0,1,2,... ,99}. The life table model
assumes that all persons are born on January 1.
Current calendar year, t = 0 marks the beginning evaluation
period, t = T marks the end of evaluation period
Calendar birth year, B = {—T, ...,0,1, ...,T — 40}
Sex, 5 = {male, female}
Race/Ethnicity, R = {non — Hispanic White, non —
Hispanic Black, other}
First hard non-fatal CVD event type,
F = {non — fatal MI, non — fatal IS}
Population type: CVD - population with a history of hard CVD
events; OTH - non-CVD population
Cause of death: CVD - cardiovascular disease death; OTH - death
from causes other than CVD
Number of years elapsed since first hard CVD event,
K = {0,1,2,3,4,5}
Living population of age a, sex s, and race/ethnicity r, born in
year b. at the beginning of the evaluation period for the cohort:
t = max (0, b)
Living population born in year b. of sex s and race/ethnicity r,
at the beginning of integer age a and calendar year t
Number of all-cause deaths in population born in year b, of sex s
and race/ethnicity r, at integer age a and calendar year t
Prevalence rate of persons with past experience of hard CVD
events at age a, sex , and race/ethnicity r
Living population born in year b, of type p, sex s, and
race/ethnicity r, at the beginning of integer age a and calendar
year t. Note that lb,o,s,r,t,cvd = 0. 'c- the EPA assumes that
people who have just been born do not have CVD history by
definition.
40 SafeWater was programmed for maximal computational efficiency and SafeWater performs a series of pre-calculations to
reduce model runtime. Therefore, the specific equations in the SafeWater code differ from the equations in this Appendix, but the
end result is mathematically consistent.
41 Total cholesterol change for the baseline life table calculations is 0 by definition.
Final PFAS Rule Economic Analysis
G-6
April 2024
-------
FINAL RULE
APRIL 2024
Table G-l: CVD Life Table Model Elements and Notation Summary
Model Element
Element Type
Definition
1b,a,s,r,t,V,c
Qa,s,r
tfa,s,r,c
Atb,a,s,t
i-b,a,s,r,t iP^-b,a,s,t)
Ya,s,r,f
Pb,a,s,r
fta,s,r,f,k
%b,a,s,r,t
Xb,a,s,r,t
^-b,a,s,r,f,t,0
n-b,a,s,r,f,t,k
mb,a,s,r,t,0
m-b,a,s,r,t,k
An,
b,a,s,r,f,t
Internally computed
quantity
Data
Data
Quantity computed
upstream
Internally computed
quantity
Data
Internally computed
quantity
Data
Internally computed
quantity
Internally computed
quantity
Internally computed
quantity
Internally computed
quantity
Internally computed
quantity
Internally computed
quantity
Internally computed
quantity
Number of deaths from cause c in population born in year b. of
type p, sex s, and race/ethnicity r, throughout integer age a and
calendar year t; deaths from cardiovascular causes occur only in
the CVD population (i.e., d
b,a,s,r,t, OTH.CVD
= 0)
General population probability of all-cause death at integer age a,
sex s, race/ethnicity r
General population probability of death from cause c at integer age
a, sex s, race/ethnicity r
A 3-tuple of modeled changes in TC/HDLC/BP for population
born in year b, of sex s, age a , in calendar year t. Each element
of the 3-tuple is set to 0 for baseline calculations for all three
biomarkers. Additionally, the change in BP is set to 0 for persons
using antihypertensive medications regardless of whether the
baseline or the regulatory alternative is evaluated.
Incidence rate of first hard CVD events for persons born in year b.
of sex s and race/ethnicity r at age a and calendar year t; this rate
is computed using the ASCVD model.
Share of first non-fatal hard CVD event type / among all first hard
CVD events at age a , sex s, race/ethnicity r
Rate of CVD deaths in CVD population born in year b. alive at the
beginning of age a, for sex s and race/ethnicity r
Probability of post-acute CVD death in age a, sex s, and
race/ethnicity r CVD population who experienced first type /
non-fatal hard CVD event k integer years ago
Incident CVD population born in year b, of sex s and
race/ethnicity r, at the beginning of integer age a and calendar
year t
Calibration factor for the incident CVD population born in year b.
of sex s and race/ethnicity r, at the beginning of integer age a and
calendar year t
Uncalibrated number of living age a, sex s, and race/ethnicity r
persons born in year b. whose first type / non-fatal hard CVD
event occurred 0 years ago, corresponding to calendar year t
Number of living age a, sex s, and race/ethnicity r persons born in
year b. whose first type / non-fatal hard CVD event occurred k
years ago, corresponding to calendar year t
Uncalibrated number of CVD deaths among those born in year b.
age a, sex s, and race/ethnicity r persons whose first hard CVD
event occurred 0 years ago, corresponding to calendar year t
Number of CVD deaths among those born in year b. age a, sex s,
and race/ethnicity r persons whose first hard CVD event occurred
k years ago, corresponding to calendar year t
Difference between regulatory alternative and baseline number of
persons born in year b, of sex s and race/ethnicity r, whose first
type / non-fatal hard CVD event occurred at age a, corresponding
to calendar year t
Final PFAS Rule Economic Analysis
G-7
April 2024
-------
FINAL RULE APRIL 2024
Table G-l: CVD Life Table Model Elements and Notation Summary
Model Element Element Type Definition
Internally computed Difference between calendar year t regulatory alternative and
^ quantity baseline number of CVD deaths among age a, sex s, and
b'a's'r't race/ethnicity r persons born in year b. who experienced their first
hard CVD event during calendar years t - 5, t - 4, ... t
Internally computed Difference between regulatory alternative and baseline number of
AiVf t quantity persons whose first type / non-fatal hard CVD event occurred
during calendar year t
Internally computed Difference between regulatory alternative and baseline number of
AMt quantity year t CVD deaths among persons whose first hard CVD event
occurred during calendar years t - 5. t - 4. ... t
Abbreviations: ASCVD - atherosclerotic cardiovascular disease; BP - blood pressure; CVD - cardiovascular disease;
HDLC - high-density lipoprotein cholesterol; TC - total cholesterol.
G.2Hard CVD Event Incidence Estimation
In this section, the EPA describes the process for estimating the probability of the first hard CVD
event ib,a,s,r,t{^Tb,a,s,t) using the ASCVD model (Section G.2.1); the prevalence of persons with
a history of hard CVD events na s r (Section G.2.2); the distribution of first hard CVD events by
type, including the share of non-fatal first hard CVD events Ya,s,r,f (Section G.2.3); and post-
acute CVD mortality rates lia,s,r,f,k within 6 years of the initial event (Section G.2.4).
G.2.1 Probability of the First Hard CVD Event
The first hard CVD event incidence estimates are generated by the Pooled Cohort ASCVD
model (Goff et al., 2014). The ASCVD model is commonly used in clinical practice to estimate
CVD risk for those aged 40-80 years. The ASCVD model predicts the 10-year probability of a
hard CVD event—fatal and non-fatal MI, fatal and non-fatal IS, or CHD death—to be
experienced by a person without a prior history of MI, IS, congestive heart failure, percutaneous
coronary intervention, coronary bypass surgery, or atrial fibrillation.
Four large longitudinal community-based epidemiologic cohort studies have been combined to
develop a geographically and racially diverse dataset used for the ASCVD model estimation:
(1) the Atherosclerosis Risk in Communities Study (Williams, 1989), (2) the Cardiovascular
Health Study (Fried et al., 1991), (3) the Coronary Artery Risk Development in Young Adults
Study (Friedman et al., 1988), and (4) the Framingham Original and Offspring Cohort Study
(Mahmood et al., 2014). Note that there are several other studies whose design is similar to the
one used in Goff et al. (2014), including D'Agostino et al. (2001), D'Agostino et al. (2000),
D'Agostino et al. (2008), D'Agostino et al. (1994), Pencina et al. (2009), Pencina et al. (2011),
Wilson et al. (1998), and Uno et al. (2011). Except for Uno et al. (2011), who also used the
Breast Cancer Survival Study (Chang et al., 2005), including D'Agostino et al. (2001),
D'Agostino et al. (2000), D'Agostino et al. (2008), D'Agostino et al. (1994), Pencina et al.
(2009), Pencina et al. (2011), Wilson et al. (1998), and Uno et al. (2011). Except for Uno et al.
(2011), who also used the Breast Cancer Survival Study (Chang et al., 2005), all of these studies
used the Framingham cohort study data that are not as diverse as the data used to estimate the
ASCVD model.
Final PFAS Rule Economic Analysis
G-8
April 2024
-------
FINAL RULE
APRIL 2024
Table G-2 shows the ASCVD model coefficient estimates used in the analysis. The predictors of
the ASCVD model include age, TC and HDLC concentrations, BP, current smoking, diagnosed
diabetes, and whether the subject is undergoing treatment for high BP. The model has been fit
separately to four population subgroups: non-Hispanic White females, non-Hispanic Black
females, non-Hispanic White males, and non-Hispanic Black males. The EPA applied sex-
specific model coefficients for non-Hispanic Blacks to estimate CVD risk in Hispanic and non-
Hispanic other race population subgroups based on validation of the ASCVD model against
published statistics as described in Section G.4.
Table G-2: ASCVD Model Coefficients
Model Coefficient
Variable Name
Non-Hispanic
Non-Hispanic Black
Non-Hispanic
Non-Hispanic Black
White Females
Females*
White Males
Males*
Ln Age (y)
-29.799
17.114
12.344
2.469
Ln Age, squared
4.884
-
-
-
Ln Total Cholesterol (mg/dL)
13.54
0.94
11.853
0.302
Ln Age x Ln Total Cholesterol
-3.114
-
-2.664
-
LnHDL-C (mg/dL)
-13.578
-18.92
-7.99
-0.307
Ln Age x LnHDL-C
3.149
4.475
1.769
-
Ln Treated Systolic BP (mm Hg)
2.019
29.291
1.797
1.916
Ln Age x Ln Treated Systolic BP
-
-6.432
-
-
Ln Untreated Systolic BP (mm
Hg)
1.957
27.82
1.764
1.809
Ln Age x Ln Untreated Systolic
BP
-
-6.087
-
-
Current Smoker (1 = Yes, 0 = No)
7.574
0.691
7.837
0.549
Ln Age x Current Smoker
-1.665
-
-1.795
-
Diabetes (1 = Yes, 0 = No)
0.661
0.874
0.658
0.645
Mean (Coefficient x Value),
^s,r fis,r
-29.18
86.61
61.18
19.54
ASCVD Baseline Survival, Ss r
0.9665
0.9533
0.9144
0.8954
Abbreviations: ASCVD - atherosclerotic cardiovascular disease; BP - blood pressure; HDLC - high-density lipoprotein
cholesterol.
Note:
*Based on the results of ASCVD model validation exercises (Section G.4), the models for non-Hispanic Black males and
females are applied to other ethnic groups.
Source: Goffetal. (2014), Table A
In order to be used for risk estimation, the ASCVD model needs to be parameterized using
values of the predictors shown in Table G-2 that are appropriate for the current age, sex, and
race/ethnicity of the cohort being evaluated. As shown in Table G-l, current age, sex, and
race/ethnicity are easily accessible indices of the CVD model. In turn, baseline values for the
other ASCVD model predictors come from several public health surveys implemented by the
Centers for Disease Control and Prevention, as detailed in Section G.5.
Final PFAS Rule Economic Analysis
G-9
April 2024
-------
FINAL RULE
APRIL 2024
To compute the 10-year probability of the first hard CVD event for a birth year b, sex s and
race/ethnicity r cohort at age a, the EPA uses the ASCVD risk equation (Goff et al., 2014, Table
G-5) adjusted to express the type of scenario being evaluated (i.e., baseline or regulatory
alternative):
Equation G-l:
D ( at A 1 c exP (ln(Tas,r+AT^a,s,t) [PT,s,r^"PaT,s,r'^(.cO\^"x-T,a,s,r P-T,s,r~%s,r Ps,r)
Kb,a,s,r,t-.t+9\i^l-b,a,s,t) ~ 1 Js,r
where
Rb,a,s,r,t:t+9(^Tb,a,s,t) probability of the first hard CVD event to occur between years t
and t + 9 for a birth year b, sex s / race/ethnicity r person whose age at
time t is a. Rb,a,s,r,t-.t+9(.Q) represents baseline 10-year first hard CVD
event risk, whereas Rb,a,s,r,t-.t+9(,&Tb,a,s,t) expresses regulatory alternative
risk consistent with a birth year b-, age a-, sex s-, calendar year t-specific
change in the baseline TC/HDLC/BP levels Arb a s
Ss r ASCVD baseline CVD event-free survival rate at 10 years, consistent with
the sex s and race/ethnicity r of the cohort being evaluated (see parameter
estimates in Table G-2);
Ta s r a vector of baseline inputs for TC, HDLC, and BP consistent with the
current age a, sex s, and race/ethnicity r of the cohort being evaluated
(see Section G.5);
/?T s r a vector of ASCVD model coefficients for the log-TC, log-HDLC, log-BP
predictors, consistent with the sex s and race/ethnicity r of the cohort
being evaluated (see parameter estimates in Table G-2);
Pax,s,r a vector of ASCVD model coefficient for the interaction between
log-current age and log-TC, log-HDLC, log-BP predictor, consistent with
the sex s and race/ethnicity r of the cohort being evaluated (see parameter
estimates in Table G-2);
X-r,a,s,r'P-T,s,r inner product of the ASCVD model coefficient vector (excluding
TC, HDLC, and BP-related coefficients) and a vector of baseline input
values (excluding TC, HDLC, and BP-related inputs), consistent with the
current age a, sex s, and race/ethnicity r of the cohort being evaluated
(see parameter estimates in Table G-2 and Section G.5); and
%s,r' Ps,r inner product of the ASCVD model coefficient vector and a vector
of average input values in the ASCVD estimation dataset (see parameter
estimates in Table G-2).
Final PFAS Rule Economic Analysis
G-10
April 2024
-------
FINAL RULE
APRIL 2024
To obtain the annual probability of the first hard CVD event, the EPA adjusts
Rb,a,s,r,t-.t+9 fab,a,s,t) as follows:
Equation G-2:
ib,a,s,r,t(.^b,a,s,t) — 1 — — ^b,a,s,r,t-t+^^b,a,s,t)^
where
<-b,a,s,r,t(^Tb,a,s,t) probability of the first hard CVD event to occur in year t for a birth
year b, sex s / race/ethnicity r person whose age at time t is a; and
Rb,a,s,r,t:t+9(^Tb,a,s,t) probability of the first hard CVD event to occur between years t
and t + 9 for a birth year b, sex s / race/ethnicity r person whose age at
time t is a.
G. 2.2 Prevalence of Post Hard CVD Events
Because the population evaluated for the first hard CVD event estimation excludes those with a
history of hard CVD events, model inputs require information on the baseline prevalence of the
past hard CVD event history in the U.S. population. The EPA used the Medical Expenditure
Panel Survey (MEPS) 2010-2017 data to estimate the prevalence of persons with a prior
experience of hard CVD events, including MI, stroke, and other acute CHD events. MEPS is a
nationally representative survey of the U.S. civilian non-institutionalized population
implemented by the Agency for Healthcare Research and Quality (AHRQ). The survey has an
overlapping panel design, tracking individuals for, at most, two years and interviewing
participants, at most, six times. MEPS collects demographic, socioeconomic, and health status
information on the first interview and in each subsequent interview asks about medical events
experienced between the current and the previous interview (generally 4-5 months), as well as
changes in employment status, health insurance coverage, and so forth. Section G.5 provides
additional information on MEPS public use files that have been used in this analysis.
The prevalence of persons with a prior experience of hard CVD events has been estimated by
dividing the number person-years in MEPS interview rounds with a reported history of MI,
stroke, or other CHD by the total number of person-years in subpopulations defined by sex and
round-specific age. The estimated ratios have been adjusted for MEPS complex survey design.
Table G-3 shows the resulting estimates of sex-, race/ethnicity-, and age category-specific
prevalence of persons with prior experience of hard CVD events, along with 95% confidence
intervals that reflect sampling uncertainty. Compared with the prevalence estimates for females,
the estimated prevalence is higher for males in all age categories and for all CVD event
categories. Among adults aged 65 or older, estimated MI, other CHD, and overall prevalence is
highest for non-Hispanic White males, while stroke prevalence is highest among non-Hispanic
Black males. Regardless of the age category, the estimated prevalence of an MI history is higher
for males, while the prevalence of a stoke history is higher for females. The prevalence of other
CHD event history is approximately three to 10 times higher compared with the prevalence of an
MI or stroke history.
Final PFAS Rule Economic Analysis
G-ll
April 2024
-------
FINAL RULE APRIL 2024
Table G-3: Estimated Past Hard CVD Event Prevalence per 100,000
Sex
Age
(years)
Race/
Ethnicity
MI
Stroke
Other CHD
Overall
632
495
5,709
6,292
Males
18-44
NH White
(410-855)
(317-673)
(5,072-6,346)
(5,620-6,965)
5,099
3,314
15,439
17,963
45-64
NH White
(4,569-5,629)
(2,804-3,823)
(14,523-16,355)
(16,930-18,995)
16,477
11,002
41,600
47,465
65 or older
NH White
(15,088-17,865)
(9,956-12,047)
(40,040-43,161)
(45,831^19,099)
436
614
3,886
4,667
Males
18-44
NH Black
(146-726)
(304-924)
(2,998-4,773)
(3,651-5,684)
4,786
5,316
12,261
16,590
45-64
NH Black
(3,928-5,644)
(4,222-6,409)
(10,801-13,720)
(14,898-18,282)
13,768
18,908
30,307
42,090
65 or older
NH Black
(11,218-16,319)
(16,185-21,631)
(26,724-33,891)
(38,368—45,812)
480
180
3,065
3,417
Males
18-44
Hispanic
(293-667)
(75-285)
(2,479-3,651)
(2,816-4,019)
4,299
3,010
9,979
12,584
45-64
Hispanic
(3,383-5,214)
(2,225-3,796)
(8,640-11,318)
(11,045-14,124)
14,071
8,254
25,866
30,548
65 or older
Hispanic
(11,569-16,573)
(6,031-10,477)
(22,420-29,313)
(26,960-34,136)
347
342
3,262
3,669
Males
18-44
NH Other
(122-572)
(75-610)
(2,330-4,194)
(2,695-4,643)
4,338
2,693
11,339
13,638
45-64
NH Other
(3,012-5,665)
(1,791-3,595)
(9,033-13,645)
(11,118-16,158)
12,256
12,354
30,516
36,932
65 or older
Other
(9,167-15,344)
(8,911-15,798)
(25,051-35,982)
(31,240^2,624)
439
830
6,262
6,954
Females
18-44
NH White
(278-600)
(608-1,052)
(5,528-6,997)
(6,223-7,685)
2,199
3,127
15,496
17,925
45-64
NH White
(1,841-2,557)
(2,595-3,659)
(14,522-16,469)
(16,791-19,059)
7,510
10,055
31,861
37,538
65 or older
NH White
(6,686-8,335)
(9,098-11,011)
(30,278-33,445)
(35,913-39,162)
393
1,092
4,628
5,612
Females
18-44
NH Black
(204-582)
(783-1,402)
(3,917-5,338)
(4,847-6,378)
3,484
6,491
15,292
19,596
45-64
NH Black
(2,808-4,160)
(5,640-7,343)
(13,915-16,670)
(17,981-21,210)
8,803
14,188
29,296
38,073
65 or older
NH Black
(7,130-10,476)
(12,304-16,071)
(26,441-32,151)
(35,102—41,045)
313
717
3,690
4,363
Females
18-44
Hispanic
(171-454)
(469-965)
(3,182-4,199)
(3,808-4,918)
2,597
3,627
10,335
12,777
45-64
Hispanic
(1,947-3,248)
(2,864-4,391)
(9,066-11,604)
(11,361-14,193)
7,513
9,469
23,149
29,186
65 or older
Hispanic
(5,953-9,073)
(7,385-11,554)
(20,350-25,948)
(26,206-32,167)
722
383
4,569
4,884
Females
18-44
NH Other
(123-1,320)
(90-675)
(3,181-5,957)
(3,502-6,266)
1,292
2,770
11,098
13,148
45-64
NH Other
(710-1,874)
(1,679-3,860)
(8,978-13,218)
(10,758-15,538)
4,150
7,321
19,001
23,463
65 or older
NH Other
(2,557-5,742)
(5,054-9,589)
(15,308-22,694)
(19,638-27,288)
Final PFAS Rule Economic Analysis
G-12
April 2024
-------
FINAL RULE APRIL 2024
Table G-3: Estimated Past Hard CVD Event Prevalence per 100,000
Sex ,Ag<\ MI Stroke Other CHD Overall
(years) Ethnicity
Abbreviations: MI - myocardial infarction (ICD9 = 410 or MIDX = 1); NH - non-Hispanic; Other CHD - other coronary heart
disease (ICD9 = 413,414,427,428 or CHDDX = 1, ANGIDX = 1, OHRTDX = 1); Stroke (ICD9 = 433,434,435,436 or
STRKDX = 1); 95% confidence interval shown in parentheses below the point estimate.
Source: The EPA analysis based onAlEPS, 2010-2017
G.2.3 Distribution of Fatal and Non-Fatal First Hard CVD Events
The ASCVD model predicts the risk of a composite hard CVD event (i.e., MI, IS, or CHD
death). However, modeling requires separate tracking of morbidity and mortality for life table
calculation purposes. In addition, acute-phase mortality and morbidity valuation depends on the
endpoint (i.e., MI or IS). Therefore, the EPA used MEPS 2010-2017 data to estimate the
distribution of first hard CVD events by type of condition (i.e., MI, stroke, and other CHD). The
EPA estimated the incidence of first hard CVD events by dividing the number of person-years in
MEPS interview rounds with reported new occurrences of MI, stroke, or other CHD by the
number of person-years in MEPS interview rounds without resorted prior experience of CVD
events, in subpopulations defined by race/ethnicity, sex and round-specific age. The EPA
adjusted the estimated ratios for MEPS complex survey design. Distribution of CVD events by
condition type was calculated based on the estimated condition-specific incidence rates.
Table G-4 shows the resulting estimates of sex-, race/ethnicity-, and age category-specific first
hard CVD event incidence, along with 95% confidence intervals that reflect sampling
uncertainty. The table also shows the distribution of first hard CVD events by event type. In
males, 15% to 17% of first hard CVD events are Mis, whereas 13% to 20% of first hard CVD
events are strokes. In females, 8% to 12% of first hard CVD events are Mis, whereas 17% to
28% of first hard CVD events are strokes. The shares of Mis and strokes increase with age for
both sexes. Among adults aged 65 or older, estimated MI, stroke, other CHD, and overall
incidence are highest for non-Hispanic White males and females.
Table G-4: Estimated First Hard CVD Event Incidence and Distribution by CVD
Event Type
Sex
Age (years)
Race/ Ethnicity
MI
Stroke
Other CHD
Overall
82
57
454
540
Males
18-44
NH White
(29-135)
(3-110)
(299-609)
(375-705)
356
333
1,536
2,048
45-64
NH White
(225-486)
(194-471)
(1,213-1,859)
(1,678-2,417)
1,326
2,001
6,233
8,125
65 or older
NH White
(679-1,973)
(1,248-2,754)
(5,035-7,431)
(6,651-9,598)
23
81
363
447
Males
18-44
NH Black
(-3-49)
(4-159)
(156-570)
(227-668)
235
805
1,039
1,862
45-64
NH Black
(64-407)
(399-1,211)
(676-1,401)
(1,339-2,385)
319
765
2,332
3,273
65 or older
NH Black
(-1-639)
(76-1,454)
(1,217-3,447)
(1,926-4,621)
52
40
135
212
Males
18-44
Hispanic
(6-99)
(-4-83)
(55-214)
(111-313)
45-64
Hispanic
276
421
735
1,142
Final PFAS Rule Economic Analysis
G-13
April 2024
-------
FINAL RULE
APRIL 2024
Table G-4: Estimated First Hard CVD Event Incidence and Distribution by CVD
Event Type
Sex
Age (years)
Race/ Ethnicity
MI
Stroke
Other CHD
Overall
(72-479)
(2-839)
(419-1,052)
(625-1,659)
951
816
2,747
3,915
65 or older
Hispanic
(285-1,618)
(349-1,283)
(1,432-4,061)
(2,440-5,390)
72
85
121
278
Males
18-44
NH Other
(-70-215)
(-54-223)
(35-207)
(63-493)
830
548
1,513
2,537
45-64
NH Other
(171-1,489)
(39-1,057)
(643-2,383)
(1,356-3,718)
665
1,232
2,940
4,251
65 or older
NH Other
(-14-1,343)
(431-2,033)
(1,496-4,383)
(2,506-5,997)
56
135
492
646
Females
18-44
NH White
(-21-134)
(54-216)
(317-668)
(437-856)
140
407
1,423
1,865
45-64
NH White
(56-225)
(193-620)
(1,109-1,737)
(1,490-2,240)
831
2,102
4,271
6,294
65 or older
NH White
(533-1,130)
(1,498-2,705)
(3,461-5,081)
(5,358-7,231)
96
57
487
597
Females
18-44
NH Black
(1-191)
(5-108)
(279-695)
(360-834)
196
530
1,168
1,754
45-64
NH Black
(74-318)
(247-812)
(793-1,543)
(1,285-2,223)
382
1,607
3,383
4,546
65 or older
NH Black
(8-756)
(762-2,453)
(2,221-4,545)
(3,179-5,913)
38
78
308
392
Females
18-44
Hispanic
(-24-100)
(25-131)
(130-487)
(190-595)
145
308
664
1,065
45-64
Hispanic
(33-257)
(76-541)
(393-936)
(699-1,432)
992
1,321
2,610
4,456
65 or older
Hispanic
(215-1,768)
(611-2,031)
(1,670-3,550)
(3,348-5,564)
47
315
315
Females
18-44
NH Other
(-46-141)
Omitted
(42-589)
(42-589)
201
399
759
1,297
45-64
NH Other
(-6-409)
(74-724)
(259-1,259)
(627-1,967)
576
1,328
2,689
4,349
65 or older
NH Other
(-43-1,195)
(381-2,276)
(1,234-4,144)
(2,463-6,234)
Abbreviations: MI - myocardial infarction (ICD9 = 410 orMIDX =1); NH - non-Hispanic, Stroke (ICD9 = 433,434,435,436
or STRKDX = 1); Other CHD - other coronary heart disease (ICD9 = 413,414,427,428 or CHDDX = 1, ANGIDX = 1,
OHRTDX = 1); 95% confidence interval shown in parentheses below the point estimate.
Final PFAS Rule Economic Analysis
G-14
April 2024
-------
FINAL RULE
APRIL 2024
The ASCVD model predicts the risk of first MI (fatal and non-fatal), IS (fatal and non-fatal), or
other fatal CHD within the next 10 years. Notably, other non-fatal CHD events are not included
among the CVD event types predicted by the ASCVD model (Goff et al., 2014). Because MEPS
data do not have sufficient information to estimate acute-phase CVD event mortality, the EPA
used AHRQ's Healthcare Cost and Utilization Project (HCUP) data on hospital mortality to
allocate CVD events into fatal and non-fatal categories. Section G.5 provides additional
information on the in-hospital mortality data.
Table G-5 shows sex- and age category-specific probability of in-hospital CVD event death
based on HCUP 2017 inpatient data (Agency for Healthcare Research and Quality, 2017a).
Probability of an in-hospital death is highest for MI events (4.64%), followed by IS events
(4.01%), and then other CHD events (1.07%). This probability grows with age across all CVD
event types and is higher for females when compared with males.
Table G-5: Probability of Hospital Death for a Hard CVD Event
Category
MI (%)
IS (%)
Other CHD (%)
Overall
4.65
4.01
1.07
Age (years)
18-44
1.43
1.91
0
45-64
2.60
2.46
0.67
65-84
5.42
3.88
1.23
85 or older
9.80
7.29
3.14
Sex
Males
4.41
3.71
1.01
Females
5.04
4.30
1.20
Abbreviations: IS - ischemic stroke (ICD10 = 163); MI - myocardial infarction (ICD10 = 121); Other CHD - other coronary
heart
disease (ICD10 = 120,122-125).
Source: HCUP 2017 (Agency for Healthcare Research and Quality, 2017a)
The EPA combined estimates in Table G-4 and Table G-5 to derive the ASCVD event
distribution over the following event types: non-fatal MI, non-fatal IS, and fatal CVD events
(i.e., fatal MI, fatal IS, and other fatal CHD events). Table G-6 shows the final sex-,
race/ethnicity-, and age category-specific estimates of the ASCVD event distribution needed as
the CVD model input. For males, the share of non-fatal MI events is 22% to 58%, the share of
non-fatal IS events is 39% to 77%, and the share of fatal CVD events is 2% to 13%. For females,
the share of non-fatal MI events is 16% to 62%, the share of non-fatal IS events is 36% to 76%,
and the share of fatal CVD events is 2% to 14%. The shares of non-fatal MI decrease with age,
whereas the share of fatal CVD events increase with age. Shares of non-fatal MI are generally
highest among non-Hispanic White males, while shares of non-fatal IS are highest among non-
Hispanic Black males. Among non-Hispanic White females, shares of non-fatal IS are highest
for those aged 45-64 years. Among non-Hispanic Black females, shares of non-fatal IS are
highest for those aged 65-84 years. Among females aged 65 or older, shares of non-fatal MI are
highest in the Hispanic population.
Final PFAS Rule Economic Analysis
G-15
April 2024
-------
FINAL RULE APRIL 2024
Table G-6: Estimated Distribution of Fatal and Non-Fatal First Hard CVD Events
Sex
Age (years)
Race/Ethnicity
Non-Fatal MI
(%)
Non-Fatal IS
(%)
Fatal CVD
Event (%)
Males
18-44
NH White
58
40
1.5
45-64
NH White
50
47
3.7
65-84
NH White
37
57
6.2
85 or older
NH White
34
53
13
Males
18-44
NH Black
22
77
1.7
45-64
NH Black
22
75
2.9
65-84
NH Black
27
66
6.4
85 or older
NH Black
25
62
13
Males
18-44
Hispanic
56
42
1.5
45-64
Hispanic
38
59
3
65-84
Hispanic
50
44
6.1
85 or older
Hispanic
47
41
12
Males
18-44
NH Other
46
53
1.6
45-64
NH Other
58
39
3.1
65-84
NH Other
33
62
5.8
85 or older
NH Other
30
58
12
Females
18-44
NH White
29
69
1.9
45-64
NH White
24
71
4.6
65-84
NH White
26
67
6.5
85 or older
NH White
24
63
13
Females
18-44
NH Black
62
36
1.7
45-64
NH Black
26
70
3.9
65-84
NH Black
18
76
6.7
85 or older
NH Black
16
70
14
Females
18-44
Hispanic
32
66
1.9
45-64
Hispanic
31
65
3.8
65-84
Hispanic
40
54
6.4
85 or older
Hispanic
37
51
12
Females
18-44
NH Other
45
53
1.8
45-64
NH Other
32
64
3.6
65-84
NH Other
28
66
6.5
85 or older
NH Other
26
61
13
Abbreviations: Fatal CVD - includes fatal MI, fatal IS, and fatal other coronary heart disease events; IS - ischemic stroke;
MI - myocardial infarction; NH - non-Hispanic.
G.2.4 Post-Acute CVD Mortality
Persons who have experienced non-fatal MI and non-fatal IS events have elevated post-acute
CVD mortality and morbidity (Roger et al., 2012). The EPA identified four studies that
examined risk factors for secondary hard CVD events. These studies differ in terms of outcomes
tracked (e.g., recurrent MI, recurrent IS, angina, heart failure, CVD, and all-cause death),
conditioning event definition (e.g., MI, IS, CHD), and the length of follow-up for which statistics
are reported (e.g., 1-year follow-up, 5-year follow-up). The data used to estimate the risks of
secondary CVD events differ with respect to average age, sex, and share of individuals who are
White among the participants:
Final PFAS Rule Economic Analysis
G-16
April 2024
-------
FINAL RULE
APRIL 2024
• Data used in Kannel et al. (1999) and D'Agostino et al. (2000) come from the Framingham
Heart Survey (Mahmood et al., 2014) and represent White males
and females approximately age 60.
• Data used in Thom et al. (2001) are from the pooled Atherosclerosis Risk in Communities
Study (Williams, 1989), Cardiovascular Health Study (Fried etal., 1991), and Framingham
Original and Offspring Cohort Study (Mahmood et al., 2014).
This pooled dataset offers representation for Black males and females, in addition to White
males and females, and captures persons aged 45 or older.
• Beatty et al. (2015) used two predominantly White male datasets developed based on the
Heart and Soul Study (Whooley et al., 2008) and the PEACE trial (PEACE Trial
Investigators, 2004), capturing persons aged 67 years and 64 years, on average, respectively.
• S. Li et al. (2019) used data for 2008 and 2012 and two types of conditioning events
(i.e., MI and IS) to assess the risk of secondary events in four large Medicare cohorts:
survivors of the first MI in 2008, survivors of the first IS in 2008, survivors of the first MI in
2012, and survivors of the first IS in 2012.42 These data represent older populations (age 80,
on average) and are not limited to a particular race/ethnicity or sex.
Of the studies that assessed risk factors for secondary hard CVD events, only three focused on
developing a risk prediction model (Beatty et al., 2015; D'Agostino et al., 2000; Kannel et al.,
1999) and only two have changes in cholesterol levels and systolic blood pressure as a primary
predictors (Beatty et al., 2015; D'Agostino et al., 2000). In these two studies, TC, HDLC, and BP
levels do not appear to significantly increase the risk of recurrent CVD events, although
D'Agostino et al. (2000) identified statistically significant relationships between the ratio of TC
to HDLC and probability of recurrent CVD events. Beatty et al. (2015) concluded that
precautionary measures and medication taken by patients who had suffered from a primary CVD
event may decrease the initial risk factors (i.e., TC, HDLC, BP) and may be a reason for the lack
of correlation between secondary CVD events and the modeled biomarkers.
In sum, studies focusing on secondary CVD events point to an elevated risk of these events
among survivors of the first hard CVD event. However, the link between these risks and TC,
HDLC, and BP levels is less clear, with limited supporting evidence coming from decades-old
data evaluated by D'Agostino et al. (2000). Therefore, the CVD model relies on the same
secondary hard CVD event rates to estimate secondary hard CVD event incidence under baseline
and regulatory alternatives. Specifically, the EPA focuses on post-acute CVD mortality as the
secondary event of interest, because other non-fatal secondary CVD events are captured in the
available unit values for first non-fatal MI and IS (see, e.g., O'Sullivan et al., 2011). The EPA
selected estimates in Thom et al. (2001) to model post-acute CVD mortality for survivors of MI
or IS at ages 40-65, because Thom et al. (2001) is the only study that analyzed this age group.
The EPA selected estimates in S. Li et al. (2019) to model post-acute CVD mortality for
survivors of MI or IS at ages 66-89, because cohorts analyzed in S. Li et al. (2019) are the
largest and most representative of the U.S. population compared with the cohorts analyzed by
other studies.
42 Note that relative to other studies with sample sizes of, at most, 10,000, the sizes of these cohorts are 20,000,
on average.
Final PFAS Rule Economic Analysis
G-17
April 2024
-------
FINAL RULE
APRIL 2024
G. 2.5 Survivors of the First Hard CVD Event at Ages 40-65
The EPA used estimates of all-cause post-acute mortality for MI survivors at the 1- and 5-year
follow-ups from Thom et al. (2001) to model post-acute CVD mortality for survivors of
non-fatal MI and non-fatal IS events at ages 45-65. While the EPA was unable to identify
comparable post-acute mortality statistics for non-fatal IS, an analysis of the Medicare
population by S. Li et al. (2019) suggests that post-acute MI mortality is a reasonable
approximation for post-acute IS mortality.43
Table G-7 shows estimated all-cause probability of death following first non-fatal MI by age
category, race/ethnicity, and sex from Thom et al. (2001), as reported in Roger et al. (2012).
These estimates are based on the analysis of pooled data from the Atherosclerosis Risk in
Communities Study (Williams, 1989), the Cardiovascular Health Study (Fried et al., 1991), and
the Framingham Original and Offspring Cohort Study (Mahmood et al., 2014). The estimates are
available only for non-Hispanic Whites and non-Hispanic Blacks.
Table G-7: Post-Acute All-Cause Mortality After the First Myocardial Infarction
Age Group D _,4. . .4 Follow-Up Period Probability of All-Cause Death (%)
/ X -Ejt.il. Ill Cltv y n.
(years) (years) Males Females
45-64 Non-Hispanic White 1 5 9
45-64 Non-Hispanic Black 1 14 8
65 or older Non-Hispanic White 1 25 30
65 or older Non-Hispanic Black 1 25 30
45-64 Non-Hispanic White 5 11 18
45-64 Non-Hispanic Black 5 22 28
65 or older Non-Hispanic White 5 46 53
65 or older Non-Hispanic Black 5 54 58
Abbreviations: MI - myocardial infarction (ICD9 = 410; ICD10 = 121).
Source: Thom etal. (2001)
Table G-8 shows estimated probabilities of post-acute CVD mortality after the first MI. The EPA
derived these probabilities by adjusting all-cause post-acute mortality probabilities reported in
Table G-7 for the ages 45-64 group44 to exclude the probability of death from non-CVD causes.
Section G.5 provides details on an estimation of integer age-, race/ethnicity- and sex-specific
probability of death from non-CVD causes based on the U.S. Life Tables, 2017 (Arias & Xu,
2019) and CVD death rates, 1999-2019 (Centers for Disease Control and Prevention, 2020c).
The last two columns of Table G-8 show annual race/ethnicity- and sex-specific post-acute CVD
death probabilities used by the CVD model in estimation of secondary mortality in years 1-5
following the first non-fatal MI or IS that occurred at ages 45-65. The EPA used post-acute
mortality data for non-Hispanic Whites to estimate mortality effects for the other race/ethnicity
groups.
43 For those aged 65 or older, S. Li et al. (2019) have estimated the probability of death within 1 year after a
non-fatal IS to be 32.07% and the probability of death within 1 year after a non-fatal MI to be 32.09%.
44 The EPA applies post-acute mortality probabilities estimated for ages 45-64 to the survivors of first MI or IS,
ages 45-65, because the magnitude of the annual death probability at age 65 is closer to the average annual
death probability for ages 45-64 than to the average annual death probability for ages 66-99.
Final PFAS Rule Economic Analysis
G-18
April 2024
-------
FINAL RULE
APRIL 2024
Table G-8: Post-Acute Mortality After the First Myocardial Infarction
All-Cause Death Probability
Integer Year (0/0)
Non-CVD Death Probability
(%)b
CVD Death Probability (%)°
MIa
Males
Females Males
Females
Males
Females
All Races/Ethnicitiesd
0
5.6
8.8 0.56
0.38
5.0
8.4
1
1.5
2.7 0.60
0.41
0.93
2.3
2
1.5
2.7 0.65
0.44
0.88
2.3
3
1.5
2.7 0.70
0.48
0.83
2.3
4
1.5
2.7 0.75
0.51
0.78
2.2
Non-Hispanic White6
0
5.0
9.0
-
4.5
8.6
1
1.5
2.3
-
0.91
1.9
2
1.5
2.3
-
0.86
1.9
3
1.5
2.3
-
0.82
1.9
4
1.5
2.3
-
0.76
1.8
Non-Hispanic Black
0
14
8.0
-
12
7.7
1
2.0
5.0
-
1.2
4.3
2
2.0
5.0
-
1.1
4.2
3
2.0
5.0
-
1.1
4.1
4
2.0
5.0
-
1.0
4.1
Abbreviations: CVD - cardiovascular disease; MEPS - Medical Expenditure Panel Survey; MI - myocardial infarction (ICD9
= 410; ICD10 = 121).
Notes:
aPost-acute death probabilities at 1- and 5-year follow-ups in Table G-9 are converted to the integer year-specific post-acute
death probabilities by assuming that the annual death probabilities in years 1^1 are identical. This assumption is supported
by data in S. Li et al. (2019), who report post-acute death probabilities at 1-, 2-, 3-, 4-, 5-, and 6-year follow-ups.
bReported annual probability of non-CVD death is a weighted average of life table age-specific probabilities for ages 45-64.
The weights are the sex-specific age distribution of the first MI survivor population, estimated using MEPS 2010-2017 data.
Tor all race/ethnicity categories, CVD death probability is the difference between all-cause death probability and non-CVD
death probability. For the non-Hispanic White and non-Hispanic Black race/ethnicity categories, the EPA obtained the
estimates
by multiplying the corresponding all-cause post-acute death probability with the all-race/ethnicity ratio of post-acute CVD
death probability to all-cause post-acute death probability.
dRace/Ethnicity-specific data for the ages 45-64 group in Table G-9 are pooled using a sex-specific race/ethnicity distribution
of the first MI survivor population, estimated using MEPS 2010-2017 data.
ePost-acute CVD death probability for non-Hispanic Whites is used to estimate mortality effects for the other race/ethnicity
groups.
Sources: Thom etal. (2001); U.S. Life Tables, 2017 (Arias &Xu, 2019); CVD death rates, 1999-2019 (Centers for Disease
Control and Prevention, 2020c)
Final PFAS Rule Economic Analysis
G-19
April 2024
-------
FINAL RULE
APRIL 2024
G.2.6Survivors of the First Hard CVD Event at Ages 66+
The EPA used the results in S. Li et al. (2019) to estimate the number of post-acute CVD deaths
for survivors of the first MI and IS events, aged 66 years or older at the time of the initial event.
Table G-9 summarizes the key results in S. Li et al. (2019) that are used to parameterize the
CVD model and the results of adjustments that the EPA made to incorporate CVD mortality
information in the model. First, the EPA estimated CVD death probabilities by subtracting non-
CVD death probabilities from all-cause post-acute mortality probabilities reported in S. Li et al.
(2019). The EPA derived the sex- and age-specific non-CVD mortality rates from U.S. Life
Tables, 2017 (Arias & Xu, 2019); CVD death rates, 1999-2019 (Centers for Disease Control and
Prevention, 2020c); and U.S. Life Tables Eliminating Certain Causes of Death, 1999-2000
(Arias et al., 2013). The EPA has averaged age- and sex-specific non-CVD death probabilities
for those age 66 or older using the demographic characteristics of the MI and IS cohorts analyzed
by S. Li et al. (2019). Second, the EPA calculated CVD mortality probability as the difference
between the all-cause death probability and the non-CVD death probability. Third, the EPA
calculated CVD mortality rate multipliers as a ratio of CVD mortality probability to the non-
CVD death probability. The EPA combined these multipliers (reported in Table G-9 for MI and
IS survivors) with age-, sex-, and race/ethnicity-specific non-CVD death rates to obtain post-
acute CVD mortality rates for each cohort included in the analysis.
Final PFAS Rule Economic Analysis
G-20
April 2024
-------
FINAL RULE
APRIL 2024
Table G-9: Post-Acute CVD Mortality Following the First Myocardial Infarction and First
Ischemic Stroke in the Population Aged 66 Years or Older
MI Survivors IS Survivors
0>
0>
Follow-
up
Period
(years)
ll-Cause Death
robability (%)a
[)n-CVD Death
"0liability (%)b
CVD Death
robability (%)°
"3
ctf
13
r
o
s
Multiplierd
ll-Cause Death
robability (%)a
[)n-CVD Death
"0liability (%)b
CVD Death
robability (%)°
[) Mortality Rai
Multiplier"1
< Ph
Z
a.
>
u
< Ph
Z
a.
>
U
0
32
4.3
27
6.4
32
4.5
28
6.1
1
16
4.6
11
2.5
15
4.8
9.9
2.07
2
15
4.9
9.6
1.9
16
5.2
10
2.1
3
14
5.2
9.04
1.7
15
5.5
9.8
1.8
4
14
5.6
8.6
1.5
15
5.9
8.9
1.5
5
14
5.9
8.04
1.4
14
6.2
8.03
1.3
Abbreviations: CVD - cardiovascular disease; IS - ischemic stroke (ICD9 = 433,434; ICD10 = 163); MI - myocardial infarction
(ICD9 = 410; ICD10 = 121).
Notes:
Tor MI, the follow-up year specific all-cause death probability is from S. Li et al. (2019) reported data for the 2008 MI survivor
cohort (N = 26,46). For IS, the follow-up year specific all-cause death probability is from S. Li et al. (2019) reported data for the
2008 IS survivor cohort (N = 17,566).
bNon-CVD annual mortality rate is based on U.S. Life Tables 2017 (Arias & Xu, 2019); CVD death rates, 1999-2019 (Centers
for Disease Control and Prevention, 2020c); and U.S. Life Tables Eliminating Certain Causes of Death, 1999-2000 (Arias et al.,
2013) for those age 66 or older. The annual age- and sex-specific death probabilities were averaged using S. Li et al. (2019) MI/IS
survivor cohort demographic characteristics.
cPost-acute CVD death probability rate is estimated by subtracting the non-CVD annual death probability from the all-cause post-
acute death probability.
dThe CVD mortality rate multiplier is defined as the difference between all-cause death probability and non-C VD death
probability divided by the non-C VD death probability. The CVD model combines the baseline rate multiplier with race/ethnicity-,
age-, and sex-specific non-C VD baseline death rates to obtain mortality rates that are appropriate for the race/ethnicity, age, and
sex of each cohort included in the analysis.
Sources: Li et al. (2019); U.S. Life Tables, 2017 (Arias & Xu, 2019); CVD death rates, 1999-2019 (Centers for Disease Control
and Prevention, 2020c); U.S. Life Tables Eliminating Certain Causes of Death, 1999-2000 (Arias et al., 2013).
G.3 Detailed CVD Model Calculations
Table G-10 provides a guide to sections containing the recurrent CVD model calculations
applicable under conditions defined by initial cohort age, current cohort age, and estimation type.
Estimation types include baseline estimation, regulatory alternative estimation, and risk
reduction estimation. Note that standard life table calculations for current cohort ages 0-39 in
Section G.3.1 apply to both the baseline and regulatory alternative estimation types. The CVD
risk reduction estimation equations in Section G.3.5 apply to ages 40+, for which the model
explicitly estimates the number of first hard CVD events and the number of post-acute CVD
deaths for survivors of the first hard CVD event.
Final PFAS Rule Economic Analysis
G-21
April 2024
-------
FINAL RULE
APRIL 2024
Table G-10: A Mapping of CVD Model Calculations by Initial Cohort Age, Current
Cohort Age, and Estimation Type
Initial
Current Cohort Age (years)
Cohort
Age
0-39
40-65
66+
(years)
Baseline Estimation
0-39
Section G.3.1
Section G.3.2, Section G.3.4
Section G.3.2, Section G.3.4
40-85+
-
Section G.3.2, Section G.3.4
Section G.3.2, Section G.3.4
Regulatory Alternative Estimation
0-39
Section G.3.1
Section G.3.3, Section G.3.4
Section G.3.3, Section G.3.4
40-85+
-
Section G.3.3, Section G.3.4
Section G.3.3, Section G.3.4
Risk Reduction Estimation
0-39
-
Section G.3.5
Section G.3.5
40-85+
-
Section G.3.5
Section G.3.5
Abbreviations: CVD - cardiovascular disease.
G.3.1 Baseline Recurrent Calculations Without Explicit
Treatment of the CVD Population
The number of deaths occurring in year t is estimated using the number of persons alive at the
start of the year, lbiCl,s,r,t-> ar|d all-cause annual probability of death, qasr:
Equation G-3:
db,a,s,r,t ~ 1a,s,r ' Ib,a,s,r,t
The number of persons surviving to the start of the next year is calculated as the difference
between the number of persons alive at the start of the year, lb,a,s,r,t-. ar|d the number of deaths
estimated to occur during the year, db a s r t \
Equation G-4:
lb,a+l,s,r,t+l ~ lb,a,s,r,t ~ db,a,s,r,t
G.3.2Baseline Recurrent Calculations with Explicit Treatment of
the CVD Population
The population of persons alive at the start of year t, lbiCl,s,r,t-> is split into CVD and non-CVD
subpopulations using externally estimated age-, race/ethnicity-, and sex-specific CVD
prevalence, na s r :
Final PFAS Rule Economic Analysis
G-22
April 2024
-------
FINAL RULE
APRIL 2024
Equation G-5:
lb,a,s,r,t,CVD — a,s,r ' Ib,a,s,r,t
Equation G-6:
lb,a,s,r,t,OTH (l ^a,s,r) ' ^b,a,s,r,t
The year t number of non-CVD deaths in the CVD and non-CVD subpopulations is estimated
by applying the annual age-, race/ethnicity-, and sex-specific probability of non-CVD death,
qa,s,r,oth> to the number of persons alive at the start of the year in each subpopulation
(lb,a,s,r,t,cvd and lb,a,s,r,t,othX respectively:
Equation G-7:
db,a,s,r,t,CVD,OTH — la,s,r,OTH ' ^b,a,s,r,t,CVD
Equation G-8:
db,a,s,r,t,OTH,OTH = qa,s,r,OTH-lb,a,s,r,t,OTHdb asr^.oth.oth = Oa.s.r.oTH 1 lb,a,s,r,t,oth
The year t number of CVD deaths in the CVD subpopulation is estimated by applying the annual
CVD death probability, qa,s,r,cvd> to the total population alive at the start of the year, lbA,s,r,t^ net
of deaths from other causes, qa,s,r,OTH^ estimated to occur during the year:
Equation G-9:
d-b,a,s,r,t,CVD,CVD — Qa,s,r,CVD ' (l — Qa.s.r,OTh) ' ^b,a,s,r,t
The number of persons surviving to the start of the next year is estimated as:
Equation G-10:
lb,a+l,s,r,t+l ~ lb,a,s,r,t ~ ^-b,a,s,r,t,CVD,CVD — ^b,a,s,r,t, OTH,OTH — ^-b,a,s,r,t, CVD, OTH
The uncalibrated number of persons experiencing their first hard CVD event in year t is
estimated by applying the baseline annual probability of first hard CVD event, ib,a,s,r,t(0). to the
start-of-the-year number of persons in the non-CVD subpopulation, lb,a,s,r,t,oth. net of non-CVD
deaths, db a s r t 0TH 0TH. The ASCVD model applies to ages 40-80 and predicts a 10-year
probability of the first hard CVD event. However, the EPA uses the ASCVD model to estimate
10-year probability of the first hard CVD event for adults ages 81+ years. For those in 85+ age
group, the EPA uses age 85 as the input to ASCVD model at the start of the evaluation period.
Finally, the EPA uses the externally estimated share of non-fatal first hard CVD events, Ya,s,r,f,
and same-year post-acute CVD mortality probability, fJ.a,s,r,f,o, to compute the number of persons
surviving their first hard type / CVD event in year t:
Final PFAS Rule Economic Analysis
G-23
April 2024
-------
FINAL RULE
APRIL 2024
Equation G-ll:
W-b,a,s,r,f,t,0 (l f^a,s,r,f,o) ' Ya,s,r,f ' ib,a,s,r,t(.0) ' (jb,a,s,r,t,OTH — a,s,r,t,OTH,OTh)
The EPA uses the externally estimated share of fatal first hard CVD events, 1 — £ fEF Ya,s,rj,
and same-year post-acute CVD mortality probability, fJ.a,s,r,f,o,t0 compute the uncalibrated
number of year t deaths in the incident CVD population at baseline:
Equation G-12:
^-b,a,s,r,t, 0 — [l + HfeF^a,s,r,f,0 l) " Ya,s,r,f\ ' ib,a,s,r,t(®) 1 (jb,a,s,r,t, OTH ^-b, a, s,r,t, OTH, OTH,
For calibration purposes, the EPA calculated the incident CVD population size, xb a s rjt, that is
consistent with the reported CVD prevalence rates, 7ra sr and na+1iS,r, and cause-specific
mortality rates, qa,s,r,cvd and qa,s,r,oth:
Equation G-13:
Xb,a,s,r,t ~ TCa+l,s,r^b,a+l,s,r,t+l ~ ^-b,a,s,r,t,CVD ^b,a,s,r,t,CVD,CVD ^b,a,s,r,t,CVD,OTH
The EPA used the incident CVD population size to estimate a calibration factor for scaling raw
ASCVD model-based results:
Equation G-14:
_ 3Cb,a,s,r,t
Xb,a,s,r,t — y ^ , ~
ZjfeFnb,a,s,r,f,t,0 mb,a,s,r,t, 0
Final PFAS Rule Economic Analysis
G-24
April 2024
-------
FINAL RULE
APRIL 2024
Using the estimated calibration factor, the EPA adjusted the raw number of persons surviving
their first hard type / CVD event in year t, nbia,s,r,f,t,o-> ar|d the raw number of year t deaths in
the incident CVD population at baseline, inbaiSrti0, to ensure that the EPA does not project a
larger number of incident events than is consistent with the CVD prevalence statistics and
mortality rates:
Equation G-15:
W-b,a,s,r,f,t,0 ~ UliTl (1, Xb,a,s,r,t) ' ^b,a,s,r,f,t,0
Equation G-16:
^-b,a,s,r,t, 0 — TYlifl (1 > Xb,a,s,r,t) ' ^b,a,s,r,t,0
Finally, the EPA uses the overall number of year t CVD deaths, dbasrtcVDCVD, net of the
number of deaths in the incident CVD population, mb a s r 10, and the size of CVD population
alive at the start of the year, lbia,s,r,t,cvd> to estimate the baseline CVD death rate in the prevalent
CVD population. This quantity is needed to support regulatory alternative estimation:
Equation G-17:
Pb,a,s,r ~ (db,a,s,r,t,CVD,CVD ~ 771b,a,s,r,t,o)/^b,a,s,r,t,CVD
G.3.3 Regulatory Alternative Recurrent Calculations with Explicit
Treatment of the CVD Population
If current cohort age a is equal to the initial cohort age, the sizes of CVD and non-CVD
subpopulations at the start of year 0 are calculated using externally estimated CVD prevalence,
nasr, and the initial population size, lbA,s,r,t• however, the current cohort age a is greater than
the initial cohort age, then the sizes of CVD and non-CVD subpopulations at the start of year t
are the same as the end-of-year t — 1 CVD and non-CVD subpopulation sizes. That is, the CVD
and non-CVD populations are computed in a recurrent manner.
Equation G-18 :
. (^-a,s,r ' Ib,a,s,r,t if CL Qx((Z t, 40)
b,a,s,r,t,cvD - \lb a_lsr t_l cVD if a > max(a - t, 40)
b,a,s,r,t,OTH
Equation G-19:
(l. 7Za,s,r^) ' Ib,a,s,r,t if & 771Qx(q t, 40)
k lb,a-l,s,r,t-l,OTH if CL > mCLX^CL — t, 40)
Final PFAS Rule Economic Analysis
G-25
April 2024
-------
FINAL RULE
APRIL 2024
The year t number of non-CVD deaths in CVD and non-CVD subpopulations is estimated by
applying the annual age-, race/ethnicity-, and sex-specific probability of non-CVD death,
qa,s,r,oth> to the number of persons alive at the start of the year in each subpopulation,
respectively:
Equation G-20:
d-b,a,s,r,t,CVD,OTH ~ 1a,s,r,OTH ' Ib,a,s,r,t,CVD
Equation G-21:
d-b,a,s,r,t,OTH,OTH ~ 1a,s,r,OTH ' Ib,a,s,r,t,OTH
The uncalibrated number of fatal and non-fatal first hard CVD events under the regulatory
alternative is estimated using the same equations (i.e., Eq. G-l 1 and Eq. G-12) as the ones used
for the baseline scenario, except for the non-zero difference between regulatory alternative and
baseline total cholesterol AT^a^t:
Equation G-22:
W-b,a,s,r,f,t,0 (l f^a,s,r,f,o) ' Ya,s,r,f ' ^-b,a,s,r,t(^^b,a,s,t) ' (jb,a,s,r,t,OTH ~ db,a,s,r,t,OTH,OTH)
Equation G-23:
1Tlb,a,s,r,t, 0 — [l + HfeF^a.s.r.f,0 l) " Ya,s,r}f \ ' ib,a,s,r,t(^-b,a,s,t) ' (j-b,a,s,r,t,OTH
db,a,s,r,t, OTH.OTh)
These estimates are used in combination with the baseline calibration factor, Xb,a,s,r,t, ar|d
the EPA-estimated regulatory alternative incident CVD population size, xb a s r t :
Equation G-24:
Xb,a,s,r,t ~ Xb,a,s,r,t(^lfeF^-b,a,s,r,f,t,0 1Tlb,a,s,r,t,o)
Final PFAS Rule Economic Analysis G-26 April 2024
-------
FINAL RULE
APRIL 2024
Using the estimated baseline calibration factor, Xb,a,s,r,t, the EPA adjusted the raw number of
persons surviving their first hard type / CVD event in year t, nbia,s,r,f,t,o-> ar|d the raw number of
year t deaths in the incident CVD population, inbaiSrti0:
Equation G-25:
W-b,a,s,r,f,t,0 ~ UliTl (1, Xb,a,s,r,t) ' ^b,a,s,r,f,t,0
Equation G-26:
^b,a,s,r,t,0 ~ UliTl (1, Xb,a,s,r,t) ' ^b,a,s,r,t,0
The number of CVD deaths at age a during year t is estimated as the sum of the number of
deaths among those whose CVD event history began before age a, pb:a,s,r ' h,a,s,r,t, ar|d the
number of deaths among those who experienced their first CVD event at age a, mb a s r tj0.
The number of deaths among those whose CVD event history began before age a is the product
of the baseline CVD death rate in the CVD subpopulation, pb,a,s,r-. and the size of the CVD
subpopulation at the start of year t, lbjajS,r,t-
Equation G-27:
db,a,s,r,t,CVD,CVD ~ Pb,a,s,r ' Ib,a,s,r,t 771b,a,s,r,t,0
Finally, the following recurrent equations are used to compute the sizes of total, CVD, and non-
CVD populations surviving through to the beginning of year t + 1:
Equation G-28:
lb,a+l,s,r,t+l ~ lb,a,s,r,t ~ db,a,s,r,t,CVD,CVD ~ ^i),a,s,r,t,077/,077/ — fD,OTH
Equation G-29:
lb,a+l,s,r,t+l,CVD ~ Ib,a,s,r,t,CVD %b,a,s,r,t ~ — fD,OTH
Equation G-30:
lb,a+l,s,r,t+l,OTH ~ Ib,a,s,r,t,OTH ~ %b,a,s,r,t ~ ^i),a,s,r,t,OT//,OT//
G.3.4 Recurrent Estimation of Post-Acute CVD Mortality
Survivors of the first type / non-fatal hard CVD event at age a in year t, nb a s r j tj0,
are followed for five future years (i.e., k = 1,2,3,4,5) to evaluate post-acute CVD mortality.
The EPA estimates the number of post-acute CVD deaths among survivors of a first hard CVD
event in year k since the initial event at age a, mb,a+k,s,r,t+k,k-. by (1) adjusting the number of
those who survived k — 1 years after the initial event, nb a+k_liS rjit+k_lik_1, for non-CVD
mortality using externally estimated non-CVD mortality rate, qa+k,s,r,OTH-.-. (2) multiplying the
result by externally estimated post-acute CVD mortality rate, /J.a+k,s,rj,kar|d (3) summing over
the first hard CVD event type /:
Final PFAS Rule Economic Analysis
G-27
April 2024
-------
FINAL RULE
APRIL 2024
Equation G-31:
^b,a+k,s,r,t+k,k ~ / \p-a+k,s,r,f,k ' (l — Ra+k,s,r,OTH) ' W-b,a+k-l,s,r,f,t+k-l,k-l\
'/eF
The EPA estimates the number of survivors of type / first hard CVD event in year k since the
initial event at age a, nb a+kiS r jit+kik, by adjusting the number of those who survived k — 1
years after the initial event, nb a+k-i,s,rj,t+k-i,k-i-. f°r mortality using externally estimated non-
CVD mortality rate, qa+k,s,r,OTH-. and post-acute CVD mortality using rate, l^a+k,s,rj,k-
Equation G-32:
W-b,a+k,s,r,f,t+k,k ~ (l — fta+k,s,r,f,k) ' (l — Ra+k,s,r,OTH) ' ^b,a+k-l,s,r,f,t+k-l,k-l
G.3.5 Risk Reduction Calculations
Assuming that the regulatory alternative is associated with a lower incidence of first hard CVD
events (via lower total cholesterol levels due to lower serum PFAS), at the end of time period t,
the number of avoided type / non-fatal first hard CVD events in the sex s and race/ethnicity r
cohort born in year b and currently age a is estimated as:
Equation G-33:
~Baseline Scenario _ Regulatory Alternative
n'Lb,a,s,r,f,t ~ 'Lb,a,s,r,f,t,0 nb,a,s,r,f,t,0
The number of avoided year t CVD deaths in the first hard CVD population in the sex s and
race/ethnicity r cohort born in year b and currently age a years is:
Equation G-34:
Z5
(^.Baselie Scenario _ Regulatory Alternative^
\mb,a,s,r,t,k mb,a,s,r,t,k )
k=0
Total number of avoided type / non-fatal first hard CVD events in year t is:
Equation G-35:
ANft —/ / / Anb,a,s,r,f,t
*—'aEA,bEB t—'sES t—'rER
Final PFAS Rule Economic Analysis
G-28
April 2024
-------
FINAL RULE
APRIL 2024
Total number of avoided CVD deaths in the first hard CVD population in year t is:
Equation G-36:
AMt- — ^Affli):a,s,r,t
*—iaeA,beB 'ses 'res
G.4ASCVD Model Validation
The validation analysis described herein relied on methodology implemented in R software and
differs slightly from SafeWater MCBC methods. Specifically, SafeWater performs a set of pre-
calculations to maximize computational efficiency and, as such, the order of analytical steps
across R and SafeWater models differs; however, results across models are mathematically
consistent. Furthermore, the R-based model version treats each integer age cohort between 85
and 99 separately, implements the CVD calculations for those aged 40-89 years only, and applies
the ASCVD model-based annual incidence at age 80 years to ages 81-89 because the ASCVD
model has been fit to those aged 40-80 years and predicts the 10-year probability of the first
CVD event.
The EPA generated life table CVD model results for race/ethnicity subpopulations under
different assumptions regarding the applicability of ASCVD coefficients for non-Hispanic
Whites and non-Hispanic Blacks to Hispanic and non-Hispanic other subpopulations. CVD
model inputs are summarized in Table G-12. The size of each subpopulation cohort was
estimated using the 2020 U.S. population size and nationally representative age / sex /
race/ethnicity distribution from the American Community Survey, 2017 (U.S. Census Bureau,
2017). The EPA evaluated the alignment among age-, sex-, and race/ethnicity-specific CVD
incidence prediction using the ASCVD model and age-, sex-, and race/ethnicity-specific CVD
incidence prediction calculated by the CVD model on the basis of race-, sex-, and age-specific
prevalence of persons with a history of CVD events based on MEPS 2010-2017 (see Section
G.2.2); U.S. Life Tables, 2017 (Arias & Xu, 2019); and CVD death rates, 1999-2019 (Centers
for Disease Control and Prevention, 2020c).
For each race/ethnicity, sex, and age combination, the EPA first computed the ratio of CVD
incidence based on reported data and incidence based on the ASCVD model. The EPA then
computed the absolute value of the deviation of this ratio from 1 and averaged the results over
age using population weights for each sex and race/ethnicity subpopulation. Table G-l 1 reports
the resulting alignment metrics for each combination of subpopulation and ASCVD model
coefficient set. Results show that the ASCVD model coefficients for the non-Hispanic Black
model are more consistent with data on CVD prevalence and mortality for Hispanic and non-
Hispanic other race subpopulations than the ASCVD model coefficients for the non-Hispanic
White model.
Final PFAS Rule Economic Analysis
G-29
April 2024
-------
FINAL RULE APRIL 2024
Table G-ll: Summary of ASCVD Model Validation
Alignment of ASCVD Model Predictions with Prevalence and
Mortality Statistics3
Sex
Race/Ethnicity
ASCVD Model Coefficients
Estimated in Non-Hispanic
White Sample
ASCVD Model Coefficients
Estimated in Non-Hispanic
Black Sample
Non-Hispanic White
0.64
-
Males
Non-Hispanic Black
-
0.22
Hispanic
0.44
0.23
Non-Hispanic Other
0.57
0.18
Non-Hispanic White
2.00
-
Females
Non-Hispanic Black
Hispanic
1.53
1.37
0.90
Non-Hispanic Other
1.44
1.07
Note:
aAlignment is represented by the population-weighted absolute value of age-specific |R - 11 within each sex and race/ethnicity
subpopulation, where R is the race/ethnicity-, age-, and sex-specific ratio of C VD incidence computed from reported data and
incidence computed from the ASCVD model.
G.5CVD Model Inputs
Table G-12 summarizes the inputs and data sources used in the CVD model, including survey
health data, model coefficients, Centers for Disease Control and Prevention life tables,
hospitalization data, and mortality incidence data.
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model
Data Element
Modeled Variability
Data Source
Notes
Percentage of
population with
high blood
pressure
Age: 10-year age
groups (ages 40-79)
Sex: males, females
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
NHANES 2011-
2016 (Centers for
Disease Control
and Prevention,
2013b, 2015a,
2015b, 2016b,
2017b, 2017c)
The EPA used the percentage of population with
high blood pressure in 10-year age groups to
estimate the number of exposed individuals with
high blood pressure who are exposed to
PFOA/PFOS in drinking water. The blood
pressure measurement NHANES datasets from
2011-2016 were combined with corresponding
respondent-specific demographic profile,
medical questionnaire, and blood pressure
questionnaire datasets to summarize the
percentage of the non-CVD population that has
high blood pressure for each age-, sex-,
and race-specific stratum.
Final PFAS Rule Economic Analysis
G-30
April 2024
-------
FINAL RULE APRIL 2024
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model
Data Element
Modeled Variability
Data Source
Notes
Percentage of
population
receiving blood
pressure
treatment
Age: 10-year age
groups (ages 40-79)
Sex: males, females
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
NHANES 2011-
2016 (Centers for
Disease Control
and Prevention,
2013b, 2015a,
2015b, 2016b,
2017b, 2017c)
To determine the percentage of the population
with controlled high blood pressure, the
percentage of the populations per age group and
sex who have high blood pressure was
multiplied by the percentage of the populations
per age group and sex who received treatment
for high blood pressure. The blood pressure
measurement NHANES datasets from
2011-2016 were combined with corresponding
respondent-specific demographic profile,
medical questionnaire, and blood pressure
questionnaire datasets to summarize the
percentage of the non-CVD population that is
being treated for having high blood pressure for
each age-, sex-, and race-specific stratum.
Treated,
untreated, and
normal systolic
blood pressure
measurements
Age: age groups 40-
59, 60+
Sex: males, females
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
Treatment status:
controlled,
uncontrolled-high,
uncontrolled-normal
NHANES 2011-
2016 (Centers for
Disease Control
and Prevention,
2013b, 2015a,
2015b, 2016b,
2017b, 2017c)
The blood pressure measurement NHANES
datasets from 2011-2016 were combined with
corresponding respondent-specific demographic
profile, medical questionnaire, and blood
pressure questionnaire datasets to summarize the
percentage of the non-CVD population that is
being treated for having high blood pressure for
each treatment status-, age-, sex-, and
race-specific stratum.
Baseline total
cholesterol level
Age: 10-year age
groups (ages 40-79)
Sex: males, females
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
NHANES 2011-
2016 (Centers for
Disease Control
and Prevention,
2013b, 2015a,
2015b, 2016b,
2017b, 2017c)
The total cholesterol NHANES datasets from
2011-2016 were combined with corresponding
respondent-specific demographic profile and
medical questionnaire datasets to summarize
weighted average total cholesterol levels in
mg/dL for each age-, sex-, and race-specific
stratum in the non-CVD population.
Baseline high
density
lipoprotein
cholesterol level
(HDLC)
Age: 10-year age
groups (ages 40-79)
Sex: males, females
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
NHANES 2011-
2016 (Centers for
Disease Control
and Prevention,
2013a, 2015a,
2015b, 2016a,
2017a, 2017c)
The HDLC NHANES datasets from 2011-2016
were combined with corresponding respondent-
specific demographic profile and medical
questionnaire datasets to summarize weighted
average HDLC levels in mg/dL for each age-,
sex-, and race-specific stratum in the non-CVD
population.
Final PFAS Rule Economic Analysis
G-31
April 2024
-------
FINAL RULE APRIL 2024
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model
Data Element
Modeled Variability
Data Source
Notes
Smoking
prevalence
Age: 10-year age
groups (ages 40-79)
Sex: males, females
Smoking status:
fraction of smokers
NHANES 2011-
2016 (Centers for
Disease Control
and Prevention,
2013d, 2015a,
2015b, 2015d,
2017c, 2017e)
The percentage of smokers and non-smokers in
each stratum were used as inputs in the ASCVD
model, providing results similar to using binary
variables representing that an individual is either
a smoker or a non-smoker and further stratifying
the sample. The smoking NHANES datasets
from 2011-2016 were combined with
corresponding respondent-specific demographic
profile and medical questionnaire datasets to
summarize the percentage of the non-CVD
population that smokes for each age-, sex-,
and race-specific stratum.
Diabetes
prevalence
Age: 10-year age
groups (ages 40-79)
Sex: males, females
Diabetes status:
fraction of diabetics
NHANES 2011-
2016 (Centers for
Disease Control
and Prevention,
2013c, 2015a,
2015b, 2015c,
2017c, 2017d)
The percentage of the population with and
without diabetes in each stratum were used as
inputs in the ASCVD model, providing results
similar to using binary variables representing
that an individual has or does not have diabetes
and further stratifying the sample. The diabetes
NHANES datasets from 2011-2016 were
combined with corresponding respondent-
specific demographic profile and medical
questionnaire datasets to summarize the
percentage of the non-CVD population that has
diabetes for each age-, sex-, and race-specific
stratum.
ASCVD model
coefficients
Sex: males, females
Race: non-Hispanic
White, non-Hispanic
Black
Goffetal. (2014),
Table A
For modeling purposes, the Hispanic
subpopulation was assigned coefficients
estimated for the non-Hispanic White
subpopulation. The model applies to ages
40-89. ASCVD regressors include age, TC,
HDLC, treated systolic BP, untreated systolic
BP, smoking status, and diabetes status.
Annual all-cause
death probability
Sex: males, females
Age: integer ages 0 ...
100
Race/Ethnicity: all,
non-Hispanic White,
non-Hispanic Black,
Hispanic
U.S. Life Tables,
2017 (Arias & Xu,
2019)
The quantity used in modeling is qx (i.e., the
probability of dying between ages x and x + 1).
Life table data for the non-Hispanic other race
category are not available; for subsequent
modeling, all-race life tables are used for
this category.
Final PFAS Rule Economic Analysis
G-32
April 2024
-------
FINAL RULE APRIL 2024
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model
Data Element
Modeled Variability
Data Source
Notes
Annual non-
CVD death
probability for
age 90+
Sex: males, females
Age: integer ages 90 ...
100
Race/Ethnicity: all,
non-Hispanic White,
non-Hispanic Black,
Hispanic
U.S. Life Tables,
2017 (Arias & Xu,
2019); U.S. Life
Tables Eliminating
Certain Causes of
Death, 1999-2000
(Arias et al., 2013)
Annual non-CVD death probability is estimated
by multiplying qx from the 2017 U.S. life tables
by the sex-specific ratio of non-CVD qx to
all-cause qx from 1999-2000 U.S. life tables
eliminating certain causes. Life table data for the
non-Hispanic other race category are not
available; for subsequent modeling, all-race life
tables are used for this category. The 1999-2000
U.S. life tables eliminating certain causes are not
race/ethnicity-specific; the U.S. general
population ratios of non-CVD qx to all-cause qx
were applied to all race/ethnicity categories.
The 1999-2000 U.S. life tables eliminating
certain causes are abridged and report 5-year
rates. The corresponding 5-year ratios are
applied to all individual years within the
5-year range.
Annual non-
CVD death
probability for
ages 40+
Sex: males, females
Age: integer ages 40 ...
89
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
U.S. Life Tables
2017 (Arias & Xu,
2019); CVD death
rates, 1999-2019
(Centers for
Disease Control
and Prevention,
2020c)
Annual non-CVD death probability is estimated
by multiplying qx from 2017 U.S. life tables by
the ratio of non-CVD qx to all-cause qx.
The non-CVD qx estimate was obtained for each
integer age by sex combination as the difference
between all-cause qx from U.S. 2017 life tables
and CVD qx from CDC 1999-2019 cause-
specific mortality rates. U.S. 2017 life table data
for the non-Hispanic other race category are not
available; life tables for the U.S. general
population are used for this category.
CVD prevalence
Sex: males, females
Age: age groups 18-
44, 45-64, 65+
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
Condition: MI, IS,
other CHD, MI + IS +
other CHD conditions
combined
MEPS 2010-2017
(Agency for
Healthcare
Research and
Quality, 2011,
2012a, 2012b,
2013a, 2013b,
2014a, 2014b,
2015a, 2015b,
2016a, 2016b,
2017b, 2017c,
2018, 2019a,
2019b, 2019c)
MEPS longitudinal files were used to obtain
survey weights, design variables, and
information on cardiovascular conditions
(including age at diagnosis) that began prior to
the start date for the survey panel. MEPS
medical conditions files were used to obtain
information on the newly diagnosed conditions
of interest. Specifically, MI events were
identified using ICD9 = 410 or MIDX = 1,
stroke events were identified using ICD9 =
433,434,435,436 or STRKDX = 1,
other CHD were identified using ICD9 =
413,414,427,428 or CHDDX = 1, ANGIDX = 1,
OHRTDX = 1. CVD prevalence was estimated
based on persons whose condition started at an
age prior to the age at which the MEPS round
interview was conducted.
Final PFAS Rule Economic Analysis
G-33
April 2024
-------
FINAL RULE APRIL 2024
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model
Data Element
Modeled Variability
Data Source
Notes
CVD incidence
in the non-CVD
population
Sex: males, females
Age: age groups 18-
44, 45-64, 65+
Race/Ethnicity: non-
Hispanic White, non-
Hispanic Black, non-
Hispanic other,
Hispanic
Condition: MI, IS,
other CHD
MEPS 2010-2017
(Agency for
Healthcare
Research and
Quality, 2011,
2012a, 2012b,
2013a, 2013b,
2014a, 2014b,
2015a, 2015b,
2016a, 2016b,
2017b, 2017c,
2018, 2019a,
2019b, 2019c)
MEPS longitudinal files were used to obtain
survey weights, design variables, and
information on cardiovascular conditions
(including age at diagnosis) that began prior to
the start date for the survey panel. MEPS
medical conditions files were used to obtain
information on the newly diagnosed conditions
of interest. Specifically, MI events were
identified using ICD9 = 410 or MIDX = 1,
stroke events were identified using ICD9 =
433,434,435,436 or STRKDX = 1,
other CHD were identified using ICD9 =
413,414,427,428 or CHDDX = 1, ANGIDX = 1,
OHRTDX = 1. CVD incidence was estimated
based on persons whose condition started at an
age that was the same as the age at which the
MEPS round interview was conducted.
In-hospital death
probability for
CVD events
Sex: males, females
Age: age groups 18-
44, 45-64, 65-84, 85+
Condition: MI, IS,
other CHD
HCUP2017
(Agency for
Healthcare
Research and
Quality, 2017a)
Hospital death probabilities were estimated from
condition-specific hospitalizations identified
using the following ICD10 codes: ICD10 = 121
for MI, I CD 10 = 163 for IS, and ICD10 = 120,
122-125 for other CHD. HCUP reports death
probabilities separately by sex or within age
groups. The EPA estimated age group- and sex-
specific hospital death probabilities by assuming
that male/female relative risk does not vary
across age groups.
1-year, 2-year,
3-year, 4-year,
and 5-year all-
cause mortality
incidence in MI
survivors ages
40-64
Sex: males, females
Race: all
Age: age groups 40-65
Condition: MI
Thometal. (2001);
MI incidence
based on the
MEPS 2010-2017
analysis, U.S. Life
Tables, 2017
(Arias & Xu,
2019)
Thom et al. (2001) sex- and race-specific
estimates for 1-year follow-up and 5-year
follow-up all-cause mortality for ages 45-64 MI
survivors are as reported in Roger et al. (2012)
(the text of the original report is not accessible).
Thom et al. (2001) generated separate estimates
for non-Hispanic White and non-Hispanic Black
persons. To derive sex-specific all-race/ethnicity
estimates, the EPA used MEPS-based
race/ethnicity- and sex-specific MI incidence for
ages 45-64 and assumed that non-Hispanic
White mortality estimates apply to other
race/ethnicity categories. To derive 2-year,
3-year, and 4-year all-cause post-MI mortality
incidence, the EPA further assumed that the
annual probability of death between 1-year
follow-up and 5-year follow-up was constant.
Finally, the EPA assumed that the resulting
estimates apply to ages 40-44 MI survivors and
age 65 MI survivors.
Final PFAS Rule Economic Analysis
G-34
April 2024
-------
FINAL RULE APRIL 2024
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model
Data Element
Modeled Variability
Data Source
Notes
1-year, 2-year,
3-year, 4-year,
5-year, and 6-
year all-cause
mortality
incidence in MI
survivors and IS
Sex: all
Race: all
Age: age group 65+
Condition: MI, IS
S. Lietal. (2019)
S. Li et al. (2019) estimates based on 2008 MI
and 2008 IS Medicare cohorts (see Figure 1 of
the paper) were used. Note that these estimates
are neither race- nor sex-specific.
survivors age
65+
1-year, 2-year,
3-year, 4-year,
and 5-year CVD
mortality
incidence in MI
survivors ages
40-65
Sex: males, females
Race: non-Hispanic
White, non-Hispanic
Black,
Age: age groups 40-65
Condition: MI
Thometal. (2001);
MI incidence
based on the
MEPS 2010-2017
analysis, U.S. Life
Tables, 2017
(Arias & Xu,
2019); CVD death
rates 1999-2019
(Centers for
Disease Control
and Prevention,
2020c)
The EPA used estimated annual age- and sex-
specific non-CVD death probability (estimated
as described above) to calculate the probability
of non-CVD death within the next 1, 2, 3, 4, and
5 years. These probabilities were averaged over
ages 45-64 using MI incidence-based weights
estimated from MEPS 2010-2017 (estimated as
described above). The EPA then subtracted these
estimates from 1-, 2-, 3-, 4-, and 5-year sex-
specific all-cause mortality incidence in MI
survivors ages 45-64 (estimated as described
above) to obtain 1-, 2-, 3-, 4-, and 5-year CVD
mortality incidence. Based on this result, the
EPA estimated the sex-specific ratios of CVD
mortality to all-cause mortality in MI survivors
1, 2, 3, 4, and 5 years after the initial event.
These ratios were applied to non-Hispanic White
and non-Hispanic Black all-cause post-MI
mortality reported in Thom et al. (2001) to
obtain post-acute CVD mortality estimates for
these races. The other race/ethnicity categories
used in modeling were assigned post-acute CVD
mortality rates for non-Hispanic Whites. Finally,
the EPA assumed that the resulting estimates
applied to ages 40-44 MI survivors and to age 65
MI survivors.
1-year, 2-year,
3-year, 4-year,
5-year, and 6-
year CVD
mortality
incidence in MI
survivors and IS
survivors ages
65+
Sex: male, female
Race: all
Age: ages 66 ... 89
Condition: MI, IS
S. Lietal. (2019);
U.S. Life Tables,
2017 (Arias & Xu,
2019); CVD death
rates, 1999-2019
(Centers for
Disease Control
and Prevention,
2020c); U.S. Life
Tables Eliminating
Certain Causes of
Death, 1999-2000
(Arias et al., 2013)
The EPA used estimated annual age- and sex-
specific non-CVD death probability (estimated
as described above) to calculate the probability
of non-CVD death within the next 1, 2, 3, 4, 5,
and 6 years. These results were averaged using
S. Li et al. (2019) 2008 MI/IS cohort age and sex
characteristics. In conjunction with all-cause
post-MI/IS mortality estimates from S. Li et al.
(2019), these estimates were used to estimate the
ratio of CVD mortality to the general population
non-CVD mortality 1, 2, 3, 4, 5, and 6 years
after the initial MI/IS event. The sex- and age-
specific probabilities of CVD death 1, 2, 3, 4, 5,
and 6 years after the initial MI/IS event were
estimated by applying these ratios to sex- and
age-specific non-CVD mortality probabilities.
Final PFAS Rule Economic Analysis
G-35
April 2024
-------
FINAL RULE
APRIL 2024
Table G-12: Summary of Inputs and Data Sources Used in the CVD Model
Data Element
Modeled Variability
Data Source
Notes
Abbreviations: ASCVD - atherosclerotic cardiovascular disease; CHD - coronary heart disease; CVD - cardiovascular disease;
HCUP - Healthcare Cost and Utilization Project; IS - ischemic stroke; MEPS - Medical Expenditure Panel Survey;
MI - myocardial infarction; NCHS - National Center for Health Statistics; NHANES - National Health and Nutrition
Examination Survey; PFOA - perfluorooctanoic acid; PFOS - perfluorooctanesulfonic acid.
Final PFAS Rule Economic Analysis
G-36
April 2024
-------
FINAL RULE
APRIL 2024
Appendix H. Cancer Benefits Model Details and
Input Data
This appendix details the cancer life table approach, the data used to estimate reduced RCC cases
resulting from changes in exposure to PFOA via drinking water, and the data used to estimate
reduced bladder cancer cases resulting from changes in exposures to disinfection byproducts
(DBPs) via drinking water. This appendix also provides baseline kidney, bladder, and liver
cancer statistics.
H.l Details on the Cancer Life Table Approach
This appendix details the life table calculations used to estimate reduced cancer cases among
population cohorts affected by reductions in PFAS and co-occurring contaminant levels at PWS
following implementation of drinking water treatment technologies.
The life table is a metric designed to represent the longevity of people from a certain population.
The inputs to the life table are the age-specific probability of death and the initial population size
(e.g., the retail population served at a given PWS). Based on this information, the life table
computes the number of persons surviving to a specific age, the number of deaths occurring at a
given age, the number of person-years lived at a given age, the number of person-years lived
beyond a given age, and age-specific life expectancy. The details of standard life table
calculations can be found in R. N. Anderson (1999). The EPA has previously used life table
approaches in regulatory analyses, including the analysis of lead-associated health effects in the
2015 Benefit and Cost Analysis for the Effluent Limitations Guidelines, Standards for the Steam
Electric Power Generating Point Source Category (U.S. EPA, 2015), and PM2.5-related health
effects in revisions to the National Ambient Air Quality Standards for ground-level ozone (U.S.
EPA, 2008). Other examples of use of a life table approach among federal agencies include the
EPA's analysis of Benefits and Costs of the Clean Air Act from 1990 to 2020 (U.S. EPA, 201 la)
and the Occupational Safety and Health Administration (OSHA) assessment of lifetime excess
lung cancer, nonmalignant respiratory disease mortality, and silicosis risks from exposure to
respirable crystalline silica (81 FR 16285, March 25, 2016; OSHA, 2010).
To estimate the health effects of changes in exposures to cancer-causing pollutants, the health
risk model tracks evolution of two populations over time - the cancer-free population and the
population living with cancer.43F45 These two populations are modeled for both the baseline
annual exposure scenario and for the regulatory alternative annual exposure scenario.
Populations in the baseline and regulatory alternative exposure scenarios are demographically
identical, but they differ in the pollutant levels to which they are exposed. The EPA assumes that
the population is exposed to baseline pollutant levels prior to technology implementation year
(i.e., change in a given pollutant equals 0) and to alternative pollutant levels that reflect the
impact of treatment implementation under the regulatory alternative. All PWSs with baseline
PFAS exceedances are assumed to upgrade their treatment by 2029 to comply with the final
regulation. To capture these effects while being consistent with the remainder of the benefit
45 When referring to the "cancer-free" population, the EPA is referring to the population that is free of the specific type of cancer
modeled in this analysis, rather than the population that is free of all cancers.
Final PFAS Rule Economic Analysis
H-l
April 2024
-------
FINAL RULE
APRIL 2024
framework, the EPA modeled changes in health outcomes resulting from changes in exposure
over an evaluation period that starts in 2024 and ends in 2105.46
The model tracks all-cause mortality and cancer experience for a set of model populations
defined by sex, location (if modeled), birth year B = 1938, ...,2024,2025, ...,2105,
and age attained by 2024 (for those alive in 2024), which is denoted by A = 0,1,2,3,... 85 + 47
Each model population is followed from age 0 in year B to age min (100,2105 — B) in year
min (B + 100,2105), using a one-year time step. For cohorts born prior to or in 2023, the model
is initialized using the location- (if modeled), age-, race/ethnicity- (if modeled), and sex-specific
number of persons estimated to be alive in 2021. For cohorts born after 2024, the model is
initialized using the location- (if modeled), race/ethnicity-, and sex-specific number of persons
age 0 estimated to be alive in 2021. Location- and sex, race/ethnicity-, and age-specific
population details are included in Appendix B.
Below, the EPA provides a list of variables included in the health risk model (Table H-l) and
describes the process for quantifying the evolution of model population defined by B and A
under baseline exposure assumptions.48 The EPA omits sex and location-specific indices because
calculation steps do not differ across sexes and locations. The EPA then describes the process for
quantifying the evolution of the population under regulatory alternative exposures. Finally, the
EPA describes the process for estimating the total calendar year y-specific health benefits. The
EPA aggregates benefits estimates over all model populations (([B, A) =
{(1938,85+),..., (2024,0), (2025,0),..., (2105,0)}).
Table H-l: Health Risk Model Variable Definitions
Variable Definition
a
Current age or age at cancer diagnosis
xa
A person's lifetime pollutant exposure under the regulatory alternative by age a
Za
A person's lifetime baseline pollutant exposure by age a
LRa
Lifetime risk of cancer per person within age interval [0, a) under the baseline conditions
IRa
Age-specific baseline annual cancer incidence rate per person
B
Birth year
A
Age in 2024 (years) for those alive in 2024, 0 for those born after 2024
P
Number of affected persons of age A in 2024 or persons aged 0 born after 2024
y
Calendar year
x„ „
A person's lifetime pollutant exposure under the regulatory alternative by age a given that this
a,y
age occurs in year y
za,y
A person's lifetime baseline pollutant exposure by age a given that this age occurs in year y
lc=0,a,y(za,y)
The baseline number of cancer-free living individuals at the beginning of age a given that this
age occurs in year y
46 Although benefits of lagged changes in lifetime cancer risk after 2105 may be attributed to changes in contaminant exposure
during the analysis period, the EPA did not model effects beyond this period.
47Note that those born after the start of the evaluation period in 2023 (i.e., during 2024-2105) are always tracked starting from
age 0. As with the CVD model, those aged 85 years or older at the start of the analysis are treated as a single cohort, with
mortality statistics averaged over ages 85-100 years and serum PFOA/PFOS set at values corresponding to age 85 years at the
beginning of evaluation.
48 SafeWater was programmed for maximal computational efficiency and SafeWater performs a series of pre-calculations to
reduce model runtime. Therefore, the specific equations in the SafeWater code differ from the equations in this Appendix, but the
end result is mathematically consistent.
Final PFAS Rule Economic Analysis
H-2
April 2024
-------
FINAL RULE APRIL 2024
Table H-l: Health Risk Model Variable Definitions
Variable Definition
dc=o av(zav) The baseline number of deaths among cancer-free individuals at age a given that this age occurs
in year y
lc=] ay(zay) The baseline number of new cancer cases at age a given that this age occurs in year y
qa Probability of a general population all-cause death at age a
Ta Share of cancer deaths among all-cause deaths at age a
Ya Baseline probability of a new cancer diagnosis at age a
k Cancer duration in years
s Cancer stage (localized, regional, distant, unstaged)
Ss=s a Age-specific share of new stage s cancers
h=s a y o (za y) The baseline number of new stage s cancers occurring at age a given that this age occurs in year
y
rs=SAjk Relative survival rate k years after stage s cancer occurrence at age a
qs=s,a,k Stage-specific probability of death in the cancer population whose cancer was diagnosed at age
a and they lived k years after the diagnosis. Current age of these individuals is a + k
ds=s a y o (za y) The baseline number of deaths in the stage s cancer population in the year of diagnosis (i.e.,
when k = 0), given the current age a and the corresponding year y
h=s a y k (za y-k) The baseline number of individuals living with the stage s cancer in the k-th year after diagnosis
in year y, given the cancer diagnosis at age a and the cumulative exposure through to that age
and year y-k
ds=s a y k (za y_k) The baseline number of deaths among those with the stage s cancer in the k-th year after
diagnosis in year y, given the cancer diagnosis at age a and the cumulative exposure through to
that age and year y-k
es=s ayk(zay-k) The baseline number of excess cancer deaths (i.e., the number of deaths in the cancer population
over and above the number of deaths expected in the general population of the same age) among
those with the stage s cancer in the k-th year after diagnosis in year y, given the cancer
diagnosis at age a and the cumulative exposure through to that age and year y-k
LRa y(zay) Recursive estimate of the lifetime risk of cancer within age interval [0, a) under the baseline
conditions, given that age a occurs in year y
RR (xa y, za y) Relative risk of cancer by age a given that this age occurs in year y, baseline exposure za y and
regulatory alternative exposure xay
LRa y (xa y) Recursive estimate of the lifetime risk of cancer within age interval [0, a) under the regulatory
alternative, given that age a occurs in year y
NCB,A,y,s The incremental number of new stage s cancer cases in year y for the model population (B, A)
LCB,A,y,s The incremental number of individuals living with stage s cancer in year y for the model
population (B,A)
EDBiAiy The incremental number of excess in stage s cancer population in year y for the model
population (B,A)
H.l.l Evolution of Model Population (B,A) under Baseline
Pollutant Exposure
Given a model population (B,A), for each current age a and calendar year y, the following
baseline exposure zay = Baseline Pollutant( y_a+(- dependent quantities are computed:
/c=o,a,y(za,y): The number of cancer-free living individuals at the beginning of age a,
in year y;
d-c=o,a,y(za,y)- The number of deaths among cancer-free individuals aged a during the year y;
Final PFAS Rule Economic Analysis
H-3
April 2024
-------
FINAL RULE
APRIL 2024
lc=i,a,y(za,y): The number of new cancer cases among individuals aged a during the
year y.
To compute each quantity above, the EPA makes assumptions about the priority of events that
terminate a person's existence in the pool of cancer-free living individuals. These events are
general population non-cancer deaths that occur with probability49 qa( 1 — Ta) and new cancer
diagnoses that occur with probability ya, which is approximated by age-specific annual cancer
incidence rate lRa. In the model, the EPA assumes that the new cancer diagnoses occur after
general population non-cancer deaths and use the following recurrent equations for ages a >
0:48F50
To initiate each set of recurrent equations for those alive in 2024, the EPA estimates the number
of cancer-free individuals at age a = 0, denoted by lc=o,o,y-A(,zo,y-A)-. that is consistent with the
number of affected persons of age A in 2024, denoted by P. To this end, Equation H-l,
Equation H-2, and Equation H-3 are estimated as find lc=o,o,y-A(,zo,y-A) = P/Tli=o(.l ~ Ri)
where P = ic=o,a,2024(^,2024)- To initiate each set of recurrent equations for those born after
2024, the EPA uses the PWS-, race/ethnicity-, sex, and scenario-specific number of persons who
died in the previous year of the analysis, thereby ensuring that the size of the modeled population
remains constant throughout the analysis period.
49 The model does not index the general population death rates using the calendar year, because the model relies
on the most recent static life tables.
50 The EPA notes that this is a conservative assumption that results in a lower bound estimate of the regulatory alternative impact
(with respect to this particular uncertainty factor). An upper bound estimate of the regulatory alternative impact can be obtained
by assuming that new cancer diagnoses occur before general population deaths. In a limited sensitivity analysis performed as part
of the Benefit and Cost Analysis for Proposed Revisions to the Effluent Limitations Guidelines and Standards for the Steam
Electric Power Generating Source Category (U.S. EPA, 2019), the EPA found that estimates generated using this alternative
assumption were approximately 5 percent larger than the estimates assuming that new cancer diagnoses occur after general
population deaths.
Final PFAS Rule Economic Analysis H-4 April 2024
-------
FINAL RULE
APRIL 2024
Consistent with available cancer survival statistics, the EPA models mortality experience in the
cancer populations lc=i,a,y{za,y) as dependent on the age-at-onset a, disease duration /t, and
cancer stage s (e.g., localized, regional, distant, unstaged). Given each age-specific share of new
cancer cases lc=i,a,y{za,y) and age-specific share of new stage s cancers 8s=sa, the EPA
calculates the number of new stage s cancers occurring at age a in year y:
Equation H-4:
^¦S=s,a,y,o(za,y) ~ &S=s,a ' ^C=l,a,y(^a,y)
For a model population (B,A) and cancer stage s, the EPA separately tracks min (85,2105 —
B) — A + 1 new stage-specific cancer populations from age-at-onset a to age min (85,2105 —
fi),49F51 Next, a set of cancer duration /c-dependent annual death probabilities is derived for
each population from available data on relative survival rates50F52 rs=s a k and general
population annual death probabilities qa+k as follows:
Equation H-5:
~ i rS=s,a,k+1 ^
Rs=s,a,k ~ -L „ v Ra+kJ
rS=s,a,k
The EPA estimates deaths in the cancer population in the year of diagnosis (i.e., when k = 0)
as follows:
Equation H-6:
ds=s,a,y,o(Za,y) ~ Rs=s,a,0 ' ^¦S=s,a,y,o(za,y)
In years that follow the initial diagnosis year (i.e., k > 0), the EPA uses the following recurrent
equations to estimate the number of people living with cancer and the annual number of deaths
in the cancer population:
Equation H-7:
^S=s,a,y,k(za,y-k) — ^¦S=s,a,y,k-l(za,y-k) ~ dS=s,a,y,k-l(^a,y-k)
Equation H-8:
d-s=s,a,y,k(za,y-k) — Rs=s,a,k ' ^¦S=s,a,y,k(za,y-k)
51 In total, there are 4 ¦ (min (85,2105 — E) — A + 1) new cancer populations being tracked for each model population.
52 Note that rs=SA k is a multiplier that modifies the general probability of survival to age a + k to reflect the fact that the
population under consideration has developed cancer k years ago.
Final PFAS Rule Economic Analysis H-5 April 2024
-------
FINAL RULE
APRIL 2024
Because the agency is interested in cancer-related deaths rather than all deaths in the cancer
population, the EPA also tracks the number of excess cancer population deaths (i.e., the number
of deaths in the cancer population over and above the number of deaths expected in the general
population of the same age). The excess deaths are computed as:
Equation H-9:
&s=s,a,y,k(Za,y-k) ~ Rs=s,a,k ' ^S=s,a,y,k(^a,y-k) ~ tfa+k ' ^S=s,a,y,k(^a,y-k)
H. 1.2 Evolution of Model Population (B,A) under the Regulatory
Alternative Pollutant Exposure
Under the baseline conditions when the change in contaminant levels is zero (i.e., before 2029),
the EPA approximates the annual cancer probability ya by age-specific annual cancer incidence
rate IRa. The EPA computes the pollutant-dependent annual new cancer cases under the
regulatory alternative conditions, /c=i ay(xay), in three steps. First, the EPA recursively
estimates LRa y(zay), the lifetime risk of cancer within age interval [0, a) under the baseline
conditions:
Equation H-10:
^^a,y(za,y) — Tc^oo ~a(zo T) ' ^=0 a ^ ^ LRq y_A(^ZQ y_A^ — 0
Second, the result of Equation H-10 is combined with the relative risk estimate RR(xay,zay),
associated with each cancer type:
Equation H-ll:
^^a,y) — RR(jXa,y> za,y)^^a,y{^a,y)
This results in a series of lifetime cancer risk estimates under the regulatory alternative. Third,
the EPA computes a series of new annual cancer case estimates under the regulatory alternative
as follows:
Equation H-12:
lc=l,a,y(xa,y) ~ (j-,^a+l,y+l{xa+l,y+l) ~ LRaiy(Xa,y)^ ' lc=0,0,y-A(z0,y-A)
H. 1.3 Health Effects and Benefits Attributable to the Regulatory
Alternatives
To characterize the overall impact of the regulatory alternatives in a given year y, for each model
population defined by (B,A), sex, and location, the EPA calculates three quantities: the
incremental number of new stage s cancer cases (NCA y s), the incremental number of individuals
living with stage s cancer (LCA y s), and the incremental number of excess deaths in the cancer
population (EDA y). The formal definitions of each of these quantities are given below:
Final PFAS Rule Economic Analysis
H-6
April 2024
-------
FINAL RULE
APRIL 2024
Equation H-13:
NCB A,y,s = [0 < y — max (2024,5) + A < min (85,2105 — B) ] ¦
(js=s,y-max (2024,B)+4,y,o(^y-max (2024,5) +A,y) ~ ^S=s,y-max (2 0 24,B)+,<4,0 (-^y-max (2024,B)+4,y)^
Equation H-14:
Z100
[0 < y — max (2024, B) + A + k < min (85,2105 — B) ]
^S=s,y-max (2024,B)+4-fc,y,fc(^y-max (2024,B)+A-k,y-k)
: (2024,B)+A-k,y,k (^y-max (2024,B)+A-k,y-k}^
Equation H-15:
^5=s,y-max (
Z100
[0 < y — max (2024, B) + A + k
k=0
< min (85,2105
-B)] X( %=s,y-max (2024,B)+4-fc,y,fc(^y-max (2024,B)+4-fc,y-fc)
ses
~ &S=s,y-max (2024,B)+4-fc,y,fc(-^y-max (2024,B)+4-fc,y-fc)^
These calculations are carried out to 2105.
H.2 Cancer Life Table Model Input Data
As noted in Section 6.6.2 of the economic analysis, the EPA relied on data sources including
SDWIS/Fed, age-, race/ethnicity- and sex-specific population from U.S. Census Bureau (2020)
(See Appendix B), the Surveillance, Epidemiology, and End Results (SEER) program database
(National Cancer Institute), and the CDC National Center for Health Statistics (NCHS) to
characterize sex-, race/ethnicity- and age group-specific general population mortality rates and
cancer incidence rates used in model simulations. Table H-2 summarizes these data sources;
Appendix B provides details on the population size estimates.
Table H-2: Summary of Data Sources Used in Cancer Lifetime Risk Models
Data Element
Modeled Variability
Data Source
Notes
Cancer incidence
rate (IR) per
100,000 persons
Age at diagnosis: 1-year
groups (ages 0 to 100)
Sex: males, females
Cancer type: Kidney
Cancer; Urinary Bladder
(Invasive & In Situ)
Cancer
Race/ethnicity: All, non-
Hispanic White, non-
Hispanic Black, Hispanic,
non-Hispanic Other
Surveillance,
Epidemiology, and End
Results (SEER) 21 cancer
incidence rates by age, sex,
and race at diagnosis for
2014-2018 (Surveillance
Research Program -
National Cancer Institute,
2020b)
Distinct SEER 21 IR data were
available forages 0, 1-4, 5-9, 10-
14, 15-19, 20-24, 25-29, 30-34,
35-39, 40-44, 45-49, 50-54, 55-
59, 60-64, 65-69, 70-74, 75-79,
80-84, 85+. The EPA assumed
that the same IR applies to all
ages within each age group. The
EPA assumed that non-Hispanic
Black iRs can be approximated
by Black iRs. The EPA assumed
Final PFAS Rule Economic Analysis
H-7
April 2024
-------
FINAL RULE
APRIL 2024
Table H-2: Summary of Data Sources Used in Cancer Lifetime Risk Models
Data Element
Modeled Variability
Data Source
Notes
that non-Hispanic Other iRs can
be approximated by all race iRs.
General
population
probability of
death
Age: 1-year groups (ages
0 to 100)
Sex: males, females
Race/ethnicity: All, non-
Hispanic White, non-
Hispanic Black, Hispanic,
non-Hispanic Other
CDC/National Center for
Health Statistics (NCHS)
United States Life Tables,
2017 (Arias &Xu, 2019)
The EPA used race/ethnicity-,
age- and sex-specific
probabilities of dying within the
integer age intervals. The EPA
assumed that non-Hispanic
Other data can be approximated
by all race data.
Share of cancer
deaths among all-
cause deaths
Age at diagnosis: 1-year
groups (ages 0 to 100)
Sex: males, females
Cancer type: Kidney
Cancer; Urinary Bladder
(Invasive & In Situ)
Cancer
Race/ethnicity: All, non-
Hispanic White, non-
Hispanic Black, Hispanic,
non-Hispanic Other
Underlying Cause of
Death, 1999-2019 on CDC
WONDER Online
Database (Centers for
Disease Control and
Prevention, 2020c)
The EPA calculated share of
cancer deaths among all-cause
deaths by race/ethnicity, age and
sex by dividing the number of
cancer deaths during 1999-2019
with the number of all-cause
deaths during 1999-2019.
Share of bladder
cancer incidence
at specific cancer
stage
Age at diagnosis: 1-year
groups (ages 0 to 100)
Sex: males, females
Cancer stage: localized,
regional, distant,
unstaged
Cancer type: Urinary
Bladder (Invasive & In
Situ) Cancer
SEER 21 distribution of
bladder cancer incidence
over stages by age and sex
at diagnosis for 2008-2018
(Surveillance Research
Program - National Cancer
Institute, 2020b)
Distinct SEER 21 data were
available for ages 0-15, 15-39,
40-64, 65-74, 75+. The EPA
assumed that the same cancer
incidence shares by stage apply
to all ages within each age
group.
Share of kidney
cancer incidence
at specific cancer
stage
Age at diagnosis: 1-year
groups (ages 0 to 100)
Sex: males, females
Cancer stage: localized,
regional, distant,
unstaged
Cancer type: Kidney
Cancer
Race/ethnicity: All, non-
Hispanic White, non-
Hispanic Black, Hispanic,
non-Hispanic Other
SEER 21 distribution of
kidney cancer incidence
over stages by
race/ethnicity, age and sex
at diagnosis for 2008-2018
(Surveillance Research
Program - National Cancer
Institute, 2020b)
Distinct SEER 21 data were
available for ages 0-15, 15-39,
40-64, 65-74, 75+. The EPA
assumed that the same cancer
incidence shares by stage apply
to all ages within each age
group. The EPA assumed that
non-Hispanic Black data can be
approximated by Black data. The
EPA assumed that non-Hispanic
Other data can be approximated
by all race data.
Relative bladder
cancer survival by
cancer stage
Age at diagnosis: 1-year
groups (ages 0 to 100)
Sex: males, females
Duration: 1-year groups
(durations 0 to 100 years)
Cancer stage: localized,
regional, distant,
unstaged
Cancer type: Urinary
Bladder (Invasive & In
Situ) Cancer
SEER 18 relative bladder
cancer survival by age at
diagnosis, sex, cancer stage
and duration with
diagnosis for 2000-2017
(Surveillance Research
Program - National Cancer
Institute, 2020a)
Distinct SEER 18 data were
available for ages at diagnosis
0-14, 15-39, 40-64, 65-74, 75+.
The EPA assumed that the same
cancer relative survival patterns
apply to all ages within each age
group. SEER 18 contained data
on relative survival among
persons that had bladder cancer
forO, 1,2,3,4, 5,6, 7, 8, 9, and
10 years. For disease durations
Final PFAS Rule Economic Analysis
H-8
April 2024
-------
FINAL RULE
APRIL 2024
Table H-2: Summary of Data Sources Used in Cancer Lifetime Risk Models
Data Element
Modeled Variability
Data Source
Notes
longer than 10 years the EPA
applied 10-year relative survival
rates.
Relative kidney
cancer survival by
cancer stage
Age at diagnosis: 1-year
groups (ages 0 to 100)
Sex: males, females
Duration: 1-year groups
(durations 0 to 100 years)
Cancer stage: localized,
regional, distant,
unstaged
Cancer type: Kidney
Cancer
Race/ethnicity: All, non-
Hispanic White, non-
Hispanic Black, Hispanic,
non-Hispanic Other
SEER 18 relative kidney
cancer survival by
race/ethnicity, age at
diagnosis, sex, cancer stage
and duration with
diagnosis for 2000-2017
(Surveillance Research
Program - National Cancer
Institute, 2020a)
Distinct SEER 18 data were
available for ages at diagnosis
0-14, 15-39, 40-64, 65-74, 75+.
The EPA assumed that the same
cancer relative survival patterns
apply to all ages within each age
group. The EPA assumed that
non-Hispanic Black data can be
approximated by Black data. The
EPA assumed that non-Hispanic
Other data can be approximated
by all race data. SEER 18
contained data on relative
survival among persons that had
kidney cancer for 0, 1, 2, 3, 4, 5,
6, 7, 8, 9, and 10 years. For
disease durations longer than
10 years the EPA applied 10-
year relative survival rates.
Abbreviations: CDC - Centers for Disease Control and Prevention; EPA - U.S. Environmental Protection Agency; IR -
incidence ratio; NCHS - National Center for Health Statistics; SEER - Surveillance, Epidemiology, and End Results.
H.3 Baseline Kidney Cancer Statistics
Table H-3 provides baseline kidney cancer incidence data used in the life table model. Kidney
cancer incidence rates per 100,000 range from 0.25 to 44 for females and from 0.16 to 96 for
males. Kidney cancer incidence rates are highest for men in their 60s, 70s, and 80s, ranging from
62 per 100,000 to 96 per 100,000. Localized kidney cancers comprise 37%-84% of all kidney
cancer incidence, whereas regional kidney cancers comprise 8.0%-34%, distant kidney cancers
comprise 6.0%-26%, and unstaged kidney cancers comprise 1.7%-l 1 % of all kidney cancer
incidence. Table H-4 provides baseline kidney cancer incidence data by race/ethnicity used in the
life table model.
Final PFAS Rule Economic Analysis
H-9
April 2024
-------
FINAL RULE APRIL 2024
Table H-3: Summary of Baseline Kidney Cancer Incidence Data Used in the Model
Females
Males
Percent of Incidence in Stage
Percent of Incidence in Stage
Age
Incidence
Localized
13
Unstaged
Incidence
Localized
13
Unstaged
per 100K
e
"Si
£
es
"S
5
per 100K
e
"Si
£
es
"S
5
<1
1.6
37
34
26
3.1
1.9
43
33
21
3.3
1-4
2.0
37
34
26
3.1
1.8
43
33
21
3.3
5-9
0.82
37
34
26
3.1
0.53
43
33
21
3.3
10-14
0.25
37
34
26
3.1
0.18
43
33
21
3.3
15-19
0.27
84
8.0
6.0
1.9
0.16
81
10
7.7
1.7
20-24
0.60
84
8.0
6.0
1.9
0.51
81
10
7.7
1.7
25-29
1.1
84
8.0
6.0
1.9
1.3
81
10
7.7
1.7
30-34
2.7
84
8.0
6.0
1.9
3.5
81
10
7.7
1.7
35-39
4.7
84
8.0
6.0
1.9
7.2
81
10
7.7
1.7
40-44
7.8
77
11
10
1.8
14
70
14
13
2.1
45-49
11
77
11
10
1.8
22
70
14
13
2.1
50-54
16
77
11
10
1.8
33
70
14
13
2.1
55-59
22
77
11
10
1.8
47
70
14
13
2.1
60-64
29
77
11
10
1.8
62
70
14
13
2.1
65-69
37
71
14
13
2.9
81
67
16
14
3.2
70-74
41
71
14
13
2.9
91
67
16
14
3.2
75-79
44
59
12
17
11
96
57
16
17
9.3
80-84
40
59
12
17
11
84
57
16
17
9.3
85+
33
59
12
17
11
68
57
16
17
9.3
Final PFAS Rule Economic Analysis
H-10
April 2024
-------
FINAL RULE
APRIL 2024
Table H-4: Summary of Race/Ethnicity-Specific Baseline Kidney Cancer Incidence Data
Used in the Model
Age
__l
1-4
5-9
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85+
__l
1-4
5-9
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85+
Females
Incidence
per 100K
1.4
2.2
0.2
0.32
0.52
1.2
2.9
4.9
12
16
22
28
37
40
46
41
33
2.4
0.88
0.84
1.1
2.4
3.8
7.4
l_l_
16_
23
38
46
49
47
46
37
Percent of Incidence in Stage
-------
FINAL RULE
APRIL 2024
Table H-4: Summary of Race/Ethnicity-Specific Baseline Kidney Cancer Incidence Data
Used in the Model
Race/Ethnicity
Age
Females
Males
Incidence
per 100K
Percent of Incidence in Stage
Incidence
per 100K
Percent of Incidence in
Stage
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Hispanic
<1
-
35
35
27
3
1.6
50
28
20
1.7
1-4
1.7
35
35
27
3
1.7
50
28
20
1.7
5-9
0.57
35
35
27
3
0.51
50
28
20
1.7
10-14
-
35
35
27
3
-
50
28
20
1.7
15-19
-
84
8.6
5.4
1.7
-
79
11
8.2
2.1
20-24
0.63
84
8.6
5.4
1.7
0.47
79
11
8.2
2.1
25-29
1
84
8.6
5.4
1.7
0.92
79
11
8.2
2.1
30-34
2.8
84
8.6
5.4
1.7
3
79
11
8.2
2.1
35-39
5.9
84
8.6
5.4
1.7
6.4
79
11
8.2
2.1
40-44
9.2
76
12
10
2.1
13
67
15
15
2.4
45-49
13
76
12
10
2.1
20
67
15
15
2.4
50-54
19
76
12
10
2.1
30
67
15
15
2.4
55-59
24
76
12
10
2.1
45
67
15
15
2.4
60-64
34
76
12
10
2.1
62
67
15
15
2.4
65-69
42
69
14
14
2.9
83
66
16
15
3.6
70-74
46
69
14
14
2.9
91
66
16
15
3.6
75-79
45
59
12
17
12
96
54
18
19
9
80-84
39
59
12
17
12
79
54
18
19
9
85+
35
59
12
17
12
70
54
18
19
9
Other
<1
1.6
37
34
26
3.1
1.9
43
33
21
3.3
1-4
2
37
34
26
3.1
1.8
43
33
21
3.3
5-9
0.82
37
34
26
3.1
0.53
43
33
21
3.3
10-14
0.25
37
34
26
3.1
0.18
43
33
21
3.3
15-19
0.27
84
8
6
1.9
0.16
81
10
7.7
1.7
20-24
0.6
84
8
6
1.9
0.51
81
10
7.7
1.7
25-29
1.1
84
8
6
1.9
1.3
81
10
7.7
1.7
30-34
2.7
84
8
6
1.9
3.5
81
10
7.7
1.7
35-39
4.7
84
8
6
1.9
7.2
81
10
7.7
1.7
40-44
7.8
77
11
10
1.8
14
70
14
13
2.1
45-49
11
77
11
10
1.8
22
70
14
13
2.1
50-54
16
77
11
10
1.8
33
70
14
13
2.1
55-59
22
77
11
10
1.8
47
70
14
13
2.1
60-64
29
77
11
10
1.8
62
70
14
13
2.1
65-69
37
71
14
13
2.9
81
67
16
14
3.2
70-74
41
71
14
13
2.9
91
67
16
14
3.2
75-79
44
59
12
17
11
96
57
16
17
9.3
80-84
40
59
12
17
11
84
57
16
17
9.3
85+
33
59
12
17
11
68
57
16
17
9.3
Final PFAS Rule Economic Analysis
H-12
April 2024
-------
FINAL RULE
APRIL 2024
Table H-5 shows relative kidney cancer survival rates53 by sex, age group at diagnosis, cancer
stage, and the number of years post diagnosis. The relative kidney cancer survival ranges from
3.2% to 100%, and generally decreases as the number of years post-diagnosis increases. The
table also shows the absolute survival probability, averaged over the age range for which the
relative survival data were available; these probabilities are a product of general population
survival probability and the relative kidney cancer survival probability by sex, age group at
diagnosis, and the number of years post-diagnosis. The life table model uses derived absolute
survival probabilities to model all-cause mortality experience in kidney cancer populations for
the baseline scenario and the regulatory alternatives. Table H-6 provides kidney cancer survival
rates by race/ethnicity used in the life table model. Finally, Table H-7 shows all-cause and
kidney cancer mortality rates used in the life table model. Kidney cancer deaths represent <1% of
all-cause mortality among females and <2% of all-cause mortality among males. Table H-8
provides all-cause and kidney cancer mortality rates by race/ethnicity used in the life table
model.
53 Relative kidney cancer survival rate is the probability of being alive K years after diagnosis at age A divided by the general
probability to survive K years for a person alive at age A without such a diagnosis.
Final PFAS Rule Economic Analysis
April 2024
H-13
-------
FINAL RULE APRIL 2024
Table H-5: Summary of Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by
Stage (Percent)
Absolute Survival (Average) by
Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
¦a
.a
13
u
o
-J
CS
S
o
"Si
C
es
"S
a
¦a
0>
#J3
a
%
s
P
¦a
0>
.a
13
u
o
-J
CS
S
o
"Si
£
C
es
"S
5
¦a
0>
ex
a
%
s
P
¦a
0>
.a
13
u
o
-J
CS
S
o
"Si
£
C
es
"S
Q
¦a
ex
a
%
s
P
¦a
.a
13
o
o
-J
CS
S
o
"Si
£
fi
es
"S
Q
¦a
a
%
s
P
Ages
<15
1 year
99
99
92
100
99
98
91
99
99
99
88
-
99
98
88
-
Ages
<15
2 years
98
97
86
100
98
97
85
99
99
96
79
-
98
95
78
-
Ages
<15
3 years
98
95
83
96
97
94
82
96
97
95
76
-
96
95
75
-
Ages
<15
4 years
97
94
81
92
97
93
81
92
97
95
74
-
96
94
73
-
Ages
<15
5 years
97
93
80
92
96
93
79
92
97
94
73
-
96
93
72
-
Ages
<15
6 years
96
93
79
92
95
93
79
92
96
94
72
-
95
93
71
-
Ages
<15
7 years
95
93
79
87
95
92
79
86
96
94
71
-
95
93
70
-
Ages
<15
8 years
95
93
78
87
95
92
78
86
96
94
70
-
95
93
69
-
Ages
<15
9 years
95
93
78
87
95
92
78
86
96
92
69
-
95
91
68
-
Ages
<15
10
years
95
93
78
87
95
92
78
86
96
92
69
-
95
90
68
-
Ages
15-39
1 year
99
93
50
90
99
92
49
89
99
92
42
91
97
90
41
89
Ages
15-39
2 years
99
85
32
83
98
84
31
82
99
85
27
84
97
83
26
83
Ages
15-39
3 years
98
80
24
77
97
79
24
76
98
78
20
83
96
76
19
81
Ages
15-39
4 years
98
75
21
77
97
74
21
76
98
74
15
83
95
72
14
81
Final PFAS Rule Economic Analysis
H-14
April 2024
-------
FINAL RULE APRIL 2024
Table H-5: Summary of Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by
Stage (Percent)
Absolute Survival (Average) by
Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages
15-39
5 years
97
73
16
77
96
72
16
76
97
71
12
79
94
69
12
77
Ages
15-39
6 years
97
72
15
77
96
71
15
76
96
69
10
72
93
67
10
70
Ages
15-39
7 years
97
71
14
77
95
70
14
76
95
68
9
69
92
65
9
67
Ages
15-39
8 years
96
70
13
77
95
69
13
76
95
66
8
66
92
64
7
64
Ages
15-39
9 years
96
69
13
77
94
68
12
76
94
65
8
66
91
62
7
63
Ages
15-39
10
years
95
69
13
77
93
68
12
76
94
65
8
66
90
62
7
63
Ages
40-64
1 year
99
91
43
73
94
87
40
70
99
92
46
78
90
84
42
71
Ages
40-64
2 years
98
85
28
67
92
80
26
63
97
86
31
69
89
78
28
63
Ages
40-64
3 years
97
80
21
64
91
75
19
60
96
81
23
64
87
73
20
58
Ages
40-64
4 years
96
77
17
61
89
72
15
57
95
77
18
61
85
69
16
54
Ages
40-64
5 years
95
74
14
60
88
69
13
55
94
74
14
58
83
65
13
51
Ages
40-64
6 years
94
71
12
56
87
66
11
52
92
71
12
55
81
62
11
48
Ages
40-64
7 years
93
69
11
55
85
63
10
50
91
68
11
52
79
58
9
45
Ages
40-64
8 years
92
66
10
52
83
60
9
47
90
65
9
50
77
55
8
43
Final PFAS Rule Economic Analysis
H-15
April 2024
-------
FINAL RULE APRIL 2024
Table H-5: Summary of Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by
Stage (Percent)
Absolute Survival (Average) by
Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages
40-64
9 years
91
64
9
50
82
57
8
45
89
63
9
48
75
53
7
40
Ages
40-64
10
years
90
63
8
50
80
56
7
44
87
60
8
45
72
50
6
38
Ages
65-74
1 year
98
89
38
66
90
82
35
61
98
90
41
67
87
80
37
60
Ages
65-74
2 years
97
82
24
58
88
75
22
53
97
84
26
60
84
73
23
52
Ages
65-74
3 years
95
76
17
53
85
68
16
47
95
78
19
54
80
66
16
45
Ages
65-74
4 years
94
73
14
49
82
64
12
43
94
74
15
48
77
60
13
39
Ages
65-74
5 years
92
69
11
47
79
59
9
40
92
70
12
44
73
55
10
35
Ages
65-74
6 years
90
66
10
46
75
55
8
38
91
67
10
42
69
52
8
32
Ages
65-74
7 years
88
63
8
44
72
51
7
36
89
65
9
37
65
48
7
27
Ages
65-74
8 years
87
61
8
39
68
48
6
31
87
63
8
37
61
44
6
26
Ages
65-74
9 years
85
57
7
35
65
43
5
27
86
61
8
34
58
41
5
23
Ages
65-74
10
years
83
53
6
34
60
39
5
25
85
57
7
32
54
37
4
20
Ages
75+
1 year
92
78
22
49
47
40
11
25
94
83
28
52
46
41
14
26
Ages
75+
2 years
91
71
12
38
46
35
6
19
93
77
17
45
44
37
8
21
Final PFAS Rule Economic Analysis
H-16
April 2024
-------
FINAL RULE APRIL 2024
Table H-5: Summary of Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by
Stage (Percent)
Absolute Survival (Average) by
Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages
75+
3 years
89
66
9
32
43
32
5
16
92
74
12
38
42
34
5
17
Ages
75+
4 years
88
61
7
29
41
29
4
13
89
70
9
32
39
31
4
14
Ages
75+
5 years
86
57
6
25
39
26
3
11
88
67
7
27
36
28
3
11
Ages
75+
6 years
84
54
5
24
36
24
2
10
87
62
6
23
34
24
2
9
Ages
75+
7 years
81
51
5
22
34
21
2
9
85
60
6
20
31
22
2
7
Ages
75+
8 years
78
50
5
19
31
20
2
8
82
57
5
19
28
20
2
7
Ages
75+
9 years
74
47
4
18
28
18
1
7
81
55
4
17
26
17
1
5
Ages
75+
10
years
72
42
3
18
25
15
1
6
79
52
4
16
23
15
1
5
Final PFAS Rule Economic Analysis
H-17
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Race/Ethnicity
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Non-Hispanic White
Ages <15
1 year
100
98
95
-
99
98
94
-
99
99
92
-
98
99
92
-
Ages <15
2 years
99
98
90
-
98
98
90
-
99
95
88
-
98
95
88
-
Ages <15
3 years
98
94
85
-
98
94
85
-
96
95
85
-
95
95
85
-
Ages <15
4 years
98
94
85
-
97
93
85
-
96
95
84
-
95
95
84
-
Ages <15
5 years
98
93
83
-
97
92
82
-
96
94
84
-
95
94
83
-
Ages <15
6 years
98
93
83
-
97
92
82
-
96
94
83
-
95
94
82
-
Ages <15
7 years
97
93
83
-
96
92
82
-
96
93
82
-
95
92
81
-
Ages <15
8 years
97
93
83
-
96
92
82
-
96
93
82
-
95
92
81
-
Ages <15
9 years
97
93
83
-
96
92
82
-
96
91
82
-
95
90
81
-
Ages <15
10
years
97
93
83
-
96
92
82
-
96
91
82
-
95
90
81
-
Ages 15-
39
1 year
100
97
58
-
99
96
58
-
99
91
52
96
97
89
51
94
Ages 15-
39
2 years
99
91
38
-
98
90
38
-
99
87
33
84
97
85
33
82
Ages 15-
39
3 years
99
85
27
-
98
84
27
-
99
83
25
84
96
81
24
82
Ages 15-
39
4 years
99
82
21
-
97
81
21
-
98
78
18
84
96
76
18
82
Ages 15-
39
5 years
98
80
18
-
97
79
18
-
97
77
14
84
95
75
14
81
Ages 15-
39
6 years
98
77
18
-
96
76
18
-
97
75
13
79
94
73
13
77
Final PFAS Rule Economic Analysis
H-18
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 15-
39
7 years
98
76
17
-
96
74
16
-
96
73
10
79
93
70
10
77
Ages 15-
39
8 years
97
74
17
-
96
73
16
-
96
72
7.2
79
92
69
7
77
Ages 15-
39
9 years
97
74
17
-
95
73
16
-
95
72
7.2
79
91
69
7
76
Ages 15-
39
10
years
96
74
17
-
94
73
16
-
94
72
7.2
79
91
69
6.9
76
Ages 40-
64
1 year
99
92
44
71
94
87
42
67
99
93
47
77
91
85
43
70
Ages 40-
64
2 years
98
85
28
65
93
80
26
61
98
87
32
69
89
79
29
63
Ages 40-
64
3 years
97
80
22
63
91
75
20
59
96
82
24
65
87
74
21
58
Ages 40-
64
4 years
96
77
17
61
90
72
16
57
95
78
18
61
85
70
16
54
Ages 40-
64
5 years
96
74
15
60
88
69
14
55
94
75
15
57
83
66
13
50
Ages 40-
64
6 years
95
71
13
57
87
65
12
52
93
72
13
54
81
63
11
48
Ages 40-
64
7 years
94
69
11
55
86
62
10
50
92
69
11
52
79
59
10
45
Ages 40-
64
8 years
93
66
10
51
84
59
8.6
46
91
67
10
49
78
57
8.2
42
Ages 40-
64
9 years
92
64
8.6
51
82
57
7.7
45
90
64
OO
00
49
76
54
7.4
41
Ages 40-
64
10
years
91
63
8.1
50
80
56
7.2
44
88
61
7.9
46
73
51
6.5
38
Final PFAS Rule Economic Analysis
H-19
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 65-
74
1 year
98
89
38
65
91
83
35
60
98
91
42
65
87
81
37
58
Ages 65-
74
2 years
97
82
24
58
88
75
22
52
97
85
26
59
84
73
23
51
Ages 65-
74
3 years
96
77
18
50
86
69
16
45
96
79
20
52
81
67
17
44
Ages 65-
74
4 years
95
75
14
46
83
65
13
40
94
75
15
47
78
61
13
38
Ages 65-
74
5 years
93
70
11
44
79
60
9.4
38
93
71
13
44
74
57
10
35
Ages 65-
74
6 years
91
67
9.3
42
76
56
7.7
35
91
69
11
43
70
53
8.2
33
Ages 65-
74
7 years
89
64
7.9
39
72
52
6.4
32
89
67
9.2
39
66
49
6.7
29
Ages 65-
74
8 years
87
61
7.2
36
68
48
5.6
28
87
65
8.5
38
62
46
6
27
Ages 65-
74
9 years
85
57
6.4
34
65
43
4.9
26
86
63
7.9
35
58
43
5.4
24
Ages 65-
74
10
years
82
54
6
33
60
39
4.4
24
85
61
6.9
31
55
39
4.4
20
Ages 75+
1 year
92
79
21
47
47
40
11
24
94
83
28
52
47
41
14
26
Ages 75+
2 years
92
72
12
37
46
36
5.9
18
94
77
17
45
44
37
8.2
21
Ages 75+
3 years
90
67
9
31
44
32
4.3
15
93
74
12
38
42
34
5.4
17
Ages 75+
4 years
89
63
6.9
28
42
29
3.2
13
91
71
9
32
39
31
3.9
14
Ages 75+
5 years
87
59
5.2
24
39
27
2.3
11
89
69
7.3
27
37
29
3
11
Final PFAS Rule Economic Analysis
H-20
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Race/Ethnicity
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 75+
6 years
85
56
4.2
23
37
24
1.8
10
89
64
6.4
23
35
25
2.5
8.9
Ages 75+
7 years
84
54
4.1
22
35
22
1.7
9.1
86
61
6.1
21
32
22
2.2
7.7
Ages 75+
8 years
82
52
4.1
19
32
21
1.6
7.6
84
58
5.9
20
28
20
2
7
Ages 75+
9 years
77
49
3.1
17
29
18
1.2
6.4
83
56
4.6
17
26
18
1.4
5.5
Ages 75+
10
years
75
44
2.9
17
26
15
1
6
82
55
3.8
16
24
16
1.1
4.7
Non-Hispanic Black
Ages <15
1 year
99
99
92
-
97
97
91
-
99
96
81
-
97
95
80
-
Ages <15
2 years
99
96
88
-
97
95
87
-
99
94
69
-
97
93
68
-
Ages <15
3 years
97
91
86
-
96
90
85
-
99
94
64
-
97
93
63
-
Ages <15
4 years
95
89
81
-
94
88
80
-
99
94
64
-
97
93
63
-
Ages <15
5 years
91
89
78
-
90
88
77
-
99
92
64
-
97
90
63
-
Ages <15
6 years
91
89
78
-
90
88
77
-
97
92
64
-
95
90
62
-
Ages <15
7 years
91
89
78
-
90
88
77
-
97
92
64
-
95
90
62
-
Ages <15
8 years
91
89
78
-
90
88
77
-
97
92
59
-
95
90
58
-
Ages <15
9 years
91
89
78
-
90
88
77
-
97
92
59
-
95
90
58
-
Ages <15
10
years
91
89
78
-
90
88
77
-
97
92
59
-
94
90
58
-
Ages 15-
39
1 year
98
83
34
-
97
81
34
-
96
86
29
-
93
84
28
-
Ages 15-
39
2 years
98
77
20
-
96
76
20
-
95
70
15
-
92
67
15
-
Final PFAS Rule Economic Analysis
H-21
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 15-
39
3 years
96
74
16
-
95
73
16
-
93
57
12
-
90
55
12
-
Ages 15-
39
4 years
95
70
14
-
94
69
14
-
92
51
9.4
-
89
49
9
-
Ages 15-
39
5 years
95
70
10
-
93
69
10
-
91
47
7.8
-
88
45
7.5
-
Ages 15-
39
6 years
94
70
10
-
93
69
10
-
90
41
5.9
-
86
40
5.6
-
Ages 15-
39
7 years
93
70
10
-
91
69
10
-
89
41
5.9
-
85
39
5.6
-
Ages 15-
39
8 years
92
70
10
-
90
69
10
-
89
41
5.9
-
84
39
5.6
-
Ages 15-
39
9 years
92
70
10
-
90
68
10
-
87
37
5.9
-
82
35
5.6
-
Ages 15-
39
10
years
90
70
10
-
88
68
10
-
87
37
5.9
-
82
35
5.6
-
Ages 40-
64
1 year
98
87
33
71
91
81
31
66
98
83
33
79
86
73
29
69
Ages 40-
64
2 years
96
78
23
64
88
71
21
58
96
77
19
67
84
67
17
58
Ages 40-
64
3 years
95
72
16
59
86
66
14
53
95
70
13
62
81
60
11
53
Ages 40-
64
4 years
93
68
12
53
84
62
11
47
93
66
8.6
57
79
56
7.3
48
Ages 40-
64
5 years
92
66
11
50
82
59
9.4
45
92
64
6.8
56
76
53
5.7
47
Ages 40-
64
6 years
91
62
9.5
48
80
55
8.4
42
90
60
6.2
54
74
50
5.1
44
Final PFAS Rule Economic Analysis
H-22
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 40-
64
7 years
90
59
9.5
48
78
52
8.3
42
89
57
6
53
71
46
4.8
43
Ages 40-
64
8 years
89
57
9
48
77
49
7.8
41
87
51
5.7
52
69
40
4.5
41
Ages 40-
64
9 years
87
54
9
41
74
46
7.7
35
86
49
4.9
48
67
38
3.8
37
Ages 40-
64
10
years
87
52
9
41
73
44
7.6
35
84
48
4.9
43
64
37
3.7
32
Ages 65-
74
1 year
96
80
34
70
87
72
31
63
97
82
32
78
82
69
27
66
Ages 65-
74
2 years
95
74
21
58
83
65
19
51
95
76
20
70
78
62
16
57
Ages 65-
74
3 years
92
66
14
54
79
57
12
46
94
68
13
59
74
54
10
47
Ages 65-
74
4 years
90
58
10
52
75
49
8.7
43
92
64
10
56
69
48
7.7
42
Ages 65-
74
5 years
88
57
8.2
52
72
46
6.6
42
92
59
7.9
51
66
43
5.7
37
Ages 65-
74
6 years
86
56
7.4
52
68
44
5.9
41
91
59
6.2
37
63
41
4.3
25
Ages 65-
74
7 years
84
56
5.6
52
64
43
4.3
39
90
57
5.9
31
59
38
3.9
21
Ages 65-
74
8 years
83
56
5.6
35
61
41
4.2
26
88
52
5.2
28
55
32
3.3
18
Ages 65-
74
9 years
80
50
5.6
27
57
35
4
19
87
48
3.7
28
51
28
2.2
17
Ages 65-
74
10
years
80
47
5.6
27
54
32
3.8
18
85
48
2
28
47
27
1.1
16
Final PFAS Rule Economic Analysis
H-23
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Race/Ethnicity
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 75+
1 year
90
65
20
58
47
34
11
30
87
73
27
49
43
36
13
24
Ages 75+
2 years
88
60
14
41
44
30
6.9
20
87
59
19
43
41
27
00
00
20
Ages 75+
3 years
85
53
10
32
41
26
4.8
15
87
54
10
37
38
24
4.5
17
Ages 75+
4 years
83
47
8.9
29
39
22
4.2
14
82
48
8.4
27
34
20
3.5
11
Ages 75+
5 years
80
43
8.3
24
36
19
3.7
11
80
41
6.4
27
32
16
2.5
11
Ages 75+
6 years
75
36
8.3
21
32
16
3.6
9
78
40
6.4
24
29
15
2.4
00
00
Ages 75+
7 years
69
35
8.3
19
28
14
3.4
7.9
73
38
5.2
14
25
13
1.8
4.7
Ages 75+
8 years
64
35
8.3
19
25
13
3.2
7.5
71
38
3.7
14
22
12
1.2
4.3
Ages 75+
9 years
61
31
8.3
19
22
11
3
7.1
70
38
3.7
-
20
11
1.1
-
Ages 75+
10
years
60
30
4.8
19
20
10
1.6
6.7
70
36
3.7
-
19
9.4
1
-
Hispanic
Ages <15
1 year
98
99
90
-
98
99
89
-
100
100
85
-
99
99
84
-
Ages <15
2 years
98
97
79
-
98
97
78
-
98
98
69
-
98
98
68
-
Ages <15
3 years
98
97
77
-
98
97
77
-
98
98
67
-
97
98
66
-
Ages <15
4 years
98
96
74
-
98
95
73
-
96
98
60
-
96
98
60
-
Ages <15
5 years
98
96
74
-
98
95
73
-
95
98
58
-
94
98
57
-
Ages <15
6 years
97
96
72
-
96
95
71
-
93
98
58
-
93
98
57
-
Ages <15
7 years
97
94
72
-
96
93
71
-
93
98
58
-
93
98
57
-
Ages <15
8 years
97
94
72
-
96
93
71
-
93
98
58
-
92
98
57
-
Ages <15
9 years
97
94
72
-
96
93
71
-
93
95
58
-
92
94
57
-
Final PFAS Rule Economic Analysis
H-24
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages <15
10
years
97
94
72
-
96
93
71
-
93
95
58
-
92
94
57
-
Ages 15-
39
1 year
99
89
53
-
99
88
53
-
99
93
49
-
98
92
48
-
Ages 15-
39
2 years
99
79
34
-
98
78
33
-
99
86
35
-
98
85
34
-
Ages 15-
39
3 years
98
72
23
-
97
71
23
-
99
74
25
-
98
73
25
-
Ages 15-
39
4 years
98
66
23
-
97
65
23
-
99
73
20
-
97
72
20
-
Ages 15-
39
5 years
98
66
14
-
97
65
14
-
98
71
19
-
96
70
18
-
Ages 15-
39
6 years
97
66
11
-
96
65
11
-
97
70
15
-
95
68
14
-
Ages 15-
39
7 years
96
66
11
-
95
65
11
-
96
70
15
-
94
68
14
-
Ages 15-
39
8 years
96
66
11
-
95
65
11
-
96
64
15
-
94
62
14
-
Ages 15-
39
9 years
96
66
11
-
95
65
11
-
96
60
15
-
93
58
14
-
Ages 15-
39
10
years
96
66
11
-
95
65
11
-
96
60
15
-
93
58
14
-
Ages 40-
64
1 year
99
91
43
79
95
87
42
76
98
92
46
77
92
86
43
72
Ages 40-
64
2 years
98
86
29
75
94
82
28
72
96
86
31
66
90
80
29
61
Ages 40-
64
3 years
97
82
21
70
93
78
20
67
95
82
24
60
88
76
22
56
Final PFAS Rule Economic Analysis
H-25
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 40-
64
4 years
96
80
18
65
91
76
17
62
93
78
19
56
85
72
17
51
Ages 40-
64
5 years
94
78
16
63
89
74
15
60
92
73
16
52
83
67
14
47
Ages 40-
64
6 years
94
75
13
58
88
71
12
55
89
71
13
50
81
64
12
45
Ages 40-
64
7 years
92
70
12
58
87
66
11
54
88
67
11
45
79
60
10
40
Ages 40-
64
8 years
91
68
11
58
85
64
10
54
86
65
10
44
76
57
8.5
39
Ages 40-
64
9 years
90
66
10
54
83
61
9.1
50
86
61
8.9
42
75
53
7.8
37
Ages 40-
64
10
years
89
66
8
54
81
60
7.4
50
83
59
8.3
42
72
51
7.1
37
Ages 65-
74
1 year
98
90
37
62
93
85
35
59
97
92
40
66
88
84
37
60
Ages 65-
74
2 years
97
86
22
53
90
81
21
50
95
85
25
55
85
76
23
49
Ages 65-
74
3 years
95
77
18
53
88
71
16
49
93
78
18
51
81
68
16
45
Ages 65-
74
4 years
94
75
13
49
85
68
12
44
92
71
15
45
79
61
13
38
Ages 65-
74
5 years
93
74
11
44
83
66
10
39
90
65
12
39
75
54
10
32
Ages 65-
74
6 years
91
73
10
44
80
64
9.1
38
88
62
11
32
71
50
8.5
26
Ages 65-
74
7 years
89
69
10
44
76
59
8.2
38
87
59
10
26
68
46
7.8
20
Final PFAS Rule Economic Analysis
H-26
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Race/Ethnicity
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 65-
74
8 years
89
67
10
44
74
56
8
37
84
56
10
26
64
42
7.6
20
Ages 65-
74
9 years
87
64
10
35
71
52
7.8
28
83
54
10
25
61
40
7.3
18
Ages 65-
74
10
years
87
58
7.6
27
69
46
6.1
21
80
45
8.9
25
56
32
6.2
17
Ages 75+
1 year
93
78
25
45
50
42
13
24
93
86
28
43
47
44
14
22
Ages 75+
2 years
90
72
13
35
48
38
6.8
18
91
78
19
32
45
39
9.3
16
Ages 75+
3 years
89
67
7.8
31
46
35
4
16
89
73
15
27
42
35
7.4
13
Ages 75+
4 years
85
60
5.8
25
43
30
2.9
13
86
67
13
20
40
31
6.1
9.3
Ages 75+
5 years
82
56
4.5
21
41
27
2.2
10
83
61
10
16
37
27
4.4
7
Ages 75+
6 years
79
55
3.6
20
38
26
1.7
9.5
82
56
7.3
14
35
24
3.1
6
Ages 75+
7 years
74
47
3.6
13
34
22
1.7
6.1
80
52
6.1
14
32
21
2.5
5.7
Ages 75+
8 years
68
44
3.6
11
31
20
1.6
5.1
75
52
5
10
29
20
1.9
3.7
Ages 75+
9 years
65
40
2.2
10
28
17
1
4.2
73
47
5
10
26
17
1.8
3.5
Ages 75+
10
years
63
33
2.2
5.2
26
14
0.9
2.1
68
43
0
10
23
14
0
3.2
Other
Ages <15
1 year
99
99
92
100
99
98
91
99
99
99
88
-
99
98
88
-
Ages <15
2 years
98
97
86
100
98
97
85
99
99
96
79
-
98
95
78
-
Ages <15
3 years
98
95
83
96
97
94
82
96
97
95
76
-
96
95
75
-
Ages <15
4 years
97
94
81
92
97
93
81
92
97
95
74
-
96
94
73
-
Ages <15
5 years
97
93
80
92
96
93
79
92
97
94
73
-
96
93
72
-
Final PFAS Rule Economic Analysis
H-27
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages <15
6 years
96
93
79
92
95
93
79
92
96
94
72
-
95
93
71
-
Ages <15
7 years
95
93
79
87
95
92
79
86
96
94
71
-
95
93
70
-
Ages <15
8 years
95
93
78
87
95
92
78
86
96
94
70
-
95
93
69
-
Ages <15
9 years
95
93
78
87
95
92
78
86
96
92
69
-
95
91
68
-
Ages <15
10
years
95
93
78
87
95
92
78
86
96
92
69
-
95
90
68
-
Ages 15-
39
1 year
99
93
50
90
99
92
49
89
99
92
42
91
97
90
41
89
Ages 15-
39
2 years
99
85
32
83
98
84
31
82
99
85
27
84
97
83
26
83
Ages 15-
39
3 years
98
80
24
77
97
79
24
76
98
78
20
83
96
76
19
81
Ages 15-
39
4 years
98
75
21
77
97
74
21
76
98
74
15
83
95
72
14
81
Ages 15-
39
5 years
97
73
16
77
96
72
16
76
97
71
12
79
94
69
12
77
Ages 15-
39
6 years
97
72
15
77
96
71
15
76
96
69
10
72
93
67
10
70
Ages 15-
39
7 years
97
71
14
77
95
70
14
76
95
68
8.9
69
92
65
8.7
67
Ages 15-
39
8 years
96
70
13
77
95
69
13
76
95
66
7.7
66
92
64
7.4
64
Ages 15-
39
9 years
96
69
13
77
94
68
12
76
94
65
7.7
66
91
62
7.4
63
Ages 15-
39
10
years
95
69
13
77
93
68
12
76
94
65
7.7
66
90
62
7.4
63
Final PFAS Rule Economic Analysis
H-28
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 40-
64
1 year
99
91
43
73
94
87
40
70
99
92
46
78
90
84
42
71
Ages 40-
64
2 years
98
85
28
67
92
80
26
63
97
86
31
69
89
78
28
63
Ages 40-
64
3 years
97
80
21
64
91
75
19
60
96
81
23
64
87
73
20
58
Ages 40-
64
4 years
96
77
17
61
89
72
15
57
95
77
18
61
85
69
16
54
Ages 40-
64
5 years
95
74
14
60
88
69
13
55
94
74
14
58
83
65
13
51
Ages 40-
64
6 years
94
71
12
56
87
66
11
52
92
71
12
55
81
62
11
48
Ages 40-
64
7 years
93
69
11
55
85
63
10
50
91
68
11
52
79
58
9.2
45
Ages 40-
64
8 years
92
66
10
52
83
60
8.7
47
90
65
9.3
50
77
55
7.9
43
Ages 40-
64
9 years
91
64
8.6
50
82
57
7.7
45
89
63
8.6
48
75
53
7.2
40
Ages 40-
64
10
years
90
63
8.1
50
80
56
7.2
44
87
60
7.7
45
72
50
6.4
38
Ages 65-
74
1 year
98
89
38
66
90
82
35
61
98
90
41
67
87
80
37
60
Ages 65-
74
2 years
97
82
24
58
88
75
22
53
97
84
26
60
84
73
23
52
Ages 65-
74
3 years
95
76
17
53
85
68
16
47
95
78
19
54
80
66
16
45
Ages 65-
74
4 years
94
73
14
49
82
64
12
43
94
74
15
48
77
60
13
39
Final PFAS Rule Economic Analysis
H-29
April 2024
-------
FINAL RULE APRIL 2024
Table H-6: Summary of Race/Ethnicity-Specific Relative and Absolute Kidney Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival
(Average)
by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 65-
74
5 years
92
69
11
47
79
59
9.4
40
92
70
12
44
73
55
10
35
Ages 65-
74
6 years
90
66
10
46
75
55
8
38
91
67
10
42
69
52
8
32
Ages 65-
74
7 years
88
63
8.1
44
72
51
6.6
36
89
65
9
37
65
48
6.6
27
Ages 65-
74
8 years
87
61
7.7
39
68
48
6
31
87
63
8.5
37
61
44
6
26
Ages 65-
74
9 years
85
57
7
35
65
43
5.3
27
86
61
7.8
34
58
41
5.3
23
Ages 65-
74
10
years
83
53
6.5
34
60
39
4.7
25
85
57
6.8
32
54
37
4.4
20
Ages 75+
1 year
92
78
22
49
47
40
11
25
94
83
28
52
46
41
14
26
Ages 75+
2 years
91
71
12
38
46
35
6.2
19
93
77
17
45
44
37
8.3
21
Ages 75+
3 years
89
66
9.4
32
43
32
4.6
16
92
74
12
38
42
34
5.5
17
Ages 75+
4 years
88
61
7.4
29
41
29
3.5
13
89
70
9.2
32
39
31
4
14
Ages 75+
5 years
86
57
5.9
25
39
26
2.7
11
88
67
7.2
27
36
28
3
11
Ages 75+
6 years
84
54
5
24
36
24
2.2
10
87
62
6.3
23
34
24
2.5
8.9
Ages 75+
7 years
81
51
4.8
22
34
21
2
9
85
60
5.8
20
31
22
2.1
7.5
Ages 75+
8 years
78
50
4.7
19
31
20
1.9
7.7
82
57
5.4
19
28
20
1.9
6.6
Ages 75+
9 years
74
47
3.6
18
28
18
1.4
6.6
81
55
4.3
17
26
17
1.4
5.4
Ages 75+
10
years
72
42
3.2
18
25
15
1.1
6.2
79
52
3.7
16
23
15
1.1
4.6
Final PFAS Rule Economic Analysis
H-30
April 2024
-------
FINAL RULE APRIL 2024
Table H-7: Summary of All-Cause and Kidney Cancer Mortality Data Used in the Model
Females
Males
Age
Rate per 100K
Percent Kidney Cancer
Rate per 100K
Percent Kidney Cancer
All-Cause
Kidney Cancer
All-Cause
Kidney Cancer
<1
537
0.04
0.007
646
0.045
0.007
1-4
22
0.070
0.31
28
0.079
0.28
5-9
11
0.084
0.80
13
0.065
0.50
10-14
12
0.043
0.36
17
0.036
0.21
15-19
29
0.042
0.15
68
0.042
0.062
20-24
46
0.063
0.14
129
0.099
0.077
25-29
61
0.075
0.12
150
0.14
0.093
30-34
82
0.13
0.16
169
0.20
0.12
35-39
111
0.23
0.21
199
0.49
0.25
40-44
159
0.40
0.25
259
1.1
0.43
45-49
246
0.91
0.37
390
2.5
0.65
50-54
376
1.8
0.47
609
4.9
0.80
55-59
545
3.1
0.57
916
8.5
0.92
60-64
785
4.7
0.60
1304
13
0.98
65-69
1166
7.1
0.61
1829
18
1.00
70-74
1844
10
0.56
2720
24
0.89
75-79
3027
14
0.47
4280
32
0.74
80-84
5193
19
0.37
7039
41
0.58
85+
-
-
0.21
-
-
0.37
Final PFAS Rule Economic Analysis
H-31
April 2024
-------
FINAL RULE APRIL 2024
Table H-8: Summary of Race/Ethnicity-Specific All-Cause and Kidney Cancer Mortality Data Used in the Model
Race/Ethnicity
Age
Females
Males
Rate per 100K
Percent Kidney Cancer
Rate per 100K
Percent Kidney Cancer
All-Cause
Kidney Cancer
All-Cause
Kidney Cancer
Non-Hispanic White
<1
453
0.0091
0.0020
554
0.043
0.0078
1-4
20
0.060
0.73
26
0.080
0.30
5-9
10
0.072
3.3
12
0.069
0.57
10-14
12
0.044
2.7
16
0.025
0.15
15-19
30
0.039
1.1
63
0.019
0.031
20-24
48
0.043
0.53
124
0.040
0.032
25-29
66
0.055
0.48
153
0.093
0.061
30-34
89
0.098
0.60
177
0.15
0.087
35-39
120
0.21
0.99
209
0.45
0.22
40-44
168
0.38
1.3
269
1.1
0.42
45-49
254
0.93
2.3
401
2.7
0.66
50-54
380
1.9
3.0
616
5.1
0.84
55-59
544
3.2
3.4
909
00
00
0.97
60-64
779
4.9
3.6
1282
13
1.0
65-69
1172
7.4
3.5
1810
19
1.0
70-74
1881
11
3.0
2732
25
0.92
75-79
3108
15
2.6
4347
33
0.76
80-84
5351
20
2.1
7225
42
0.59
85+
-
-
2.6
-
-
0.36
Non-Hispanic Black
<1
1042
0.031
0.0029
1249
0.029
0.0024
1-4
36
0.11
0.77
45
0.10
0.23
5-9
16
0.15
4.3
20
0.076
0.37
Final PFAS Rule Economic Analysis
H-32
April 2024
-------
FINAL RULE APRIL 2024
Table H-8: Summary of Race/Ethnicity-Specific All-Cause and Kidney Cancer Mortality Data Used in the Model
Race/Ethnicity
Age
Females
Males
Rate per 100K
Percent Kidney Cancer
Rate per 100K
Percent Kidney Cancer
All-Cause
Kidney Cancer
All-Cause
Kidney Cancer
10-14
17
0.053
1.8
25
0.098
0.39
15-19
34
0.095
2.1
111
0.14
0.12
20-24
63
0.18
1.7
202
0.38
0.19
25-29
86
0.21
1.4
232
0.40
0.17
30-34
121
0.31
1.5
262
0.55
0.21
35-39
173
0.39
1.3
312
0.96
0.31
40-44
249
0.44
1.0
397
1.5
0.38
45-49
377
1.1
1.8
572
2.9
0.51
50-54
579
1.8
1.9
892
5.0
0.56
55-59
844
3.2
2.1
1398
8.6
0.61
60-64
1193
4.7
2.1
2052
14
0.66
65-69
1656
7.3
2.2
2791
19
0.68
70-74
2399
9.5
2.0
3820
24
0.63
75-79
3616
13
1.9
5464
31
0.57
80-84
5700
18
1.6
8058
37
0.45
85+
-
-
2.4
-
-
0.36
Hispanic
<1
435
0.055
0.013
513
0.070
0.014
1-4
19
0.063
0.82
23
0.056
0.25
5-9
9
0.080
3.6
11
0.053
0.49
10-14
11
0.042
2.3
14
0.018
0.13
15-19
23
0.016
0.49
58
0.034
0.058
20-24
34
0.041
0.66
106
0.082
0.078
25-29
39
0.038
0.50
111
0.11
0.10
Final PFAS Rule Economic Analysis
H-33
April 2024
-------
FINAL RULE APRIL 2024
Table H-8: Summary of Race/Ethnicity-Specific All-Cause and Kidney Cancer Mortality Data Used in the Model
Race/Ethnicity
Age
Females
Males
Rate per 100K
Percent Kidney Cancer
Rate per 100K
Percent Kidney Cancer
All-Cause
Kidney Cancer
All-Cause
Kidney Cancer
30-34
50
0.092
0.98
117
0.15
0.13
35-39
65
0.23
2.0
137
0.40
0.29
40-44
95
0.47
2.8
180
0.97
0.54
45-49
149
0.80
3.0
275
2.3
0.84
50-54
232
1.6
3.8
438
4.2
0.95
55-59
355
3.1
4.7
665
7.4
1.1
60-64
550
4.8
4.7
982
12
1.2
65-69
840
6.9
4.3
1402
17
1.2
70-74
1328
10
4.1
2113
23
1.1
75-79
2251
14
3.3
3343
30
0.90
80-84
3960
19
2.6
5411
34
0.63
85+
-
-
2.7
-
-
0.44
Other
<1
409
0.22
0.053
498
0.000
0.000
1-4
19
0.070
0.97
24
0.10
0.43
5-9
10
0.041
1.7
12
0.053
0.46
10-14
11
0.014
0.79
13
0.054
0.40
15-19
23
0.027
0.90
48
0.052
0.11
20-24
33
0.024
0.43
80
0.034
0.043
25-29
37
0.051
0.74
85
0.095
0.11
30-34
47
0.13
1.4
93
0.076
0.082
35-39
62
0.15
1.4
113
0.26
0.23
40-44
88
0.30
2.0
154
0.63
0.41
45-49
139
0.55
2.3
234
1.4
0.61
Final PFAS Rule Economic Analysis
H-34
April 2024
-------
FINAL RULE APRIL 2024
Table H-8: Summary of Race/Ethnicity-Specific All-Cause and Kidney Cancer Mortality Data Used in the Model
Females
Males
Race/Ethnicity
Age
Rate per 100K
Rate per 100K
All-Cause
Kidney Cancer
Percent Kidney Cancer
All-Cause
Kidney Cancer
Percent Kidney Cancer
50-54
210
0.96
2.6
354
2.9
0.82
55-59
298
1.7
3.2
527
5.1
0.97
60-64
438
2.2
2.7
754
8.1
1.1
65-69
661
3.6
2.9
1081
9.6
0.89
70-74
1066
6.2
3.0
1623
13
0.83
75-79
1849
8.0
2.3
2661
17
0.64
80-84
3363
12
1.9
4522
23
0.51
85+
-
-
2.1
-
-
0.35
Final PFAS Rule Economic Analysis
H-35
April 2024
-------
FINAL RULE
APRIL 2024
H.4 Baseline Bladder Cancer Statistics
Table H-9 provides baseline bladder cancer incidence data used in the life table model. Bladder
cancer incidence rates per 100,000 range from 0.17 to 76 for females and from 0.11 to 357 for
males. Bladder cancer incidence rates are highest for men in their 60s, 70s, and 80s, ranging
from 67 per 100,000 to 357 per 100,000. Localized bladder cancers comprise 66%-90% of all
bladder cancer incidence, whereas regional bladder cancers comprise 4.5%-8.6%, distant bladder
cancers comprise 3.1%-14%, and unstaged bladder cancers comprise 0%-6.8% of all bladder
cancer incidence.
Table H-9: Summary of Baseline Bladder Cancer Incidence Data Used in the Model
Females
Males
Percent of Incidence in Stage
Percent of Incidence in Stage
Age
Incidence
Localized
13
Unstaged
Incidence
Localized
13
Unstaged
per 100K
e
"Si
"S
5
per 100K
e
"Si
"S
5
<1
_
77
4.5
14
4.5
_
66
23
11
0
1-4
_
77
4.5
14
4.5
_
66
23
11
0
5-9
_
77
4.5
14
4.5
_
66
23
11
0
10-14
_
77
4.5
14
4.5
_
66
23
11
0
15-19
_
82
8.2
5.1
4.9
0.11
90
4.8
3.1
2.5
20-24
0.17
82
8.2
5.1
4.9
0.30
90
4.8
3.1
2.5
25-29
0.26
82
8.2
5.1
4.9
0.51
90
4.8
3.1
2.5
30-34
0.50
82
8.2
5.1
4.9
1.1
90
4.8
3.1
2.5
35-39
0.89
82
8.2
5.1
4.9
2.1
90
4.8
3.1
2.5
40-44
1.5
83
8.6
6.1
2.7
4.2
85
7.4
4.9
2.5
45-49
2.9
83
8.6
6.1
2.7
00
00
85
7.4
4.9
2.5
50-54
6.6
83
8.6
6.1
2.7
19
85
7.4
4.9
2.5
55-59
11
83
8.6
6.1
2.7
38
85
7.4
4.9
2.5
60-64
18
83
8.6
6.1
2.7
67
85
7.4
4.9
2.5
65-69
29
84
7.9
5.6
2.8
114
86
6.7
4.3
2.9
70-74
43
84
7.9
5.6
2.8
176
86
6.7
4.3
2.9
75-79
58
80
7.1
5.8
6.8
245
85
6.2
4.1
5.2
80-84
71
80
7.1
5.8
6.8
315
85
6.2
4.1
5.2
85+
76
80
7.1
5.8
6.8
357
85
6.2
4.1
5.2
Final PFAS Rule Economic Analysis
H-36
April 2024
-------
FINAL RULE
APRIL 2024
Table H-10 shows relative bladder cancer survival rates54 by sex, age group at diagnosis, cancer
stage, and the number of years post diagnosis. The relative bladder cancer survival ranges from
0% to 100%, and generally decreases as the number of years post-diagnosis increases. The table
also shows the absolute survival probability, averaged over the age range for which the relative
survival data were available; these probabilities are a product of general population survival
probability and the relative bladder cancer survival probability by sex, age group at diagnosis,
and the number of years post-diagnosis. The life table model uses derived absolute survival
probabilities to model all-cause mortality experience in bladder cancer populations for the
baseline scenario and the regulatory alternative. Finally, Table H-l 1 shows all-cause and bladder
cancer mortality rates used in the life table model. Bladder cancer deaths <1% of all-cause
mortality among females and <2% of all-cause mortality among males.
54 Relative bladder cancer survival rate is the probability of being alive K years after diagnosis at age A divided by the general
probability to survive K years for a person alive at age A without such a diagnosis.
Final PFAS Rule Economic Analysis
April 2024
H-37
-------
FINAL RULE APRIL 2024
Table H-10: Summary of Relative and Absolute Bladder Cancer Survival Used in the Model
0>
Females
Males
Age at Diagnos
S
H
s.
p
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
1
!£
_o
"o
it
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages
15-39
1
year
98
79
20
90
97
79
20
90
99
85
46
100
97
83
45
98
Ages
15-39
2
years
97
58
4
83
96
57
4
83
99
67
23
97
96
65
22
95
Ages
15-39
3
years
96
47
0
80
95
46
0
79
98
60
14
95
96
58
13
92
Ages
15-39
4
years
95
39
0
80
94
39
0
79
97
58
11
91
95
56
11
89
Ages
15-39
5
years
95
32
0
80
93
32
0
79
96
56
11
91
94
54
11
89
Ages
15-39
6
years
94
28
0
80
93
27
0
79
96
56
9
91
93
54
9
89
Ages
15-39
7
years
94
28
0
80
92
27
0
79
96
56
7
91
93
54
7
88
Ages
15-39
8
years
93
28
0
80
92
27
0
78
95
56
7
91
92
54
7
88
Ages
15-39
9
years
93
28
0
80
91
27
0
78
94
52
5
91
91
51
4
88
Ages
15-39
10
years
93
28
0
80
91
27
0
78
93
52
5
85
90
50
4
82
Ages
40-64
1
year
97
73
34
84
92
69
32
80
98
78
36
85
90
72
33
78
Ages
40-64
2
years
95
53
15
81
90
50
14
76
96
57
16
79
87
52
15
72
Ages
40-64
3
years
94
45
9
77
88
42
9
72
94
48
11
75
85
43
10
67
Ages
40-64
4
years
93
40
7
76
87
37
7
70
93
43
9
73
83
38
8
65
Final PFAS Rule Economic Analysis
H-38
April 2024
-------
FINAL RULE APRIL 2024
Table H-10: Summary of Relative and Absolute Bladder Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages
40-64
5
years
92
37
5
74
85
34
5
69
91
40
8
71
81
35
7
63
Ages
40-64
6
years
91
36
5
74
84
33
5
68
90
38
7
68
79
33
7
60
Ages
40-64
7
years
90
34
4
73
82
31
4
66
89
37
7
66
77
32
6
57
Ages
40-64
8
years
89
32
4
71
80
29
4
64
88
36
7
64
75
30
6
54
Ages
40-64
9
years
88
31
4
70
79
28
3
63
87
35
7
61
73
29
6
51
Ages
40-64
10
years
87
31
4
70
77
27
3
62
86
34
7
61
71
28
6
51
Ages
65-74
1
year
95
67
25
72
88
62
24
66
97
74
32
81
86
66
29
72
Ages
65-74
2
years
92
48
11
67
83
44
10
61
94
55
16
75
82
48
13
65
Ages
65-74
3
years
90
38
8
63
80
34
7
57
92
47
11
72
77
39
9
60
Ages
65-74
4
years
88
34
6
60
77
30
5
52
89
42
8
69
73
34
6
56
Ages
65-74
5
years
86
31
5
58
73
26
5
50
88
39
6
66
70
31
5
52
Ages
65-74
6
years
85
28
5
56
71
23
4
47
86
36
6
64
66
27
4
49
Ages
65-74
7
years
84
27
4
54
68
22
3
44
84
34
5
61
62
25
4
45
Ages
65-74
8
years
82
25
4
52
64
20
3
41
82
32
5
57
58
23
4
40
Final PFAS Rule Economic Analysis
H-39
April 2024
-------
FINAL RULE APRIL 2024
Table H-10: Summary of Relative and Absolute Bladder Cancer Survival Used in the Model
0>
Females
Males
Age at Diagnos
S
H
s.
p
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
Relative Survival by Stage
(Percent)
Absolute Survival (Average) by
Stage (Percent)
1
!£
_o
"o
it
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages
65-74
9
years
81
25
3
51
61
19
2
39
80
30
4
56
54
20
3
38
Ages
65-74
10
years
79
25
3
51
58
18
2
37
79
29
4
56
50
19
3
36
Ages
75+
1
year
86
48
17
39
44
25
9
20
92
60
22
59
45
30
11
29
Ages
75+
2
years
81
36
8
32
40
18
4
16
87
44
10
51
42
21
5
24
Ages
75+
3
years
77
30
6
27
38
15
3
13
84
38
7
45
38
17
3
21
Ages
75+
4
years
76
28
5
24
36
13
2
11
81
35
5
40
35
15
2
17
Ages
75+
5
years
73
26
4
22
33
12
2
10
79
33
5
37
33
14
2
15
Ages
75+
6
years
71
24
4
22
31
11
2
9
76
32
4
34
30
13
2
13
Ages
75+
7
years
69
22
3
20
29
9
1
8
74
29
3
31
27
11
1
11
Ages
75+
8
years
68
21
3
18
27
8
1
7
72
28
3
29
25
10
1
10
Ages
75+
9
years
66
21
2
18
25
8
1
7
70
28
3
26
22
9
1
8
Ages
75+
10
years
65
18
2
18
23
6
1
6
68
28
3
23
20
8
1
7
Final PFAS Rule Economic Analysis
H-40
April 2024
-------
FINAL RULE APRIL 2024
Table H-ll: Summary of All-Cause and Bladder Cancer Mortality Data Used in the Model
Females
Males
Age
Rate per 100K
Percent Bladder Cancer
Rate per 100K
Percent Bladder Cancer
All-Cause
Bladder Cancer
All-Cause
Bladder Cancer
<1
537
0
0
646
0.0090
0.0014
1-4
22
0.002
0.010
28
0.0011
0.0040
5-9
11
0.002
0.017
13
0.0009
0.0068
10-14
12
0.004
0.030
17
0.0034
0.0202
15-19
29
0.002
0.006
68
0.0033
0.0049
20-24
46
0.008
0.016
129
0.016
0.012
25-29
61
0.035
0.057
150
0.029
0.019
30-34
82
0.067
0.082
169
0.10
0.060
35-39
111
0.22
0.19
199
0.28
0.14
40-44
159
0.47
0.30
259
0.77
0.30
45-49
246
0.92
0.37
390
2.0
0.52
50-54
376
1.6
0.43
609
4.4
0.72
55-59
545
2.8
0.51
916
00
00
0.96
60-64
785
4.7
0.60
1304
16
1.2
65-69
1166
8.0
0.69
1829
27
1.5
70-74
1844
15
0.82
2720
49
1.8
75-79
3027
27
0.88
4280
88
2.1
80-84
5193
43
0.83
7039
146
2.1
85+
-
-
0.54
-
-
1.6
Final PFAS Rule Economic Analysis
H-41
April 2024
-------
FINAL RULE
APRIL 2024
H.5 Baseline Liver Cancer Statistics
Table H-12 provides baseline liver cancer incidence data used in the life table model. Liver
cancer incidence rates per 100,000 range from 0.089 to 32 for females and from 0.10 to 72 for
males. Liver cancer incidence rates are highest for men in their 60s, 70s, and 80s, ranging from
58 per 100,000 to 72 per 100,000. Localized liver cancers comprise 35%-44% of all liver cancer
incidence, whereas regional liver cancers comprise 20%-28%, distant liver cancers comprise
17%-29%, and unstaged liver cancers comprise 4.1%-26% of all liver cancer incidence.
Table H-12: Summary of Baseline Liver Cancer Incidence Data Used in the Model
Females
Males
Percent of Incidence in Stage
Percent of Incidence in
Stage
Age
Incidence
per 100K
Localized
Regional
Distant
Unstaged
Incidence per
100K
Localized
Regional
Distant
Unstaged
<1
1.3
50
24
18
7.1
1.8
46
28
22
4.1
1-4
0.52
50
24
18
7.1
0.77
46
28
22
4.1
5-9
0.12
50
24
18
7.1
0.13
46
28
22
4.1
10-14
0.089
50
24
18
7.1
0.10
46
28
22
4.1
15-19
0.16
39
23
29
00
00
0.16
35
27
28
10
20-24
0.19
39
23
29
00
00
0.19
35
27
28
10
25-29
0.27
39
23
29
00
00
0.41
35
27
28
10
30-34
0.44
39
23
29
00
00
0.74
35
27
28
10
35-39
0.80
39
23
29
00
00
1.3
35
27
28
10
40-44
1.4
43
24
21
12
2.4
41
27
18
14
45-49
2.2
43
24
21
12
5.9
41
27
18
14
50-54
4.6
43
24
21
12
16
41
27
18
14
55-59
9.5
43
24
21
12
36
41
27
18
14
60-64
16
43
24
21
12
58
41
27
18
14
65-69
23
44
23
21
13
72
41
27
18
14
70-74
24
44
23
21
13
64
41
27
18
14
75-79
30
36
20
18
26
63
38
24
17
21
80-84
32
36
20
18
26
66
38
24
17
21
85+
29
36
20
18
26
56
38
24
17
21
Final PFAS Rule Economic Analysis
H-42
April 2024
-------
FINAL RULE
APRIL 2024
Table H-13: Summary of Race/Ethnicity-Specific Baseline Liver Cancer Incidence Data
Used in the Model
Age
__l
1-4
5-9
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85+
__l
1-4
5-9
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80-84
85+
Females
Incidence
per 100K
1.3
0.4
0.2
0.2
0.3
0.4
0.7
1.3
1.8
3.8
7.7
12.2
16.8
18.4
23.4
25.1
23.3
0.7
0.9
1.1
2.5
4.9
11.6
23.8
30.3
22.9
22
25.9
25.7
Percent of Incidence in Stage
.N
13
O
O
-J
52
52
52
52
39
39
39
39
39
41
41
41
41
41
40
40
33
33
33
38
38
38
38
42
42
42
42
42
41
41
41
41
41
42
42
32
32
32
a
s
e
"Si
(2
23
23
23
23
26
26
26
26
26
25
25
25
25
25
24
24
20
20
20
31
31
31
31
21
21
21
21
21
26
26
26
26
26
24
24
23
23
23
s
a
18
18
18
18
28
28
28
28
28
22
22
22
22
22
23
23
19
19
19
25
25
25
25
27
27
27
27
27
19
19
19
19
19
21
21
21
21
21
a>
ex
cs
%
s
P
12
12
12
12
12
12
12
27
27
27
10
10
10
10
10
13
13
13
13
13
12
12
24
24
24
Males
Incidence
per 100K
1.7
0.7
0.2
0.2
0.2
0.3
0.6
0.8
1.6
3.8
10.6
27.5
46
56.8
52.2
53.6
56.6
47.8
0.6
0.6
1.2
1.9
2.3
12.8
41.8
86.8
118
85.6
57.6
48
38.7
Percent of Incidence in
Stage
.N
13
O
O
-J
45
45
45
45
37
37
37
37
37
41
41
41
41
41
40
40
38
38
38
47
47
47
47
27
27
27
27
27
38
38
38
38
38
40
40
35
35
35
CS
S
e
"Si
(2
28
28
28
28
24
24
24
24
24
27
27
27
27
27
27
27
24
24
24
27
27
27
27
30
30
30
30
30
29
29
29
29
29
28
28
22
22
22
Final PFAS Rule Economic Analysis
H-43
April 2024
-------
FINAL RULE
APRIL 2024
Table H-13: Summary of Race/Ethnicity-Specific Baseline Liver Cancer Incidence Data
Used in the Model
Race/Ethnicity
Age
Females
Males
Incidence
per 100K
Percent of Incidence in Stage
Incidence
per 100K
Percent of Incidence in
Stage
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Hispanic
<1
1.3
50
24
18
8
1.5
46
30
21
3
1-4
0.7
50
24
18
8
1
46
30
21
3
5-9
0
50
24
18
8
0
46
30
21
3
10-14
0
50
24
18
8
0
46
30
21
3
15-19
0
37
20
33
10
0
37
24
26
13
20-24
0
37
20
33
10
0.2
37
24
26
13
25-29
0.2
37
20
33
10
0.4
37
24
26
13
30-34
0.3
37
20
33
10
0.6
37
24
26
13
35-39
0.7
37
20
33
10
1.1
37
24
26
13
40-44
1.2
46
21
19
14
2.5
43
25
16
15
45-49
2.5
46
21
19
14
8.7
43
25
16
15
50-54
6.4
46
21
19
14
26.9
43
25
16
15
55-59
14.1
46
21
19
14
54.7
43
25
16
15
60-64
25.1
46
21
19
14
81.4
43
25
16
15
65-69
40.2
48
22
16
14
105
42
26
16
16
70-74
46.2
48
22
16
14
102
42
26
16
16
75-79
59.4
40
19
15
26
105.7
38
23
16
23
80-84
64.4
40
19
15
26
106
38
23
16
23
85+
58.1
40
19
15
26
97.7
38
23
16
23
Other
<1
1.3
50
24
18
7
1.8
46
28
22
4
1-4
0.5
50
24
18
7
0.8
46
28
22
4
5-9
0.1
50
24
18
7
0.1
46
28
22
4
10-14
0.1
50
24
18
7
0.1
46
28
22
4
15-19
0.2
39
23
29
9
0.2
35
27
28
10
20-24
0.2
39
23
29
9
0.2
35
27
28
10
25-29
0.3
39
23
29
9
0.4
35
27
28
10
30-34
0.4
39
23
29
9
0.7
35
27
28
10
35-39
0.8
39
23
29
9
1.3
35
27
28
10
40-44
1.4
43
24
21
12
2.4
41
27
18
14
45-49
2.2
43
24
21
12
5.9
41
27
18
14
50-54
4.6
43
24
21
12
15.6
41
27
18
14
55-59
9.5
43
24
21
12
35.6
41
27
18
14
60-64
16.1
43
24
21
12
57.8
41
27
18
14
65-69
22.5
44
23
21
13
71.7
41
27
18
14
70-74
23.9
44
23
21
13
63.8
41
27
18
14
75-79
29.6
36
20
18
26
63.2
38
24
17
21
80-84
32.3
36
20
18
26
66
38
24
17
21
85+
29.3
36
20
18
26
56.3
38
24
17
21
Final PFAS Rule Economic Analysis
H-44
April 2024
-------
FINAL RULE
APRIL 2024
Table H-14 shows relative liver cancer survival rates55 by sex, age group at diagnosis, cancer
stage, and the number of years post diagnosis. The relative liver cancer survival ranges from 0%
to 94%, and generally decreases as the number of years post-diagnosis increases. The table also
shows the absolute survival probability, averaged over the age range for which the relative
survival data were available; these probabilities are a product of general population survival
probability and the relative liver cancer survival probability by sex, age group at diagnosis, and
the number of years post-diagnosis. The life table model uses derived absolute survival
probabilities to model all-cause mortality experience in liver cancer populations for the baseline
scenario and the regulatory alternative. Finally, Table H-16 shows all-cause and liver cancer
mortality rates used in the life table model. Liver cancer deaths <1% of all-cause mortality
among females and among males.
Table H-14: Summary of Relative Liver Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages <15
1 year
92
88
71
83
94
89
77
74
Ages <15
2 years
89
85
60
76
91
85
63
64
Ages <15
3 years
88
84
54
68
89
81
59
64
Ages <15
4 years
87
83
48
68
87
79
57
64
Ages <15
5 years
86
81
48
63
87
78
56
64
Ages <15
6 years
86
80
48
63
86
78
55
64
Ages <15
7 years
86
79
48
63
86
78
53
60
Ages <15
8 years
85
76
48
63
86
78
53
60
Ages <15
9 years
84
75
48
63
86
78
52
60
Ages <15
10 years
84
75
48
63
86
78
52
60
Ages 15-
39
1 year
87
64
44
71
78
46
32
55
Ages 15-
39
2 years
77
50
23
65
69
32
19
45
Ages 15-
39
3 years
72
42
15
63
61
28
13
39
Ages 15-
39
4 years
67
37
13
57
59
23
10
37
Ages 15-
39
5 years
65
34
11
54
55
22
9
34
Ages 15-
39
6 years
63
31
11
51
54
19
8
33
Ages 15-
39
7 years
60
29
11
48
51
18
7
33
55 Relative liver cancer survival rate is the probability of being alive K years after diagnosis at age A divided by the general
probability to survive K years for a person alive at age A without such a diagnosis.
Final PFAS Rule Economic Analysis
H-45
April 2024
-------
FINAL RULE
APRIL 2024
Table H-14: Summary of Relative Liver Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 15-
39
8 years
58
28
11
48
50
17
6
32
Ages 15-
39
9 years
57
28
11
48
49
17
6
32
Ages 15-
39
10 years
57
27
11
48
46
17
5
31
Ages 40-
64
1 year
75
49
25
44
70
40
16
35
Ages 40-
64
2 years
61
31
12
33
55
25
7
23
Ages 40-
64
3 years
52
23
8
28
46
18
4
17
Ages 40-
64
4 years
47
20
6
23
40
15
3
14
Ages 40-
64
5 years
43
17
5
21
36
13
2
11
Ages 40-
64
6 years
40
16
4
19
33
11
2
10
Ages 40-
64
7 years
38
15
4
18
31
10
2
9
Ages 40-
64
8 years
36
14
3
17
29
10
2
8
Ages 40-
64
9 years
35
13
3
16
28
9
2
8
Ages 40-
64
10 years
34
13
3
15
27
9
1
7
Ages 65-
74
1 year
70
43
21
35
69
41
17
32
Ages 65-
74
2 years
54
25
9
23
54
26
7
20
Ages 65-
74
3 years
45
18
6
17
45
18
4
14
Ages 65-
74
4 years
39
14
4
14
38
14
3
11
Ages 65-
74
5 years
34
12
3
12
33
11
2
9
Ages 65-
74
6 years
30
10
3
10
29
10
2
7
Ages 65-
74
7 years
28
10
2
8
26
8
2
6
Ages 65-
74
8 years
25
9
2
7
24
8
1
6
Ages 65-
74
9 years
23
8
2
7
23
7
1
5
Final PFAS Rule Economic Analysis
H-46
April 2024
-------
FINAL RULE
APRIL 2024
Table H-14: Summary of Relative Liver Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage
(Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 65-
74
10 years
22
7
2
6
21
6
1
5
Ages 75+
1 year
52
28
14
22
55
30
12
22
Ages 75+
2 years
37
15
6
12
40
17
5
13
Ages 75+
3 years
28
10
3
8
29
11
3
8
Ages 75+
4 years
23
7
2
5
23
8
2
5
Ages 75+
5 years
19
6
2
4
19
6
1
4
Ages 75+
6 years
16
4
1
3
15
5
1
3
Ages 75+
7 years
14
4
1
2
13
3
1
2
Ages 75+
8 years
12
3
1
2
11
3
1
2
Ages 75+
9 years
11
2
1
2
9
2
1
2
Ages 75+
10 years
10
2
1
2
8
2
0
2
Final PFAS Rule Economic Analysis
H-47
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
Race/Ethnicity
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Non-Hispanic White
Ages <15
1 year
91
85
69
100
94
90
75
100
Ages <15
2 years
85
84
61
100
89
87
59
100
Ages <15
3 years
85
84
59
100
89
83
53
100
Ages <15
4 years
85
84
54
100
88
82
49
100
Ages <15
5 years
84
79
54
100
87
82
49
100
Ages <15
6 years
84
79
54
100
87
80
49
100
Ages <15
7 years
84
77
54
100
87
80
49
100
Ages <15
8 years
84
77
54
100
87
80
49
100
Ages <15
9 years
82
73
54
100
87
80
49
100
Ages <15
10 years
82
73
54
100
87
80
49
100
Ages 15-39
1 year
89
71
47
67
80
68
44
67
Ages 15-39
2 years
80
58
23
60
71
46
24
52
Ages 15-39
3 years
76
49
19
60
66
41
18
48
Ages 15-39
4 years
70
42
15
56
63
33
13
44
Ages 15-39
5 years
67
38
14
53
63
31
12
42
Ages 15-39
6 years
63
32
14
51
61
27
11
42
Ages 15-39
7 years
62
32
12
48
60
26
10
42
Ages 15-39
8 years
61
32
12
48
60
26
7
39
Ages 15-39
9 years
58
32
12
48
59
26
7
39
Ages 15-39
10 years
58
30
12
48
55
26
7
36
Ages 40-64
1 year
73
51
26
45
69
41
17
34
Final PFAS Rule Economic Analysis
H-48
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
0>
£
Females
Males
s
ex
a
H
s.
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
a
"ea
0>
ex
<
p
1
!£
o
"o
li
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 40-64
2 years
59
33
13
33
55
25
7
23
Ages 40-64
3 years
51
24
8
27
46
18
4
17
Ages 40-64
4 years
46
21
6
23
40
15
3
14
Ages 40-64
5 years
42
19
5
20
36
13
2
12
Ages 40-64
6 years
40
17
4
19
33
12
2
11
Ages 40-64
7 years
37
15
4
18
31
11
1
10
Ages 40-64
8 years
36
14
3
17
29
10
1
9
Ages 40-64
9 years
34
14
3
16
28
9
1
8
Ages 40-64
10 years
33
14
2
15
27
9
1
7
Ages 65-74
1 year
68
41
23
30
68
41
18
31
Ages 65-74
2 years
54
24
10
21
53
25
7
19
Ages 65-74
3 years
44
18
5
15
43
18
4
13
Ages 65-74
4 years
39
14
3
12
37
13
3
10
Ages 65-74
5 years
35
12
2
11
32
10
2
8
Ages 65-74
6 years
31
10
1
10
28
9
2
7
Ages 65-74
7 years
30
10
1
9
25
8
2
6
Ages 65-74
8 years
28
9
1
8
23
7
1
5
Ages 65-74
9 years
26
9
1
7
21
7
1
5
Ages 65-74
10 years
24
8
1
6
20
6
1
5
Ages 75+
1 year
48
27
13
18
54
29
13
19
Ages 75+
2 years
35
14
6
9
38
16
5
11
Final PFAS Rule Economic Analysis
H-49
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
%
0>
£
Females
Males
"8
s
ex
a
H
s.
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
"45
si
&
a
"ea
ex
p
1
!£
0
"o
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 75+
3 years
25
9
3
6
28
10
3
7
Ages 75+
4 years
21
6
2
4
22
7
2
4
Ages 75+
5 years
18
5
2
2
18
4
1
3
Ages 75+
6 years
16
4
1
2
15
3
1
3
Ages 75+
7 years
14
3
1
1
12
3
1
2
Ages 75+
8 years
12
2
1
1
11
3
1
2
Ages 75+
9 years
10
2
1
1
8
2
1
2
Ages 75+
10 years
9
2
1
1
6
1
0
2
Ages <15
1 year
96
100
100
100
88
85
100
100
Ages <15
2 years
96
100
100
100
88
81
100
100
Ages <15
3 years
96
100
100
100
88
73
100
100
Ages <15
4 years
91
100
100
100
79
73
100
100
si
Ages <15
5 years
91
100
100
100
79
73
100
100
CO
Ages <15
6 years
91
100
100
100
79
73
100
100
C
si
O.
Ages <15
7 years
91
100
100
100
79
73
100
100
3
Ages <15
8 years
84
100
100
100
79
73
100
100
s
0
z
Ages <15
9 years
84
100
100
100
79
73
100
100
Ages <15
10 years
84
100
100
100
79
73
100
100
Ages 15-39
1 year
76
57
43
100
73
29
19
45
Ages 15-39
2 years
68
40
29
100
64
24
12
42
Ages 15-39
3 years
60
38
14
100
54
18
8
33
Final PFAS Rule Economic Analysis
H-50
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 15-39
4 years
57
35
14
100
50
13
5
33
Ages 15-39
5 years
54
31
14
100
45
11
5
33
Ages 15-39
6 years
54
31
14
100
44
7
5
33
Ages 15-39
7 years
52
27
14
100
40
6
5
33
Ages 15-39
8 years
50
27
14
100
38
6
5
33
Ages 15-39
9 years
50
27
14
100
38
6
5
33
Ages 15-39
10 years
50
27
14
100
36
6
100
33
Ages 40-64
1 year
72
43
20
43
65
33
15
31
Ages 40-64
2 years
58
27
11
29
50
19
7
19
Ages 40-64
3 years
49
20
7
23
39
13
3
13
Ages 40-64
4 years
43
16
4
20
33
11
2
10
Ages 40-64
5 years
38
14
3
19
28
9
2
9
Ages 40-64
6 years
35
13
2
17
26
8
2
7
Ages 40-64
7 years
34
12
2
15
23
7
2
6
Ages 40-64
8 years
32
11
2
15
22
6
2
5
Ages 40-64
9 years
31
10
2
15
19
5
2
4
Ages 40-64
10 years
30
10
2
15
18
5
2
4
Ages 65-74
1 year
71
39
18
36
67
36
14
29
Ages 65-74
2 years
57
24
7
25
52
24
7
18
Ages 65-74
3 years
51
18
4
18
42
16
4
12
Ages 65-74
4 years
42
13
4
16
36
14
2
9
Final PFAS Rule Economic Analysis
H-51
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
Race/Ethnicity
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 65-74
5 years
36
11
4
14
32
11
1
7
Ages 65-74
6 years
31
10
3
10
28
10
1
7
Ages 65-74
7 years
26
10
3
7
24
8
1
4
Ages 65-74
8 years
21
7
3
7
22
6
1
4
Ages 65-74
9 years
19
7
3
5
21
6
0
4
Ages 65-74
10 years
19
4
3
5
18
5
0
4
Ages 75+
1 year
50
19
15
22
51
21
9
24
Ages 75+
2 years
34
10
7
16
36
11
4
12
Ages 75+
3 years
25
9
4
10
27
8
4
9
Ages 75+
4 years
25
6
1
5
19
7
3
6
Ages 75+
5 years
19
4
0
5
14
7
2
3
Ages 75+
6 years
16
4
0
3
13
7
2
2
Ages 75+
7 years
14
2
0
3
11
5
100
0
Ages 75+
8 years
14
1
0
3
10
5
100
0
Ages 75+
9 years
14
100
0
3
8
100
100
0
Ages 75+
10 years
13
100
0
3
8
100
100
0
Hispanic
Ages <15
1 year
93
93
70
100
93
87
82
100
Ages <15
2 years
92
86
60
100
92
83
68
100
Ages <15
3 years
88
86
52
100
89
78
66
100
Ages <15
4 years
88
84
49
100
87
75
63
100
Ages <15
5 years
87
82
49
100
87
74
61
100
Final PFAS Rule Economic Analysis
H-52
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages <15
6 years
87
82
49
100
86
74
61
100
Ages <15
7 years
85
82
49
100
86
74
59
100
Ages <15
8 years
85
75
49
100
86
74
59
100
Ages <15
9 years
85
75
49
100
86
74
56
100
Ages <15
10 years
85
75
49
100
86
74
56
100
Ages 15-39
1 year
88
64
38
80
79
49
43
55
Ages 15-39
2 years
74
49
20
72
71
37
29
42
Ages 15-39
3 years
69
41
13
62
63
29
18
36
Ages 15-39
4 years
62
37
11
57
59
28
15
33
Ages 15-39
5 years
60
34
9
57
53
26
13
28
Ages 15-39
6 years
60
32
9
57
53
24
11
28
Ages 15-39
7 years
50
32
9
57
51
24
11
28
Ages 15-39
8 years
50
25
9
57
49
24
11
28
Ages 15-39
9 years
50
25
9
57
48
24
11
28
Ages 15-39
10 years
50
25
9
57
46
24
6
28
Ages 40-64
1 year
75
48
25
42
68
43
16
36
Ages 40-64
2 years
59
30
13
33
53
26
7
23
Ages 40-64
3 years
51
22
8
27
43
18
4
16
Ages 40-64
4 years
45
19
6
22
37
14
3
12
Ages 40-64
5 years
41
16
5
19
32
12
3
10
Ages 40-64
6 years
37
15
5
18
29
11
3
9
Final PFAS Rule Economic Analysis
H-53
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
0>
£
Females
Males
s
ex
a
H
s.
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
a
"ea
0>
ex
<
p
1
!£
o
"o
li
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 40-64
7 years
35
15
4
18
27
10
2
8
Ages 40-64
8 years
33
13
4
17
26
9
2
8
Ages 40-64
9 years
32
13
4
15
24
9
2
7
Ages 40-64
10 years
31
13
4
13
23
8
2
7
Ages 65-74
1 year
67
43
20
39
66
41
17
32
Ages 65-74
2 years
49
24
11
22
49
25
8
20
Ages 65-74
3 years
38
16
8
17
39
17
5
14
Ages 65-74
4 years
31
12
6
14
32
11
3
11
Ages 65-74
5 years
26
10
5
11
27
10
2
9
Ages 65-74
6 years
23
8
5
7
23
9
1
6
Ages 65-74
7 years
21
7
5
6
21
8
1
5
Ages 65-74
8 years
19
6
5
6
18
7
1
4
Ages 65-74
9 years
17
6
2
6
17
7
1
4
Ages 65-74
10 years
16
6
2
6
14
6
100
4
Ages 75+
1 year
52
28
14
27
51
30
11
24
Ages 75+
2 years
36
15
5
15
34
16
4
14
Ages 75+
3 years
27
9
3
10
24
10
3
9
Ages 75+
4 years
20
6
2
7
16
8
2
6
Ages 75+
5 years
15
4
2
5
13
6
1
4
Ages 75+
6 years
13
3
1
3
11
4
1
3
Ages 75+
7 years
11
1
1
3
9
2
1
3
Final PFAS Rule Economic Analysis
H-54
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
Race/Ethnicity
Age at Diagnosis
Follow-Up Time
Females
Males
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 75+
8 years
9
100
100
3
8
2
1
3
Ages 75+
9 years
8
100
100
3
6
2
0
3
Ages 75+
10 years
7
100
100
3
4
2
0
3
Other
Ages <15
1 year
88
71
83
88
94
89
77
74
Ages <15
2 years
85
60
76
85
91
85
63
64
Ages <15
3 years
84
54
68
84
89
81
59
64
Ages <15
4 years
83
48
68
83
87
79
57
64
Ages <15
5 years
81
48
63
81
87
78
56
64
Ages <15
6 years
80
48
63
80
86
78
55
64
Ages <15
7 years
79
48
63
79
86
78
53
60
Ages <15
8 years
76
48
63
76
86
78
53
60
Ages <15
9 years
75
48
63
75
86
78
52
60
Ages <15
10 years
75
48
63
75
86
78
52
60
Ages 15-39
1 year
64
44
71
64
78
46
32
55
Ages 15-39
2 years
50
23
65
50
69
32
19
45
Ages 15-39
3 years
42
15
63
42
61
28
13
39
Ages 15-39
4 years
37
13
57
37
59
23
10
37
Ages 15-39
5 years
34
11
54
34
55
22
9
34
Ages 15-39
6 years
31
11
51
31
54
19
8
33
Ages 15-39
7 years
29
11
48
29
51
18
7
33
Ages 15-39
8 years
28
11
48
28
50
17
6
32
Final PFAS Rule Economic Analysis
H-55
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
0>
£
Females
Males
s
ex
a
H
s.
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
a
"ea
0>
ex
<
p
1
!£
o
"o
it
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 15-39
9 years
28
11
48
28
49
17
6
32
Ages 15-39
10 years
27
11
48
27
46
17
5
31
Ages 40-64
1 year
49
25
44
49
70
40
16
35
Ages 40-64
2 years
31
12
33
31
55
25
7
23
Ages 40-64
3 years
23
8
28
23
46
18
4
17
Ages 40-64
4 years
20
6
23
20
40
15
3
14
Ages 40-64
5 years
17
5
21
17
36
13
2
11
Ages 40-64
6 years
16
4
19
16
33
11
2
10
Ages 40-64
7 years
15
4
18
15
31
10
2
9
Ages 40-64
8 years
14
3
17
14
29
10
2
8
Ages 40-64
9 years
13
3
16
13
28
9
2
8
Ages 40-64
10 years
13
3
15
13
27
9
1
7
Ages 65-74
1 year
43
21
35
43
69
41
17
32
Ages 65-74
2 years
25
9
23
25
54
26
7
20
Ages 65-74
3 years
18
6
17
18
45
18
4
14
Ages 65-74
4 years
14
4
14
14
38
14
3
11
Ages 65-74
5 years
12
3
12
12
33
11
2
9
Ages 65-74
6 years
10
3
10
10
29
10
2
7
Ages 65-74
7 years
10
2
8
10
26
8
2
6
Ages 65-74
8 years
9
2
7
9
24
8
1
6
Ages 65-74
9 years
8
2
7
8
23
7
1
5
Final PFAS Rule Economic Analysis
H-56
April 2024
-------
FINAL RULE APRIL 2024
Table H-15: Summary of Race/Ethnicity-Specific Relative Liver Cancer Survival Used in the Model
&
8
Females
Males
"8
a
©X
es
H
O.
Relative Survival by Stage (Percent)
Relative Survival by Stage (Percent)
"45
si
&
Q
"S
©X
<
P
1
!£
0
"o
ti
Localized
Regional
Distant
Unstaged
Localized
Regional
Distant
Unstaged
Ages 65-74
10 years
7
2
6
7
21
6
1
5
Ages 75+
1 year
28
14
22
28
55
30
12
22
Ages 75+
2 years
15
6
12
15
40
17
5
13
Ages 75+
3 years
10
3
8
10
29
11
3
8
Ages 75+
4 years
7
2
5
7
23
8
2
5
Ages 75+
5 years
6
2
4
6
19
6
1
4
Ages 75+
6 years
4
1
3
4
15
5
1
3
Ages 75+
7 years
4
1
2
4
13
3
1
2
Ages 75+
8 years
3
1
2
3
11
3
1
2
Ages 75+
9 years
2
1
2
2
9
2
1
2
Ages 75+
10 years
2
1
2
2
8
2
0
2
Final PFAS Rule Economic Analysis
H-57
April 2024
-------
FINAL RULE APRIL 2024
Table H-16: Summary of All-Cause and Liver Cancer Mortality Data Used in the Model
Females
Males
Age
Rate per 100K
Percent Liver Cancer
Rate per 100K
Percent Liver Cancer
All-Cause
Liver Cancer
All-Cause
Liver Cancer
<1
579
0.071
0.012
702
0.06
0.009
1-4
25
0.066
0.270
31
0.12
0.39
5-9
12
0.000
0.000
14
0.027
0.19
10-14
13
0.005
0.040
19
0.010
0.05
15-19
33
0.025
0.08
78
0.038
0.049
20-24
47
0.053
0.11
136
0.08
0.06
25-29
60
0.10
0.17
148
0.18
0.12
30-34
80
0.19
0.24
165
0.35
0.21
35-39
113
0.34
0.30
204
0.70
0.34
40-44
168
0.72
0.43
281
1.6
0.56
45-49
253
1.5
0.59
419
4.6
1.1
50-54
378
3.1
0.81
631
11
1.7
55-59
558
5.5
1.0
933
20
2.2
60-64
833
8.6
1.0
1361
29
2.1
65-69
1256
12
1.0
1963
33
1.7
70-74
1996
17
0.83
2977
38
1.3
75-79
3270
23
0.70
4704
45
1.0
80-84
5550
28
0.50
7623
52
0.69
85+
-
-
0.24
-
-
0.35
Final PFAS Rule Economic Analysis
H-58
April 2024
-------
FINAL RULE APRIL 2024
Table H-17: Summary of Race/Ethnicity-Specific All-Cause and Liver Cancer Mortality Data Used in the Model
Race/Ethnicity
Age
Females
Males
Rate per 100K
Percent Liver Cancer
Rate per 100K
Percent Liver Cancer
All-Cause
Liver Cancer
All-Cause
Liver Cancer
Non-Hispanic
White
<1
486
0.083
0.017
600
0.12
0.019
1-4
22
0.088
0.40
28
0.15
0.51
5-9
11
0
0
13
0.038
0.29
10-14
13
0.009
0.074
19
0.018
0.096
15-19
35
0.042
0.12
73
0.065
0.090
20-24
48
0.090
0.19
127
0.089
0.070
25-29
61
0.11
0.18
144
0.15
0.10
30-34
82
0.18
0.22
164
0.23
0.14
35-39
114
0.32
0.28
203
0.48
0.24
40-44
166
0.62
0.38
279
1.2
0.41
45-49
249
1.3
0.51
412
3.7
0.89
50-54
369
2.7
0.73
615
9.2
1.5
55-59
547
4.7
0.86
907
17
1.9
60-64
820
7.4
0.90
1326
24
1.8
65-69
1251
10
0.82
1932
28
1.4
70-74
2015
15
0.72
2972
33
1.1
75-79
3322
20
0.60
4747
41
0.87
80-84
5670
25
0.44
7774
48
0.62
85+
-
-
0.21
-
-
0.32
Non-Hispanic
Black
<1
1148
0.17
0.015
1386
0
0
1-4
39
0
0
49
0
0
5-9
17
0
0
22
0
0
Final PFAS Rule Economic Analysis
H-59
April 2024
-------
FINAL RULE APRIL 2024
Table H-17: Summary of Race/Ethnicity-Specific All-Cause and Liver Cancer Mortality Data Used in the Model
Race/Ethnicity
Age
Females
Males
Rate per 100K
Percent Liver Cancer
Rate per 100K
Percent Liver Cancer
All-Cause
Liver Cancer
All-Cause
Liver Cancer
10-14
18
0
0
28
0
0
15-19
38
0
0
121
0
0
20-24
67
0
0
221
0.13
0.059
25-29
93
0.16
0.18
250
0.38
0.15
30-34
131
0.32
0.25
278
0.78
0.28
35-39
192
0.49
0.25
339
1.3
0.39
40-44
288
1.0
0.36
455
2.6
0.57
45-49
427
2.3
0.54
673
6.9
1.0
50-54
625
4.8
0.77
1015
17
1.7
55-59
894
00
00
0.98
1513
35
2.3
60-64
1280
13
1.0
2185
53
2.4
65-69
1815
16
0.89
3012
58
1.9
70-74
2650
19
0.72
4212
52
1.2
75-79
4007
24
0.60
6073
48
0.79
80-84
6198
29
0.47
8873
54
0.61
85+
-
-
0.27
-
-
0.35
Hispanic
<1
469
0
0
556
0
0
1-4
21
0.076
0.36
26
0.16
0.62
5-9
10
0
0
12
0.024
0.20
10-14
12
0
0
16
0
0
15-19
26
0
0
70
0
0
20-24
35
0
0
117
0.024
0.021
25-29
40
0.028
0.070
116
0.080
0.068
Final PFAS Rule Economic Analysis
H-60
April 2024
-------
FINAL RULE APRIL 2024
Table H-17: Summary of Race/Ethnicity-Specific All-Cause and Liver Cancer Mortality Data Used in the Model
Race/Ethnicity
Age
Females
Males
Rate per 100K
Percent Liver Cancer
Rate per 100K
Percent Liver Cancer
All-Cause
Liver Cancer
All-Cause
Liver Cancer
30-34
50
0.13
0.26
123
0.22
0.18
35-39
70
0.29
0.42
151
0.46
0.30
40-44
103
0.68
0.66
207
1.6
0.77
45-49
160
1.7
1.0
311
5.4
1.7
50-54
247
3.3
1.3
476
14
2.9
55-59
380
7.0
1.8
713
27
3.8
60-64
595
12
2.0
1059
39
3.7
65-69
922
19
2.1
1546
49
3.2
70-74
1468
28
1.9
2356
57
2.4
75-79
2463
41
1.7
3702
71
1.9
80-84
4241
48
1.1
5873
79
1.3
85+
-
-
0.55
-
-
0.74
Other
<1
419
0
0
510
0
0
1-4
21
0
0
26
0
0
5-9
11
0
0
12
0
0
10-14
12
0
0
15
0
0
15-19
27
0
0
55
0
0
20-24
33
0
0
83
0
0
25-29
36
0.056
0.15
83
0.39
0.47
30-34
47
0.21
0.44
92
0.95
1.0
35-39
64
0.39
0.61
118
2.2
1.9
40-44
93
1.0
1.1
164
4.1
2.5
45-49
145
2.0
1.4
246
8.9
3.6
Final PFAS Rule Economic Analysis
H-61
April 2024
-------
FINAL RULE APRIL 2024
Table H-17: Summary of Race/Ethnicity-Specific All-Cause and Liver Cancer Mortality Data Used in the Model
Females
Males
Race/Ethnicity
Age
Rate per 100K
Rate per 100K
All-Cause
Liver Cancer
Percent Liver Cancer
All-Cause
Liver Cancer
Percent Liver Cancer
50-54
216
3.3
1.5
366
16
4.4
55-59
314
6.9
2.2
545
27
5.0
60-64
474
12
2.5
797
37
4.7
65-69
727
18
2.5
1169
47
4.0
70-74
1178
28
2.3
1785
61
3.4
75-79
1999
42
2.1
2933
75
2.6
80-84
3573
57
1.6
4885
95
1.9
85+
-
-
0.71
-
-
0.93
Final PFAS Rule Economic Analysis
H-62
April 2024
-------
FINAL RULE
APRIL 2024
H.6RCC Valuation Data
The EPA identified the study selected for use in evaluating potential medical costs avoided as a
result of the final PFAS rule and regulatory alternatives, Ambavane et al. (2020), as part of a
targeted kidney cancer valuation literature search. The scope of the search covered cost of illness
(COI) and willingness to pay literature published in English language peer reviewed sources
during 2010-2021.56 The searches were executed in the Google Scholar article database. The
EPA reviewed 153 references retrieved by the willingness to pay-oriented searches and the top
348 references retrieved by the COI-oriented searches.57
The search did not identify any suitable kidney cancer willingness to pay studies. However, there
were seven additional studies containing COI information. Of those, four were cost-effectiveness
studies that focused only on medication costs. The remaining three studies focused on the overall
medical care costs but had methodological issues that prevented the EPA from using them as the
basis for kidney cancer morbidity valuation:
• Hollenbeak et al. (2011) reported 5-year RCC cost estimates based on Medicare data from
early 2000s; however, even after adjusting for medical care price inflation, these RCC cost
estimates were too low relative to the costs reported by more recent cost-effectiveness
studies.
• Bhattachaijee et al. (2017) annual cost estimates were based on the Medical Expenditure
Panel Survey 2002-2011 data for persons experiencing kidney cancer but included
expenditures for conditions other than kidney cancer.
• Mitchell et al. (2020) reported Medicare costs for various first line kidney cancer treatment
types, but not the frequency and duration with which these treatments were typically applied.
Detailed notes on the 8 studies reviewed by the EPA are provided in Table H-12.
50 The query terms used for willingness to pay-oriented and COI-oriented searches are available upon request.
57The EPA applied exclusion-term based automated screening to the raw Google Scholar result sets; exclusion terms are available
upon request. Hie number of references listed in this document reflect the size of the result sets after the automated screening was
applied. There were 153 references in the willingness to pay-oriented search result set and 1,342 references in the COI-oriented
search result set. The EPA reviewed all 153 references in the willingness to pay-oriented results set and top 348 references in the
COI-oriented results set. The references in the COI-oriented results set were prioritized using Okapi BM25 metric applied to
article titles and Google Scholar ranks.
Final PFAS Rule Economic Analysis
H-63
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
Ambavane et
al. (2020)
Lifetime
treatment costs
of several
treatment
sequences
(first and
second line
drug costs,
administration
costs, disease
management,
and adverse
effects
management)
Incidence-
based
Accounting for
first and second
line, drug costs +
administration
costs + disease
management costs
per month + single
time AE
management cost
(not accounting for
mean AE
disutility/month) =
$189,594.76/month
+ $48,122; annual
cost = $2.3 million
(without including
monthly disutility).
Dollar values
reported in 2018$.
-26% U.S.;
-35%
Canada/Western
Europe/North
Europe; -39%
rest of world
779,
majority
male and
white with
baseline
median age
of 62 years
Cohort data
from the
CheckMate
214 trial
Not stated
Discrete event
simulation model
estimates lifetime
costs and survival
among patients.
Recent US-based
costs; risk data are
bias toward older
white males and 26%
of trial participants
were from U.S.;
provides costs but
not information on
baseline treatment
frequencies.
Hollenbeak
et al. (2011)
Payments
made by
Medicare for
all-cause
medical
treatments
including
inpatient stays,
emergency
room visits,
outpatient
procedures,
office visits,
home health
visits, durable
medical
Prevalence-
based, by
year since
diagnosis
Mean costs per
patient per month
(PPPM) in the first
year were $3,673
for patients with
RCC. PPPM costs
were higher for
RCC patients with
more advanced
stage (i.e., regional
or distant) disease.
Average
cumulative total
costs for RCC
patients were
$33,605 per patient
USA, individual
scale
4,938
patients
with RCC
and 9,876
non-HMO
noncancer
comparison
group. The
sample was
limited to
non-HMO
patients
aged 65
years or
older who
were
Surveillance,
Epidemiology,
and End
Results
Program
(SEER)-
Medicare
database,
which
combines
tumor registry
data from the
National
Cancer
Institutes
(NCI) SEER
1995-2002
Estimated all-cause
health care costs
associated with RCC
using SEER-
Medicare data. Using
the method of Bang
and Tsiatis (2000),
estimated cumulative
costs at 1 and 5 years
by estimating
average costs for
each patient in each
month up to 60
months following
diagnosis. Medicare
population; costs
Final PFAS Rule Economic Analysis
H-64
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
equipment,
and hospice
care, but
excluding
outpatient
prescription
drugs
in the first year
following
diagnosis and
$59,397 per patient
in the first 5 years
following
diagnosis. Costs
available for first
five years and
separated by stage.
diagnosed
with a first
primary
RCC (SEER
site recode
59, kidney
and renal
pelvis)
between
1995 and
2002
program for
patients who
are covered by
Medicare with
their Medicare
billing records
within 5-years of
diagnosis; data from
2005.
Mitchell et
al. (2020)
Medicare costs
for first-line
and
maintenance
treatment
Cost
accounting-
based
First-line
treatments for
kidney cancer
range from
$30,538 to
$31,190, while
maintenance
treatments range
from $7,722 to
$8,997. These
costs represent the
average monthly
cost of treatment.
USA, individual
scale
Not
specified
Medicare costs
for first-line
and
maintenance
treatments for
cancers with
the highest
incidence in
the US that had
published
NCCN
Evidence
Blocks as of
December 31,
2018; costs
based on
Medicare
prices from the
January 2019
Medicare ASP
file
2018
Calculated Medicare
costs for all first-line
and maintenance
treatments for 30
cancers with the
highest incidence in
the US that had
published NCCN
Evidence Blocks as
of December 31,
2018. Categorized
each treatment as
either "time-limited"
or "time-unlimited."
For time-unlimited
treatments (all
kidney cancer
treatments fall into
this category),
calculated the
average monthly cost
of treatment. No
Final PFAS Rule Economic Analysis
H-65
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
information on
treatment duration.
Bhattachaijee
etal. (2017)
Total
healthcare
expenditure,
which includes
inpatient,
outpatient,
emergency
room,
prescription
drugs, home
health agency,
dental care,
vision care,
and other
expenditures.
The study
included
different
sources of
payment such
as direct
payments from
individuals,
private
insurance,
Medicare,
Prevalence-
based
The annual average
total healthcare
expenditures
($15,078 vs.
$8,182; P<.001)
for adults with
kidney cancer were
significantly higher
compared with
propensity-score-
matched adults
with other forms of
cancer. The
average inpatient
($6755 vs. $1959)
and prescription
drug ($3485 vs.
$1570)
expenditures were
significantly higher
for adults with KC
compared with
matched controls.
Dollar values
reported in 2011$.
USA, individual
scale
Adults aged
21 or older
who did not
die during
the calendar
year of
MEPS data
and had
positive
total
healthcare
expenditures
(N = 541 for
time-
unlimited
treatments,
N = 845 for
time-limited
treatments-
analysis
includes
~30 cancer
types).
Cancer
stage not
specified.
Medical
Expenditure
Panel Survey
2002-2011
Used a retrospective,
cross-sectional,
propensity-score-
matched, case-
control study design
using 2002 to 2011
MEPS data to
determine impacts of
health and functional
status and co-
occurring chronic
conditions.
Developed OLS
regressions on log-
transformed
expenditures for total
and subtypes of
health expenditures.
Calculated
percentage change in
expenditure. Very
small sample of -100
persons; non-
incremental annual
average healthcare
expenditures among
Final PFAS Rule Economic Analysis
H-66
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
Medicaid,
Workers'
Compensation,
and
miscellaneous
other sources.
All
expenditures
inflated using
medical CPI.
those with RCC that
could include care
for other health
issues; no stage and
no variation by time
since diagnosis;
focus on those with
positive
expenditures.
Wan et al.
(2019)
Compares
cost-
effectiveness
of kidney
cancer
treatments:
nivolumab
plus
ipilimumab vs
sunitinib
Incidence-
based
Provides total cost
of regimen, other
values reported in
Incremental Cost-
Effectiveness Ratio
/QALY; cost
effectiveness
analysis of two
different
treatments for RCC
USA, individual
scale
1096
patients
with mRCC
from
clinical trial
modeled to
receive the
drug
CheckMate
214, Centers
for
Medicare &
Medicaid
Services
2018
A Markov model
was developed to
compare the lifetime
cost and
effectiveness of
nivolumab plus
ipilimumab vs
sunitinib in the first-
line treatment of
mRCC using
outcomes data from
the CheckMate 214
phase 3 randomized
clinical trial, which
included 1096
patients with mRCC
(median age, 62
years) and compared
nivolumab plus
ipilimumab vs
sunitinib as first-line
treatment of mRCC.
Final PFAS Rule Economic Analysis
H-67
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
In the analysis,
patients were
modeled to receive
sunitinib or
nivolumab plus
ipilimumab for 4
doses followed by
nivolumab
monotherapy,
provides costs of
treatment but does
not provide the
frequency with
which these
treatments are
applied in the general
population.
Reinhorn et
al. (2019)
Compares
cost-
effectiveness
of kidney
cancer
treatments:
nivolumab and
ipilimumab
versus
sunitinib
Incidence-
based
Cost effectiveness
analysis of two
different
treatments for
RCC; study
centered on
specific drug cost
and was limited by
data availability
USA, individual
scale
Markov
model-
simulated
population
with each
model cycle
representing
1 month
over a 10-
year time
horizon
CheckMate
214
2017
A Markov model
was developed to
compare the costs
and effectiveness of
nivolumab and
ipilimumab with
those of sunitinib in
the first-line
treatment of
intermediate- to
poor-risk advanced
RCC. Health
outcomes were
measured in life-
years and quality-
adjusted life-years
(QALYs). Drug costs
were based on
Final PFAS Rule Economic Analysis
H-68
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
Medicare
reimbursement rates
in 2017. Study
extrapolated survival
beyond the trial
closure using
Weibull distribution.
Model robustness
was addressed in
univariable and
probabilistic
sensitivity analyses.
Provides costs of
treatment but does
not provide the
frequency with
which these
treatments are
applied in the general
population
Perrin et al.
(2015)
Compares
cost-
effectiveness
of kidney
cancer
treatments:
everolimus vs
axitinib;
provides costs
per patient
from
simulated data
Incidence-
based
Cost effectiveness
analysis of two
different
treatments for RCC
USA, individual
scale
Simulated
population
of advanced
RCC
patients
MarketScan
Commercial
Claims and
Encounters and
Medicare
Supplemental
database
2004-2011
A Markov model
was developed to
simulate a cohort of
sunitinib-refractory
advanced RCC
patients and estimate
the cost of treating
patients with
everolimus vs
axitinib. The
following health
states were included:
stable disease
without adverse
events (AEs), stable
Final PFAS Rule Economic Analysis
H-69
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
disease with AEs,
disease progression
(PD), and death. The
model included the
following resources:
active treatments,
post-progression
treatments, AEs,
physician and nurse
visits, scans and
tests, and palliative
care. Resource
utilization inputs
were derived from a
US claims database
analysis.
Additionally, a 3%
annual discount rate
was applied to costs,
and the robustness of
the model results was
tested by conducting
sensitivity analyses,
including those on
dosing scheme and
post-progression
treatment costs.
Provides costs of
treatment but does
not provide the
frequency with
which these
treatments are
applied in the general
population.
Final PFAS Rule Economic Analysis
H-70
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
Racsa et al.
(2015)
Compares
cost-
effectiveness
of kidney
cancer
treatments:
two tyrosine
kinase
inhibitors;
provides
original dollar
estimates for
different
medications
Incidence-
based
Cost effectiveness
analysis of two
different
treatments for RCC
USA, individual
scale
1,438 RCC
patients
aged 19 to
89 years,
with
medical and
pharmacy
insurance
through
commercial
or Medicare
plans
Humana
Research
Database
2009-2012
Study used claims
data to conduct an
observational,
retrospective cohort
study of individuals
aged 19 to 89 years,
with commercial or
Medicare insurance,
advanced RCC, and
at least one
pharmacy claim for
sunitinibor
pazopanib between 1
November 2009 and
31 December 2012.
Treatment
characteristics
(treatment
interruption,
adherence, duration,
and discontinuation),
survival, and costs
were measured up to
12 months. Statistical
models were
adjusted for age,
gender, geographic
region, race, and
RxRisk-Vscore.
Provides costs of
treatment but does
not provide the
frequency with
which these
treatments are
Final PFAS Rule Economic Analysis
H-71
April 2024
-------
FINAL RULE APRIL 2024
Table H-18: Studies Reviewed Related to Kidney Cancer Medical Treatment Costs
Study
Reference
Valuation
Target
Focus
Result(s) Type &
Quality
Geographic
Scope & Scale
Population
Datasets
Data
Collection
Year
Methodology and
Other Notes
applied in the general
population; addresses
a younger
population.
Abbreviations: AE - adverse event; CPI - consumer price index; HMO - Health Maintenance Organization; MEPS - Medical Expenditure Panel Survey; mRCC metastatic
renal cell carcinoma; KC - kidney cancer; NCCN - National Comprehensive Cancer Network; NCI- National Cancer Institute; OLS - ordinary least squares; PD - disease
progression; PPPM - per patient per month; QALYs - quality adjusted life years; RCC - renal cell carcinoma; SEER - Surveillance, Epidemiology, and End Results Program.
Final PFAS Rule Economic Analysis
H-72
April 2024
-------
FINAL RULE
APRIL 2024
Appendix I. Trihalomethane Co-Removal Model
Details and Analysis
1.1 Data Analysis
The EPA analyzed Information Collection Rule Treatment Study Database (ICR TSD) data to
predict time-based removal efficacy of total organic carbon (TOC) and four regulated
trihalomethanes (THM4) from pilot and rapid small-scale column tests (RSSCTs). In all, the
EPA extracted 182 datasets from the ICR TSD database, which included some quarterly RSSCTs
and some long-term pilots. The EPA used RSSCT scaling factors identified in the original
datasets to scale predictions to expected full-scale operational time, rather than short duration
experimental time.
This appendix focuses on estimates of THM4 production because it forms the basis of potential
reductions in health risks resulting from reducing PFAS levels under all regulatory scenarios.
Note that the same approaches described in this appendix were used to estimate TOC removal.
The EPA developed a Python program to standardize the data analysis and produce graphics.
Figure 1-1 shows example data from one study (SystemID 1003, RSSCT) to demonstrate the
approach for estimating THM4 reduction. Each dataset provided influent and effluent
concentrations for TOC and THM4 formation potential for a 10-min empty bed contact time
(EBCT). Most datasets also included 20-min EBCT effluent concentrations. If data were not
available for 20-min EBCT effluent concentrations, then only 10-min EBCT data were included
in the analysis. For all datasets and EBCTs, the EPA used a logistic function to estimate the
expected breakthrough curve over time (effluent concentrations vs. time). Since the logistic
function is non-linear, the EPA used the Python function scipy.optimize.curvefit to estimate
equation parameters.
Final PFAS Rule Economic Analysis
1-1
April 2024
-------
FINAL RULE
APRIL 2024
70 ¦
60 -
X-
7*7
50 -
CI
=1
2 40-
S 30
c
0
u
1 20 H
10 -
0 -
THM4 Influent
65.87 avg. C,
THM4 lOmin EBCT
THM4 20min EBCT
THM4 10 min Logistic
THM4 20 min Logistic
20
40
100
120
140
60 80
Time (days)
Figure 1-1: Example Breakthrough Curve for THM4
from the ICR Dataset with Logistic Fit Functions Shown
160
The logistic function is provided as:
Equation 1-1:
C(t) = Cf(Ae~rt + l)~n+1
where C is effluent concentration, Cf is the final concentration (concentration units), A, r and n
are additional fit parameters and t is time (in days). The EPA generated a set of fit parameters for
each of the datasets and EBCTs. The logistic function provides a continuous function throughout
a period and can be used to estimate effective effluent concentrations beyond the original test
period. This assumes that Cf could be estimated effectively and represents the long-term
effective removal after breakthrough (i.e., that an equilibrium removal was achieved). Figure 1-2
shows the projected removal percentage for bed replacement intervals from 30 days (1 month) to
730 days (2 years). Percent removal for each data pair was calculated as:
Final PFAS Rule Economic Analysis
1-2
April 2024
-------
FINAL RULE
APRIL 2024
Equation 1-2:
%Removal = 100 * (1 C^ )
Cinf,avg
where, C(t) is the result of the logistic function over time, and Cinf avg is the average influent
concentration for each species.
100
80 -
2
pE 60 -
"fo
>
o
£
£ 40-
20 -
0
Figure 1-2: Example Percent Removal Results vs.
Time based on Logistic Plots Shown in Figure 1-1
THM4 - Avg. Cone. = 65.87
10
EBCT
EBCT
100 200 300 400 500
Replacement Interval (days)
600
700
Final PFAS Rule Economic Analysis
1-3
April 2024
-------
FINAL RULE
APRIL 2024
TOC - 10 min EBCT
27.2±8.8 % removed
TOC - 20 min EBCT
35.5±13.6 % removed
100 200 300 400 500
Replacement Interval (days)
THM4 - 10 min EBCT
29.9± 15.6 % removed
THM4 - 20 min EBCT
40.0±18.4 % removed
600
700
Figure 1-3: Mean Percentage Removal (Shaded Area ± 1 Standard Deviation)
Final PFAS Rule Economic Analysis
1-4
April 2024
-------
FINAL RULE
APRIL 2024
The percent removal formula provides a conservative estimate for removal over each EBCT. The
EPA assumes that the percent removal at the carbon removal day is the best removal that was
achieved, where breakthrough curves demonstrate that additional removal may be achieved for
earlier portions of the operational carbon life. For longer operational times, this early removal
capacity for each species becomes a diminishingly small percentage of removal percentage.
The EPA used the percentage removal at V2 year intervals for V2, 1, 1 V2, and 2 years in the
co-removal benefits analysis. Information about the source water (pre-categorized type from the
ICR, ground water or surface water) and averages of influent concentrations of TOC, and THM4
were stored with results, which were used during further analyses.
Figure 1-3 represents the mean percentage removal for TOC, THM4 over time with shaded areas
representing mean ±1 standard deviation. Figure 1-4 also shows a probability density function
representation of concentration reduction following treatment after 2 years of carbon operations
(i.e., GAC replacement time). These plots demonstrate the variability in the results.
0.05 -
0.04 -
>0.03-
S
0.02 -
0.01 -
0.00 -
0 50 100 150 200
THM4 (Infl.-Effl.) Concentration (/ng/L)
Figure 1-4: Probability Density Function of Concentration Difference
at 2 Years of Carbon Life (Subdivided by TOC level)
1.2 Discussion of Other Models
The EPA explored another existing model to determine THM4 removal (ATHM4) resulting from
granular activated carbon (GAC) treatment. The Water Treatment Plant (WTP) model uses the
ICR TSD data along with other datasets and includes specific process selection inputs such as
GAC units (U.S. EPA, 2001). In contrast with the logistic model detailed in Section 1.1, the WTP
model cannot be run with the GAC unit in isolation. Within the Water Treatment Plant model,
the GAC unit process equation relies on TOC and ultraviolet absorbance (UVA) changes and
does not directly predict THMs. Additional data needed to use the WTP model include types of
chemicals used, dosing concentrations, contact times, and full process train information, which
the EPA did not have outside of the DBP ICR for national scale estimates. Comparing the
models, the logistic equations for GAC treatment were generally in the same form. However, in
THM4 Density Plot
— All TOC: 10-min (N: 182)
All TOC: 20-min (N: 182)
1-2.0 TOC: 10-min (N: 20)
J \ 1-2.0 TOC: 20-min (N: 20)
:' i \ 2-3.5 TOC: 10-min (N: 103)
/ :\ j 2-3.5 TOC: 20-min (N: 103)
- ¦ - 3.5-5 TOC: 10-min (N: 44)
/ : ; - ¦ - 3.5-5 TOC: 20-min (N: 44)
above5 TOC: 10-min (N: 15)
/ '• I above5 TOC: 20-min (N: 15)
Final PFAS Rule Economic Analysis
1-5
April 2024
-------
FINAL RULE
APRIL 2024
this analysis, the EPA fit the THM4 results reported in the ICR dataset directly. In contrast, the
WTP model would need to have simulated all various treatment trains, including GAC, to
calculate TOC levels followed by a conversion with then another model equation to predict the
ATHM4. Both the simulation of treatment trains to calculate TOC levels and conversion to
predict the ATHM4 would add uncertainty to this approach. While these equations result in the
same shape of function to find predictions, the logistic model approach outlined in Section 1.1
uses a singular step with singular uncertainty that was data driven.
1.3 THM4 Reduction Results
All systems used free chlorine for the THM4 formation potential experiments in the ICR TSD.
However, the hold time to replicate the distribution system (DS) varied based on the typical
disinfectant used in the PWS. Table 1-1 shows the ATHM4 differences based on source water
type, EBCTs, and disinfectant type of the parent system. Table 1-2 to Table 1-5 shows the
ATHM4 differences based on GAC replacement intervals (1/2, 1, 1 V2, and 2 years), disinfectant
type (free chlorine versus chloramine), source water type (ground versus surface water), and
TOC range (1-2.0, 2-3.5, 3.5-5, and above 5 mg/L).
Final PFAS Rule Economic Analysis
1-6
April 2024
-------
FINAL RULE APRIL 2024
Table 1-1: ICR TSD Predictions for ATHM4 Based on Disinfectant
Disinfectant Source Pilot/ RSSCT ATHM4 with 10 min ATHM4 with 20 min ATHM4 with 10 min ATHM4 with 20 min
Type Type Count EBCT (%) EBCT (%) EBCT Qig/L) EBCT Qig/L)
Chloramine GW 21 30.5 ± 10.5 29.6 ± 15.3 43.0 ± 32.2 38.1 ±32.2
Chloramine, SW 102 26.6 ± 12.8 36.7 ± 14.5 29.0 ± 24.3 37.7 ± 26.2
Free Chlorine GW 16 34.7 ±24.3 35.3 ± 17.6 18.8 ±13.5 18.8 ±10.7
Free Chlorine SW 43 35.40 ± 17.8 54.7 ±20.8 20.2 ± 17.5 32.9 ±31.2
Abbreviations: EBCT - empty bed contact time; GW - ground water; RSSCT - rapid small-scale column test; SW - surface water; THM4 - four regulated trihalomethanes.
Final PFAS Rule Economic Analysis
1-7
April 2024
-------
FINAL RULE
APRIL 2024
Table 1-2: ICR TSD Predictions for ATHM4 for V2 Year GAC Replacement Based on Disinfectant Type, EBCT, and Source
Water Type
Source
Water
Type
TOC
Range
(mg/L)
ATHM4 with 10 min
ATHM4 with 20 min
ATHM4 with 10 min
ATHM4 with 20 min
Disinfectant
Count
EBCT (%Reduction
EBCT(%
EBCT (jig/L
EBCT (jig/L
Type
(N)
±1 Standard
Reduction ± 1
Reduction ± 1
Reduction ± 1
Deviation)
Standard Deviation)
Standard Deviation)
Standard Deviation)
1-2.0
3
38.09 ± 14.59
48.46 ±21.42
16.02 ±6.77
20.42 ±9.85
GW
2-3.5
4
51.61 ± 11.77
70.85 ± 1.40
31.79 ± 18.76
50.07 ±43.63
Vi
year
3.5-5
6
34.84 ±4.41
39.33 ±2.39
34.04 ± 17.05
42.42 ± 27.47
Chloramine
Above 5
8
33.41 ±6.39
34.53 ± 14.62
86.59 ±20.77
84.86 ±30.12
1-2.0
5
33.69 ±27.18
43.68 ±30.09
16.49 ±8.62
22.78 ± 12.69
SW
2-3.5
59
36.87 ± 15.24
57.29 ± 17.23
29.15 ± 17.83
44.57 ±23.77
3.5-5
31
36.11 ± 11.62
52.84 ± 13.91
49.95 ±33.55
72.35 ±41.99
Above 5
7
40.79 ±5.04
51.16 ±8.68
73.81± 20.77
90.92 ±21.64
1-2.0
5
55.33 ±22.41
59.13 ±20.53
28.74 ± 19.06
25.74 ± 12.18
GW
2-3.5
10
33.81 ± 17.98
48.58 ± 19.85
18.95 ±9.83
27.45 ± 12.81
Free chlorine
3.5-5
1
87.56
49.50
41.99
23.73
1-2.0
7
60.83 ± 25.20
84.69 ±25.89
13.91 ±8.54
20.28 ± 12.94
SW
2-3.5
30
49.21 ± 19.68
74.65 ± 15.39
32.04 ±23.71
50.60 ± 36.79
3.5-5
6
42.78 ± 10.26
63.53 ± 17.68
30.57 ±24.87
42.46 ±31.69
Abbreviations: EBCT - empty bed contact time; GAC - granular activated carbon; GW - ground water; ICR TSD - Information Collection Rule Treatment Study Database;
SW - surface water; THM4 - four regulated trihalomethanes; TOC - total organic carbon.
Final PFAS Rule Economic Analysis
1-8
April 2024
-------
FINAL RULE
APRIL 2024
Table 1-3: ICR TSD Predictions for ATHM4 for One Year GAC Replacement Based on Disinfectant Type, EBCT, and
Source Water Type
Disinfectant
Type
Source
Water
Type
TOC
Range
(mg/L)
Count
(N)
ATHM4 with 10
min EBCT
(%Reduction ± 1
Standard Deviation)
ATHM4 with 20
min EBCT
(%Reduction ± 1
Standard Deviation)
ATHM4 with 10
min EBCT (jig/L
Reduction ± 1
Standard Deviation)
ATHM4 with 20
min EBCT (jig/L
Reduction ± 1
Standard Deviation)
1-2.0
3
32.14 ± 14.75
33.55 ± 16.87
13.55 ±6.76
14.16 ±7.68
GW
2-3.5
4
39.39 ± 17.79
55.20 ±7.81
21.38 ±7.40
38.25 ±32.05
3.5-5
6
31.61 ±4.48
32.56 ±3.55
30.76 ± 15.12
33.06 ± 15.17
Chloramine
Above 5
8
31.33 ±6.43
27.57 ± 16.09
81.10 ± 19.88
66.03 ±35.55
1
year
1-2.0
5
22.40 ± 16.25
33.48 ±23.63
11.13 ± 6.38
17.24 ±9.33
SW
2-3.5
59
29.59 ± 13.50
44.65 ± 15.02
23.82 ± 15.60
34.77 ± 18.39
3.5-5
31
30.88 ± 12.05
42.95 ± 13.96
43.06 ±30.99
58.76 ±35.32
Above 5
7
36.90 ±4.72
42.70 ± 9.72
66.85 ± 19.58
75.13 ± 18.43
1-2.0
5
45.26 ±20.71
48.48 ± 18.62
23.75 ± 16.84
21.17 ± 10.73
GW
2-3.5
10
28.46 ± 17.25
36.76 ± 17.66
16.17 ±9.50
21.35 ± 11.95
Free Chlorine
3.5-5
1
93.04
49.50
44.61
23.73
1-2.0
7
49.44 ±21.75
73.99 ±25.56
11.00 ±6.30
17.02 ±9.75
SW
2-3.5
30
39.04 ± 17.75
61.02 ± 16.94
25.33 ±20.13
41.75 ±34.79
3.5-5
6
36.29 ± 14.08
55.21 ±21.66
26.15 ±20.67
35.33 ±25.67
Abbreviations: EBCT - empty bed contact time; GAC - granular activated carbon; GW - ground water; ICR TSD - Information Collection Rule Treatment Study Database;
SW - surface water; THM4 -four regulated trihalomethanes; TOC - total organic carbon.
Final PFAS Rule Economic Analysis
1-9
April 2024
-------
FINAL RULE
APRIL 2024
Table 1-4: ICR TSD Predictions for ATHM4 for 1 Vi Year GAC Replacement Based on Disinfectant Type, EBCT, and
Source Water Type
Source
Water
Type
TOC
Range
(mg/L)
ATHM4 with 10
ATHM4 with 20
ATHM4 with 10
ATHM4 with 20
Disinfectant
Count
min EBCT (%
min EBCT (%
min EBCT (jig/L
min EBCT (jig/L
Type
(N)
reduction ± 1
reduction ± 1
reduction ± 1
reduction ± 1
standard deviation)
standard deviation)
standard deviation)
standard deviation)
1-2.0
3
30.17 ± 14.81
27.31 ± 13.19
12.73 ±6.76
11.52 ±6.02
GW
2-3.5
4
35.06 ±20.01
48.68 ± 11.30
17.79 ±5.62
33.79 ±28.67
3.5-5
6
30.54 ±4.61
30.32 ±5.21
29.67 ± 14.51
29.96 ± 11.22
Chloramine
Above 5
8
30.64 ±6.45
25.26 ± 16.63
79.29 ± 19.61
59.80 ±37.61
1 '/2
year
1-2.0
5
18.19 ± 13.29
28.56 ± 19.06
9.21 ±6.28
14.93 ±8.17
SW
2-3.5
59
26.99 ± 13.11
39.59 ± 14.66
21.94 ± 14.98
30.94 ± 16.92
3.5-5
31
29.14 ± 12.31
39.60 ± 14.37
40.78 ± 30.26
54.13 ±33.41
Above 5
7
35.61 ±4.79
39.86 ± 10.48
64.55 ± 19.30
69.85 ± 18.23
1-2.0
5
41.91 ±20.19
44.95 ± 17.99
22.10 ± 16.10
19.66 ± 10.25
GW
2-3.5
10
26.68 ± 17.09
32.73 ± 17.45
15.26 ±9.44
19.27 ± 11.88
Free chlorine
3.5-5
1
94.96
49.50
45.53
23.73
1-2.0
7
45.53 ±21.01
68.48 ±25.48
10.02 ±5.61
15.42 ±8.41
SW
2-3.5
30
35.66 ± 17.51
55.85 ± 18.31
23.10 ± 19.09
38.58 ±34.59
3.5-5
6
34.14 ± 15.63
52.45 ±23.08
24.69 ± 19.35
32.96 ±23.79
Abbreviations: EBCT - empty bed contact time; GAC - granular activated carbon; GW - ground water; ICR TSD - Information Collection Rule Treatment Study Database;
SW - surface water; THM4 - four regulated trihalomethanes; TOC - total organic carbon.
Final PFAS Rule Economic Analysis
1-10
April 2024
-------
FINAL RULE
APRIL 2024
Table 1-5: ICR TSD Predictions for ATHM4 for Two Year GAC Replacement Based on Disinfectant Type, EBCT, and
Source Water Type
Source
Water
Type
TOC
Range
(mg/L)
ATHM4 with 10
ATHM4 with 20
ATHM4 with 10
ATHM4 with 20
Disinfectant
Count
min EBCT (%
min EBCT (%
min EBCT (jig/L
min EBCT (jig/L
Type
(N)
reduction ± 1
reduction ± 1
reduction ± 1
reduction ± 1
standard deviation)
standard deviation)
standard deviation)
standard deviation)
1-2.0
3
29.18 ± 14.84
24.02 ± 11.12
12.31 ±6.75
10.13 ±5.09
GW
2-3.5
4
32.87 ±21.16
45.31 ± 13.18
15.99 ±5.85
31.51 ±27.06
3.5-5
6
30.00 ±4.69
29.20 ± 6.06
29.13 ± 14.21
28.40 ±9.32
Chloramine
Above 5
8
30.30 ±6.47
24.10 ± 16.91
78.37 ± 19.48
56.66 ±38.69
2
1-2.0
5
16.08 ± 12.47
26.09 ± 16.95
8.25 ±6.42
13.76 ±7.67
year
SW
2-3.5
59
25.69 ± 13.10
36.81 ± 14.64
21.00 ± 14.73
28.86 ± 16.36
3.5-5
31
28.27 ± 12.46
37.92 ± 14.65
39.63 ±29.92
51.80 ±32.56
Above 5
7
34.97 ±4.86
38.44 ± 10.92
63.39 ± 19.18
67.20 ± 18.30
1-2.0
5
40.23 ± 19.94
43.17 ± 17.68
21.26 ± 15.73
18.90 ± 10.01
GW
2-3.5
10
25.79 ± 17.03
30.70 ± 17.46
14.79 ±9.42
18.23 ± 11.89
Free chlorine
3.5-5
1
95.92
49.50
46.00
23.73
1-2.0
7
43.57 ±20.76
65.69 ±25.67
9.52 ±5.27
14.61 ±7.76
SW
2-3.5
30
33.97 ± 17.48
53.22 ± 19.21
21.99 ± 18.59
36.97 ±34.54
3.5-5
6
33.06 ± 16.43
51.06 ±23.81
23.95 ± 18.71
31.77 ±22.87
Abbreviations: EBCT - empty bed contact time; GAC - granular activated carbon; GW - ground water; ICR TSD - Information Collection Rule Treatment Study Database;
SW - surface water; THM4 - four regulated trihalomethanes; TOC - total organic carbon.
Final PFAS Rule Economic Analysis
1-11
April 2024
-------
FINAL RULE
APRIL 2024
1.4 Sampling Points from the Fourth Six Year Review Plants
with Granular Activated Carbon Treatment
To examine the Six Year Review 4 (SYR4) THM4 data,
the EPA extracted and matched sampling point IDs for the years that represent before and after
GAC treatment. Only sampling point IDs with the same number of samples before and after
GAC treatment were used to determine THM4 averages. To calculate a single location
comparison, the EPA selected one sampling point ID for each public water system identification
(PWSID). Entry point (EP) sampling point types were used when available. When unavailable,
the EPA used the first sampling point type. Table 1-6 shows an example of sampling point IDs,
sampling point types, and number of samples available for one PWSID in the SYR4 dataset.
Table 1-6: Sampling Point IDs for each PWSID were Extracted and Matched for the
Years that Represent Before/After GAC Treatment (Example: PWSID AL0000577)
Sampling Point ID
Sampling Point Type
# Of Samples (2017, 2019)
ATHM4 (jig/L)a
12967
WS
29 (4, 4)
8.5
12970
WS
29 (4, 4)
8.9
12972
WS
29 (4, 4)
8.5
12974
WS
29 (4 ,4)
9.3
12975
EP
32 (4, 4)
5.7
12976
WS
29 (4, 4)
15.8
12977
DS
32 (4, 4)
10.4
12978
WS
28 (4, 4)
9.4
12979
WS
29 (4, 4)
9.8
12980
DS
24 (3, 0)
-
12981
DS
26 (4, 0)
-
12983
DS
26 (4, 0)
-
13022
WS
25 (4, 4)
11.9
13044
DS
6 (0, 4)
-
13089
MR
2(1.0)
-
Abbreviations: DS - distribution system; EP - entry point; MR - point of maximum residence; WS - water system facility
point.
Notes:
aATHM4 was not calculated for sampling point IDs that did not have sample data for the years that represent either before or
after GAC treatment.
Final PFAS Rule Economic Analysis
1-0
April 2024
-------
FINAL RULE
APRIL 2024
Appendix J. Value of a Statistical Life Updating
The EPA follows U.S. EPA (2010) to estimate the economic value of avoiding premature
mortality. To obtain a Value of Statistical Life (VSL) suitable for valuation of mortality risk
reductions during 2024-2105, the EPA relies on the base value estimate of $4.8 million ($1990,
1990 income year), which is the central tendency of the Value of Statistical Life distribution
recommended for use in the EPA's regulatory impact analyses (U.S. EPA, 2010). The EPA
adjusted the base Value of Statistical Life estimate for inflation and income growth as follows:
Vt ?n? ? —
Equation J-l:
^2022 I Yt
't,2022 — "1990,1990 D
i 1
(fS
Vliqqn/
1990 vi1990^
Where:
^t,2022 VSL value ($2022) updated for use in evaluation year t, t = 2024 ... 2050;
^1990,1990 Base VSL value of $4,800,000 ($1990, 1990 income year);
P2022 Gross Domestic Product (GDP) price deflator index value in 2022;
P1990 GDP price deflator index value in 1990;
Yt Projected income per capita ($2012) in evaluation year t,t = 2024 ... 2050;
liggo Historical income per capita ($2012) in 1990;
e VSL income elasticity of 0.4 as recommended by U.S. EPA (U.S. EPA, 2010).
The EPA used disposable personal annual income to represent U.S. income per capita. Because
the PFAS analysis spans a future time period from 2024 to 2105, the EPA relied on the long-term
personal disposable income projections from the U.S. Energy Information Administration
(2021). The long-term personal income projections are available annually from 2020 to 2050.
The EPA's SafeWater model requires a single income growth factor to project the 2024 Value of
Statistical Life (in $2022) to future years (2025 through 2105). Based on the Value of Statistical
Life estimates calculated using Equation J-l, the EPA calculated the compound annual growth
rate, CAGR, of Value of Statistical Life values from 2024 to 2050 as follows:
Equation J-2:
(V20S0 2022\^2050^2024)
CAGR = ' - 1
V2024.2022/
The EPA used the calculated CAGR value to approximate Value of Statistical Life growth
during the analysis period (2024 to 2105) based on the 2022 Value of Statistical Life value
estimated using Equation J-l.
Equation J-3:
^t,2022 = ^2024,2022 " (1 + CAGR)1 2024
Final PFAS Rule Economic Analysis J-l April 2024
-------
FINAL RULE
APRIL 2024
Table J-l summarizes the projected Value of Statistical Life estimates through 2050 and the
approximated Value of Statistical Life estimates through 2105.
Table J-l: Estimated Value of Statistical Life Series
Year
Historical
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Projected
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Income Growth
Factor (Ratio of
Projected
PDYPP to
Historical 1990
PDYPP to the
Power of 0.4)
Projected Value of
Statistical Life
($2022)
Approximated
Value of Statistical
Life
($2022)
1990
30,327
-
1
9,597,133
-
2024
-
47,987
1.201330302
11,529,327
11,529,327
2025
-
48,917
1.210595048
11,618,242
11,601,616
2026
-
49,760
1.218899284
11,697,939
11,674,358
2027
-
50,616
1.2272399
11,777,985
11,747,556
2028
-
51,496
1.235732098
11,859,486
11,821,214
2029
-
52,407
1.244430191
11,942,963
11,895,333
2030
-
53,393
1.253742955
12,032,338
11,969,916
2031
-
54,326
1.262455217
12,115,951
12,044,968
2032
-
55,258
1.271073774
12,198,665
12,120,490
2033
-
56,207
1.279765868
12,282,084
12,196,485
2034
-
57,145
1.288265959
12,363,660
12,272,957
2035
-
58,072
1.296586905
12,443,518
12,349,909
2036
-
58,985
1.304696423
12,521,346
12,427,343
2037
-
59,874
1.312534459
12,596,568
12,505,262
2038
-
60,753
1.320206338
12,670,196
12,583,670
2039
-
61,643
1.327910067
12,744,130
12,662,570
2040
-
62,513
1.335367798
12,815,703
12,741,964
2041
-
63,408
1.342991031
12,888,864
12,821,856
2042
-
64,346
1.350901532
12,964,782
12,902,249
2043
-
65,282
1.358723314
13,039,849
12,983,146
2044
-
66,210
1.366414095
13,113,658
13,064,550
2045
-
67,148
1.374127034
13,187,681
13,146,465
2046
-
68,095
1.381844195
13,261,743
13,228,894
2047
-
69,069
1.389721143
13,337,339
13,311,839
2048
-
70,076
1.397792319
13,414,799
13,395,304
2049
-
71,066
1.405655221
13,490,261
13,479,292
2050
-
72,024
1.413208106
13,562,747
13,563,808
2051
-
-
-
-
13,648,853
2052
-
-
-
-
13,734,431
2053
-
-
-
-
13,820,546
2054
-
-
-
-
13,907,201
2055
-
-
-
-
13,994,399
Final PFAS Rule Economic Analysis
J-2
April 2024
-------
FINAL RULE APRIL 2024
Table J-l: Estimated Value of Statistical Life Series
Year
Historical
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Projected
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Income Growth
Factor (Ratio of
Projected
PDYPP to
Historical 1990
PDYPP to the
Power of 0.4)
Projected Value of
Statistical Life
($2022)
Approximated
Value of Statistical
Life
($2022)
2056
-
-
-
-
14,082,144
2057
-
-
-
-
14,170,439
2058
-
-
-
-
14,259,287
2059
-
-
-
-
14,348,693
2060
-
-
-
-
14,438,660
2061
-
-
-
-
14,529,190
2062
-
-
-
-
14,620,288
2063
-
-
-
-
14,711,957
2064
-
-
-
-
14,804,201
2065
-
-
-
-
14,897,023
2066
-
-
-
-
14,990,428
2067
-
-
-
-
15,084,418
2068
-
-
-
-
15,178,997
2069
-
-
-
-
15,274,169
2070
-
-
-
-
15,369,938
2071
-
-
-
-
15,466,308
2072
-
-
-
-
15,563,282
2073
-
-
-
-
15,660,863
2074
-
-
-
-
15,759,057
2075
-
-
-
-
15,857,866
2076
-
-
-
-
15,957,295
2077
-
-
-
-
16,057,347
2078
-
-
-
-
16,158,027
2079
-
-
-
-
16,259,338
2080
-
-
-
-
16,361,284
2081
-
-
-
-
16,463,869
2082
-
-
-
-
16,567,098
2083
-
-
-
-
16,670,973
2084
-
-
-
-
16,775,500
2085
-
-
-
-
16,880,683
2086
-
-
-
-
16,986,525
2087
-
-
-
-
17,093,030
2088
-
-
-
-
17,200,203
2089
-
-
-
-
17,308,049
2090
-
-
-
-
17,416,570
Final PFAS Rule Economic Analysis
April 2024
-------
FINAL RULE
APRIL 2024
Table J-l: Estimated Value of Statistical Life Series
Historical
Projected
Income Growth
Personal
Personal
Factor (Ratio of
Approximated
Value of Statistical
Life
($2022)
Disposable
Disposable
Projected
Projected Value of
Year Income Per
Income Per
PDYPP to
Statistical Life
Capita
Capita
Historical 1990
($2022)
(PDYPP,
(PDYPP,
PDYPP to the
$2012)
$2012)
Power of 0.4)
2091
.
17,525,772
2092
-
17,635,659
2093
-
17,746,234
2094
-
17,857,503
2095
-
17,969,470
2096
-
18,082,138
2097
-
18,195,513
2098
-
18,309,599
2099
-
18,424,400
2100
-
18,539,921
2101
-
18,656,167
2102
-
18,773,141
2103
-
18,890,848
2104
-
19,009,294
2105
.
19,128,482
Acronym: PDYPP- personal disposable income per capita.
Table J-2 summarizes the data employed in updating the values used to monetize reductions in
mortality and morbidity risks in the population exposed to PFOA and PFOS in drinking water.
The EPA uses the Value of Statistical Life to monetize reduced mortality benefits and uses the
COI to monetize reduced morbidity benefits. The details on morbidity valuation for birth weight,
CVD, RCC, and bladder cancer analyses are provided in the respective sections of the main
document.
Table J-2: Summary of Inputs and Data Sources Used for Valuation
Data Element
Modeled
Variability
Data Source
Notes
Base Value of
Statistical Life
Value of
Statistical Life
income elasticity
Medical Care
CPI
None
None
Time: Annual,
1990..2023
U.S. EPA,
2010
U.S. EPA,
2010
BLS 2022
(U.S. Bureau of
Labor
The base value of 4,800,000 ($1990) was used
as recommended by the U.S. EPA Guidelines
for Preparing Economic Analyses.
Income growth adjustments were done using
income elasticity 0.4 per recommendations in
the U.S. EPA Guidelines for Preparing
Economic Analyses.
Medical cost inflation adjustments were done
using annual CPI for medical care (U.S. city
average, all urban consumers, series number
CUUR0000SAM).
Final PFAS Rule Economic Analysis
April 2024
-------
FINAL RULE
APRIL 2024
Table J-2: Summary of Inputs and Data Sources Used for Valuation
Data Element
Modeled
Variability
Data Source
Notes
Employment
Cost Index
GDP Price
Deflator Index
Historical
income per
capita
Projected
income per
capita
Time: Quarterly,
2001..2022
Time: Annual,
1990..2022
Time: Annual,
1990..2022
Time: Annual,
2020..2050
Statistics,
2022a)
BLS 2022
(U.S. Bureau of
Labor
Statistics,
2022b)
BEA 2023
(U.S. Bureau of
Economic
Analysis, 2023)
BEA 2021
(U.S. Bureau of
Economic
Analysis, 2021)
U.S. EIA 2021
(U.S. Energy
Information
Administration,
2021)
Opportunity cost inflation adjustments were
done using quarterly index for total
compensation for all civilian workers in all
industries and occupations (series number
CIS 10100000000001).
Value of Statistical Life inflation adjustments
were done using annual GDP price deflator
index.
Disposable personal annual income per capita
(series number A229RC0A052NBEA). Data are
in $2022. The series were converted to constant
$2012 to align with US EIA 2021 projections
using BLS 2022 CPI series.
The U.S. EIA long-term projections focus on
components of potential growth, fiscal balances
and debt accumulation, domestic saving and
investment balances, and external balances are
covered and interest rates consistent with those
projections. The projection horizon is 2050. The
EPA used the ratio of projected real disposable
personal income (in constant $2012, series
number 18-AEO2021.55.ref2021-dl 13020a) to
project population size (series number
18-AEQ2021.42.ref2021-dl 13020a).
Abbreviations: BEA - Bureau of Economic Analysis; BLS - Bureau of Labor Statistics; CPI -
EIA - Energy Information Administration; GDP - gross domestic product.
consumer price index;
Final PFAS Rule Economic Analysis
April 2024
-------
FINAL RULE
APRIL 2024
Appendix K. Benefits Sensitivity Analyses
This appendix provides details on the sensitivity analyses implemented by the EPA to evaluate
the impact of the exposure-response assumptions in the CVD benefits model and the impact of
Perfluorononanoic Acid (PFNA) inclusion in the birth weight benefits model. Section K. 1
describes hypothetical regulatory alternatives evaluated in the sensitivity analyses. Section K.2
provides details on estimation of blood serum PFOA, PFOS, and PFNA. Section K.3 summarizes
the CVD exposure response scenarios and presents the associated results. Section K.4
summarizes the birth weight dose response scenarios and results. Section K.5 summarizes the
RCC exposure response scenarios and results.
The sensitivity analyses described herein relied on methodology implemented in R software (R
Core Team, 2021) and differ slightly from SafeWater MCBC methods. Specifically, SafeWater
performs a set of pre-calculations to maximize computational efficiency and, as such, the order
of analytical steps across R and SafeWater models differs; however, results across models are
mathematically consistent. The R-based model version treats each integer age cohort between 85
and 99 separately, implements the CVD calculations for those aged 40-89 years only, and applies
the ASCVD model-based annual incidence at age 80 years to ages 81-89 because the ASCVD
model has been fit to those aged 40-80 years and predicts the 10-year probability of the first
CVD event.
K.l Overview of the Hypothetical Exposure Reduction
Table K-l shows the details of the two hypothetical exposure reductions for the sensitivity
analyses. For both alternatives, the EPA assumed the same population served size of 100,000
distributed over age-, sex-, and race-ethnicity categories using national-level demographic data
(see Appendix B). Hypothetical exposure reduction 1 assumes a reduction of 1 ppt in PFOA and
a reduction of 1 ppt in PFOS. Hypothetical exposure reduction 2 assumes a reduction of 1 ppt in
PFNA,58 in addition to the reductions specified for hypothetical exposure reduction 1. Additional
sensitivity analysis assumptions (other than those pertaining to the exposure-response scenarios
in Section K.3 and Section K.4), such as evaluation period, population growth, etc., align with
those used in the economic analysis. The EPA notes that uncertainty was not characterized for
these sensitivity analysis scenarios. All parameters treated as uncertain in the economic analysis
were set to their central estimate values (see Appendix L).
The EPA notes that relative magnitudes of reductions in PFOA, PFOS, and PFNA may differ
from those evaluated in the economic analysis. At EPs where PFOA, PFOS, and PFNA
concentrations exceed their respective final MCLs, the EPA expects reductions of 1 ppt or
greater. Multiple data sources, including UCMR 3 and state-collected finished drinking water
data, demonstrate that PFNA has been detected between 0.22 ppt and 94.2 ppt. In UCMR 3,
0.28% of participating systems (14 total) had PFNA detections greater than/equal to the MRL
(20 ppt), while state monitoring efforts showed that the number of systems in each state with
PFNA detections ranged between 0.0% and 16.5%. The EPA chose to evaluate unit reductions
(i.e., 1 ppt each) to demonstrate the effects of and make comparisons between unit changes in
58 Note that the inclusion of PFNA under Alternative 2 was only relevant to BW sensitivity analysis because there is evidence
that PFNA reductions can improve BW. There is a lack of supporting evidence for an impact for CVD and RCC benefits.
Final PFAS Rule Economic Analysis
K-l
April 2024
-------
FINAL RULE
APRIL 2024
PFOA, PFOS, and PFNA exposure (U.S. EPA, 2024d). Caution should be exercised in
quantifying the potential magnitude of change in the national benefits estimates based on the
results of these sensitivity analyses, although conclusions about the directionality of these effects
can be inferred.
Table K-l: Overview of Hypothetical Exposure Reductions
Hypothetical Exposure Reduction
Parameter Description | 2
(PFOA+PFOS) (PFOA+PFOS+PFNA)
Population served at the start of the evaluation period 100,000 100,000
Reduction in PFOA concentration (ppt) 1 1
Reduction in PFOS concentration (ppt) 1 1
Reduction in PFNA concentration (ppt) 0 1
Abbreviations: PFNA - perfluorononanoic acid; PFOA - periluorooctanoic acid; PFOS - perfluorooctane sulfonic acid.
K.2 Estimation of Blood Serum PFOA, PFOS, and PFNA
The EPA used PFOA and PFOS drinking water concentrations as inputs to its Pharmacokinetic
(PK) model to estimate blood serum PFOA and PFOS concentrations for adult males and
females. See the EPA's Github repository for PK modeling59 and the Final Human Health
Toxicity Assessments for PFOA and PFOS for further information on the PFOA/PFOS model
(U.S. EPA, 2024b; U.S. EPA, 2024c). Application of the PK model in the context of the benefits
estimation is detailed in Section 6.3 of the economic analysis.
To estimate blood serum PFNA based on its drinking water concentration, the EPA used a first-
order single-compartment model whose behavior was previously demonstrated to be consistent
with PFOA pharmacokinetics in humans (Bartell et al., 2010). Equation K-l-Equation K-4
summarize this model (Bartell, 2003; Bartell, 2017; Lu & Bartell, 2020):
59 https://gitliub.coin/USEPA/OW-PFOS-PFOA-MCLG-support-PK-models
Final PFAS Rule Economic Analysis
K-2
April 2024
-------
FINAL RULE
APRIL 2024
Equation K-l:
W*S
Coo — B + ¦
1000
Equation K-2:
Ct = Coo + (B-Coo) * e~kt
Equation K-3:
k = ln(2) /t1/2
Equation K-4:
k-Vd
Where:
Coo = steady-state serum PFNA concentration (ng/mL);
Ct = serum concentration at time t (ng/mL);
t = time since beginning of / change in the water exposure (days);
B = background serum PFNA concentration (ng/mL). The EPA used an estimate of 0.411
ng/mL for 2017-2018 from Centers for Disease Control and Prevention (2022);
W = drinking water PFNA concentration (ppt);
S = steady-state serum/water concentration ratio (unitless);
k = first order elimination rate constant for PFNA from serum (days-1), defined as a
function of half-life in Equation K-3 (Bartell, 2003);
tXj2 = PFNA half-life in serum (days). Following Lu and Bartell (2020) model assumptions,
the EPA used an estimate of 3.9 years from Zhang et al. (2013) (weighted average estimate),
after converting it to 1,424.5 days;
/ = fraction of PFNA absorbed (unitless). Following Lu and Bartell (2020) model
assumptions, the EPA used 100% absorption;
Q = water intake (L/kg body weight per day). Consistent with assumptions used for serum
PFOA and PFOS, the EPA used a water intake of 0.013 L/kg of body weight per day (U.S. EPA,
201 lb) in order to compute the PFNA dose from drinking water sources; and
Vd = volume of distribution (L/kg body weight per day), a proportionality constant relating
the total amount of a chemical in the body to the concentration in plasma (Hoffman et al., 2011).
Following Lu and Bartell (2020) model assumptions, the EPA used an estimate of 0.17 L/kg
body weight from Zhang et al. (2013).
Using this model, the EPA evaluated lifetime baseline and lifetime regulatory alternative
exposure scenarios described in Section 6.3 of the economic analysis and used the difference
between the two as an input to the downstream analysis of health effects.
Final PFAS Rule Economic Analysis K-3 April 2024
-------
FINAL RULE
APRIL 2024
K.3 CVD Sensitivity Analyses
CVD sensitivity analyses rely on hypothetical exposure reduction 1 (i.e., 1 ppt reduction in
PFOA and 1 ppt reduction in PFOS) to explore the impact of the following changes in the CVD
exposure-response modeling:
• The use of single study-based TC effect estimates, rather than the EPA meta-analysis-based
effect estimates. To this end, the EPA used estimates from a large NHANES study (Dong et
al., 2019) and estimates from a longitudinal study of diabetes prevention program outcomes
study (P.-I. D. Lin et al., 2019);
• Inclusion of HDLC effects from the CVD analysis; and,
• Exclusion of BP effects from the CVD analysis.
Table K-2 summarizes the exposure-response scenarios, while Table K-3 provides details on the
slope factors used in this sensitivity analysis.
Table K-2: Overview of CVD Exposure-Response Scenarios
Exposure-Response „ . „ ... „
* c, . * Scenario Definition
Economic analysis scenario using the EPA meta-analysis for TC, Liao et al.
(2020) for BP, and excluding HDLC impacts.
Scenario using Dong et al. (2019) for TC, Liao et al. (2020) for BP, and
excluding HDLC impacts.
Scenario using P.-I. D. Lin et al. (2019) for TC, Liao et al. (2020) for BP, and
excluding HDLC impacts.
Scenario using the EPA meta-analysis for TC and HDLC, and Liao et al. (2020)
for BP.
Scenario using Dong et al. (2019) for TC and HDLC, and Liao et al. (2020) for
BP.
Scenario using P.-I. D. Lin et al. (2019) for TC and HDLC, and Liao et al. (2020)
for BP.
Scenario using the EPA meta-analysis for TC and excluding HDLC and BP
impacts. This scenario is most comparable to the U.S. EPA (2021a) analysis
implemented for the SAB review.
Scenario using Dong et al. (2019) for TC and excluding HDLC and BP impacts.
Scenario using P.-I. D. Lin et al. (2019) for TC and excluding HDLC and BP
impacts.
Scenario using the EPA meta-analysis for TC and HDLC, and excluding BP
impacts.
Scenario using Dong et al. (2019) for TC and HDLC, and excluding BP impacts.
Scenario using P.-I. D. Lin et al. (2019) for TC and HDLC, and excluding BP
impacts.
Abbreviations: BP - blood pressure; CVD - cardiovascular disease; EA - economic analysis; HDLC - high-density
lipoprotein cholesterol; PFOA - pertluorooctanoic acid; PFOS - perfluorooctane sulfonic acid; SAB - Science Advisory
Board; TC - total cholesterol.
1-EA
2-Dong
3-Lin
4-EA (+HDLC)
5-Dong (+HDLC)
6-Lin (+HDLC)
7-EA (-BP)
8-Dong (-BP)
9-Lin (-BP)
10-EA (-BP +HDLC)
11-Dong (-BP +HDLC)
12-Lin (-BP +HDLC)
Final PFAS Rule Economic Analysis
K-4
April 2024
-------
FINAL RULE
APRIL 2024
Table K-3: Exposure-Response Information for CVD Biomarkers
Linear Slope Estimate (mg/dL per 1 ng/mL)
Source Contaminant
TC HDLC BP
Serum PFOA
1.57
0.11
EPA meta-analysis3
Serum PFOS
(CI95: 0.02,3.13)
0.08
(CI95:-0.01,0.16)
(CI95: -0.22, 0.43)
0.05
(CI95:-0.01, 0.11)
-
Serum PFOA
1.48
-0.03
Dong et al. (2019)
Serum PFOS
(CI95: 0.18, 2.78)
0.40
(CI95: 0.13, 0.67)
(CI95: -0.44, 0.39)
0.01
(CI95:-0.08, 0.11)
-
Serum PFOA
1.63
-0.13
P.-I. D. Lin et al.
(CI95: -0.84, 2.42)
(CI95:-0.37,0.107)
(2019)
Serum PFOS
0.13
(CI95: -0.005,0.27)
-0.02
(CI95: -0.06, 0.02)
-
Liao et al. (2020)
Serum PFOS
0.044
—
(CI95: 0.006,0.083)
Abbreviations: BP - systolic blood pressure; CI95 - 95% CI; CVD - cardiovascular disease; HDLC - high-density lipoprotein
cholesterol; PFOA - perfluorooctanoic acid; PFOS - perfluorooctane sulfonic acid; TC - total cholesterol.
Notes:
aSee Section 6.5.2 of the economic analysis.
Table K-4 shows the results of the CVD sensitivity analysis. The EPA made the following
observations:
• Relative to the annualized CVD benefits estimated using the EPA meta-analysis-based slope
factors, using the Dong et al. (2019) slope factors increases the annualized CVD benefits by
12.2%, while using the P.-I. D. Lin et al. (2019) slope factors increases the annualized CVD
benefits by 6.3%%.
• Inclusion of HDLC effects decreases annualized CVD benefits by 20.5% if the EPA meta-
analysis slope factors are used. The use of Dong et al. (2019) and the P.-I. D. Lin et al.
(2019) instead of the EPA meta-analysis slope factors increases annualized benefits by 2.4%
and 18.4%), respectively. The wide variation in the impact of HDLC inclusion may be
explained by high variance in the slope factor estimates. The EPA notes, however, that none
of the PFOA/PFOS-HDLC slope factors are statistically significant at the 5% level.
• Exclusion of BP effects decreases annualized CVD benefits by 2.5% if the EPA meta-
analysis slope factors are used. However, estimates decrease by 2.2% and 2.3% if the Dong
et al. (2019) and the P.-I. D. Lin et al. (2019), respectively, slope factors are used.
The relative magnitudes of reductions in PFOA and PFOS used in this sensitivity analysis may
differ from those implied by the regulatory alternatives evaluated in the economic analysis.
Therefore, the potential magnitude of changes in national CVD benefits due to alternative
TC/HDLC exposure-response assumptions as well as exclusion of the BP effects may differ from
the ones estimated in this sensitivity analyses.
Final PFAS Rule Economic Analysis
K-5
April 2024
-------
FINAL RULE APRIL 2024
Table K-4: Summary of CVD Sensitivity Analysis for Hypothetical Exposure Reduction 1 (PFOA+PFOS)
Result Description3 Exposure-Response Scenario13 0
1-EA
M
U
-J
p
#J3 U
U
-J
p
£
CO
£
CO
1
£
CO
CLh
CLh
CLh
a
0
a
1
fl
J
1
5
n
i
-t
H J
p B
« +
£
s
J
1
<
w
1
r-
M
a
0
a
1
90
a
J
1
o\
< p
U X
1 T
o +
o P
P w
+
.3 P
as
Average reduction in serum PFOA
0.091
0.091
0.091
0.091
0.091
0.091
0.091
0.091
0.091
0.091
0.091
0.091
concentration (ng/mL)
Average reduction in serum PFOS
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
0.084
concentration (ng/mL)
Average reduction in TC
0.150
0.168
0.160
0.150
0.168
0.160
0.150
0.168
0.160
0.150
0.168
0.160
concentration (mg/dL)
Average reduction in HDLC
0.000
0.000
0.000
0.014
-0.002
-0.014
0.000
0.000
0.000
0.014
-0.002
-0.014
concentration (mg/dL)
Average reduction in BP (mmHg)
0.004
0.004
0.004
0.004
0.004
0.004
0.000
0.000
0.000
0.000
0.000
0.000
Non-fatal first MI (total cases
2.745
3.084
2.920
1.973
3.187
3.654
2.708
3.048
2.883
1.936
3.150
3.618
avoided)d
Non-fatal first IS (total cases
3.965
4.455
4.218
3.005
4.583
5.130
3.909
4.399
4.161
2.948
4.526
5.073
avoided)d
CVD deaths (total cases avoided)d
0.778
0.875
0.828
0.641
0.893
0.958
0.755
0.852
0.804
0.618
0.870
0.935
PDV, non-fatal first MI (2%
0.142
0.159
0.151
0.101
0.165
0.189
0.140
0.157
0.149
0.100
0.163
0.188
discount rate, millions $2022)
PDV, non-fatal first IS (2%
0.058
0.065
0.062
0.043
0.067
0.076
0.057
0.064
0.061
0.043
0.066
0.075
discount rate, millions $2022)
PDV, CVD deaths (2% discount
6.387
7.169
6.790
5.089
7.341
8.023
6.226
7.009
6.629
4.928
7.181
7.862
rate, millions $2022)
PDV, total CVD benefits (2%
6.587
7.394
7.003
5.234
7.573
8.288
6.424
7.230
6.839
5.070
7.409
8.124
discount rate, millions $2022)
Annualized CVD benefits (2%
0.164
0.184
0.174
0.130
0.189
0.206
0.160
0.180
0.170
0.126
0.185
0.202
discount rate, millions $2022)
Abbreviations: PFOA - perfluorooctanoic acid; PFOS - perfluorooctane sulfonic acid; TC - total cholesterol; HDLC - high-density lipoprotein cholesterol; BP -
systolic blood pressure; CVD - cardiovascular disease; EA - economic analysis; SAB - Science Advisory Board; MI - myocardial infarction; IS - ischemic stroke;
PDV - present discounted value.
Notes:
aSee Table K-l
bSee Table K-3
cNegative values refer to increases in a particular result (e.g., the HDLC reduction of -0.002 mg/dL in Scenario 2-Dong refers to an increase in HDLC).
Final PFAS Rule Economic Analysis
K-6
April 2024
-------
FINAL RULE APRIL 2024
Table K-4: Summary of CVD Sensitivity Analysis for Hypothetical Exposure Reduction 1 (PFOA+PFOS)
Result Description3 Exposure-Response Scenario13 0
1-EA
©X
u
-J
Q
ac U
u
-J
p
&
CO
&
Cfl
1
&
CO
a.
a.
Su
a.
fl
0
Q
1
3
i
5
n
i
H J
p B
£
fl
J
<
w
1
r-
©X
c
0
Q
1
90
s
J
1
On
^ p
W X
1 T
o +
e O
P w
tH +
.3 P
3*
-t
V.©
dTotal over the period of analysis.
Final PFAS Rule Economic Analysis
K-7
April 2024
-------
FINAL RULE
APRIL 2024
K.4 Birth Weight Sensitivity Analyses
Birth weight sensitivity analyses rely on the two hypothetical exposure reductions described in
Table K-l to explore the impact of the following changes in the birth weight exposure-response
modeling:
• Early pregnancy birth weight effects using first trimester estimates from Steenland et al.
(2018) for PFOA and Dzierlenga, Crawford, and Longnecker (2020) for PFOS; and
• Inclusion of PFNA-birth weight effects using estimates from two studies (Lenters et al.,
2016; Valvi et al., 2017), in addition to the PFOA-birth weight and PFOS-birth weight
effects analyzed in the economic analysis.
Table K-5 summarizes the exposure-response scenarios, while Table K-6 provides details on the
slope factors used in this sensitivity analysis.
Table K-5: Overview of Birth Weight Exposure-Response Scenarios
Exposure-
Response Scenario Definition
Scenario
Economic analysis scenario using Steenland et al. (2018) for PFOA, Dzierlenga, Crawford,
and Longnecker (2020) for PFOS
Scenario using first trimester estimates from Steenland et al. (2018) for PFOA and
Dzierlenga, Crawford, and Longnecker (2020) for PFOS
Scenario using Steenland et al. (2018) for PFOA, Dzierlenga, Crawford, and Longnecker
(2020) for PFOS, Lenters et al. (2016) forPFNA
Scenario using Steenland et al. (2018) for PFOA, Dzierlenga, Crawford, and Longnecker
(2020) for PFOS. Valvi et al. (2017) forPFNA
Abbreviations: PFNA - perfluorononanoic acid; PFOA - periluorooctanoic acid;
PFOS - periluorooctane sulfonic acid.
1-EA
2-First Trimester
3-EA+Lenters
4-EA+Valvi
Final PFAS Rule Economic Analysis
K-8
April 2024
-------
FINAL RULE
APRIL 2024
Table K-6: Exposure-Response Information for Birth Weight
Linear Slope Estimate
Source
(g birth weight per 1 ng/mL)
Serum PFOA Serum PFOS Serum PFNA
Dzierlenga, Crawford, and Longnecker (2020)
First trimester - Steenland et al. (2018)
First trimester - Dzierlenga, Crawford, and
Longnecker (2020)
Lenters et al. (2016)
Valvi et al. (2017)
Steenland et al. (2018)
Abbreviations: CI95 - 95% confidence interval; PFNA - perfluorononanoic acid; PFOA - perfhiorooctanoic acid;
PFOS - perfhiorooctane sulfonic acid.
Table K-7 shows the results of the birth weight sensitivity analysis. The EPA made the following
observations:
• Using early pregnancy study-based dose-response estimates could reduce annualized benefits
by 66%.
• Inclusion of a 1 ppt PFNA reduction could increase annualized birth weight benefits by a
factor of 5.6 to 7.8, relative to the scenario that quantifies a 1 ppt reduction in PFOA and a 1
ppt reduction in PFOS only.
• The range of estimated PFNA-related increases in benefits is driven by the exposure-
response, with smaller estimates produced using the slope factors from Lenters et al. (2016),
followed by Valvi et al. (2017). The EPA notes that the PFNA slope factor estimates used are
orders of magnitude larger than the slope factor estimates used to evaluate the impacts of
PFOA/PFOS reductions. The EPA also notes that the PFNA slope factor estimates used are
not precise, with 95% CIs covering wide ranges that include zero (i.e., serum PFNA slope
factor estimates used are not statistically significant at 5% level).
The relative magnitudes of reductions in PFOA, PFOS, and PFNA used in this sensitivity
analysis may differ from those implied by the regulatory alternatives evaluated in the economic
analysis. Therefore, the potential magnitude of increase in the national birth weight benefits
estimates due to inclusion of PFNA effects may differ from the one estimated in this sensitivity
analyses.
Final PFAS Rule Economic Analysis K-9 April 2024
-------
FINAL RULE
APRIL 2024
Table K-7: Summary of Birth Weight Sensitivity Analysis
Hypothetical Exposure Reduction3 /
Exposure-Response Scenario13
Result Description (PFOA+PFOS) (PFOA+PFOS+PFNA)
1-EA 2-First 3-EA+Lenters 4-EA+Valvi
Trimester
Average reduction in serum PFOA
0.089
0.089
0.089
0.089
concentration (ng/mL)
Average reduction in serum PFOS
0.081
0.081
0.081
0.081
concentration (ng/mL)
Average reduction in serum PFNA
0.000
0.000
0.136
0.136
concentration (ng/mL)
Total increase in birth weight (g)
1.180
0.404
6.654
9.320
Total number of births affected0
102,268
102,268
102,268
102,268
Total number of surviving births affected0
101,804
101,803
101,806
101,808
Birth weight-related deaths (total cases
0.616
0.211
3.462
4.841
avoided)0
PDV, birth weight-related deaths (2%
3.943
1.349
22.023
30.779
discount rate, millions $2022)
PDV, birth weight-related morbidity (2%
0.117
0.040
0.656
0.918
discount rate, millions $2022)
PDV, total birth weight benefits (2%
4.061
1.389
22.679
31.697
discount rate, millions $2022)
Annualized birth weight benefits (2%
0.101
0.035
0.565
0.790
discount rate, millions $2022)
Abbreviations: PDV - present discounted value; PFNA - perfluorononanoic acid; PFOA - perfluorooctanoic acid; PFOS -
perfluorooctane sulfonic acid.
Notes: See Appendix P for results presented at 3 and 7 percent discount rates.
aSee Table K-l
bSee Table K-5
cTotal over the period of analysis.
Final PFAS Rule Economic Analysis
K-10
April 2024
-------
FINAL RULE
APRIL 2024
K.6 RCC Sensitivity Analyses
RCC sensitivity analyses rely on the first hypothetical exposure reduction described in Table K-l
to explore the impact of the following changes in the RCC exposure-response modeling:
• The use of the serum PFOA central tendency slope from Vieira et al. (2013), as derived by
the EPA (U.S. EPA, 2024c); and
• The use of the serum PFOA central tendency slopes from Vieira et al. (2013) excluding a
very high exposure group, as derived by the EPA (U.S. EPA, 2024c).
Table K-8 summarizes the exposure-response scenarios, while Table K-9 provides details on the
slope factors used in this sensitivity analysis.
Table K-8: Overview of RCC Exposure-Response Scenarios
Exposure-Response Scenario Definition3
Scenario
. . Economic analysis scenario using the serum PFOA central tendency slope from Shearer
etal. (2021)
2-Vieira Scenario using the serum PFOA central tendency slope from Vieira et al. (2013)
3-VieiraExcludeHigh
Scenario using the serum PFOA central tendency slope from Vieira et al. (2013),
excluding a very high exposure group
Abbreviations: PFOA - perfluorooctanoic acid; RCC - renal cell carcinoma.
Note:
aAll exposure-response scenarios include the 3.94% population attributable fraction (PAF)-based cap on the magnitude of
relative risk reductions, as described in Section 6.6.
Table K-9: Exposure-Response Information for RCC
Source
Linear Slope Estimate, Serum PFOA
(per 1 ng/mL)
Shearer et al. (2021), as derived by the EPA (U.S. EPA, 2024c)
0.00178
(CI95: 0.00005, 0.00352)
Vieira et al. (2013), as derived by the EPA (U.S. EPA, 2024c)
0.00007
(CI95: 0.000001, 0.00014)
Vieira et al. (2013) excluding very high exposure group from
0.00025
Vieira et al. (2013), as derived by the EPA (U.S. EPA, 2024c)
(CI95: 0.00001, 0.00048)
Abbreviations: CI95 - 95% CI; PFOA - perfluorooctanoic acid; RCC - renal cell carcinoma.
Final PFAS Rule Economic Analysis
K-ll
April 2024
-------
FINAL RULE
APRIL 2024
Table K-10 shows the results of the RCC sensitivity analysis. The EPA made the following
observations:
• Using the slope factor based on Vieira et al. (2013) could reduce annualized benefits
by 96%;
• Using the slope factor based on Vieira et al. (2013) excluding a very high exposure group
could reduce annualized benefits by 86%.
The EPA also notes that the population attributable fraction (PAF)-based cap of 3.94% on the
RCC relative risk reductions associated with a 1 ppt reduction in PFOA is rarely binding for the
economic analysis scenario presented below and never binding for the sensitivity analysis
scenarios. For larger PFOA reduction magnitudes, the PAF-based cap could become binding,
which would attenuate the differences across the sensitivity analysis scenarios.
Table K-10: Summary of RCC Sensitivity Analysis
Exposure-Response Scenario
a
Result Description
1-EA
2-Vieira
3- VieiraExcludeHigh
Average reduction in serum PFOA
0.085
0.085
0.085
concentration (ng/mL)
Non-fatal RCC (cases avoided)
9.329
0.365
1.295
RCC-related deaths (cases avoided)13
3.762
0.147
0.522
PDV, Non-fatal RCC (2% discount
2.270
0.089
0.315
rate, millions $2022)
PDV, RCC-related deaths (2%
22.477
0.878
3.118
discount rate, millions $2022)
PDV, total RCC benefits (2%
24.747
0.967
3.433
discount rate, millions $2022)
Annualized RCC benefits (2%
0.616
0.024
0.086
discount rate, millions $2022)
Abbreviations: PDV - present discounted value; PFOA - perfluorooctanoic acid; RCC - renal cell carcinoma.
Notes: See Appendix P for results presented at 3 and 7 percent discount rates.
aSee Table K-8.
bTotal over the period of analysis.
Final PFAS Rule Economic Analysis
K-12
April 2024
-------
FINAL RULE
APRIL 2024
Appendix L. Uncertainty Characterization Details
and Input Data
L.l Cost Analysis Uncertainty Characterization
In addition to occurrence uncertainty, the national cost estimates reflect two other sources of
uncertainty. The first is the total organic carbon concentration, which affects PFAS treatment
selection and is a factor for the DBP co-benefits analysis. The second is the unit cost curve
selection. The following subsections provide additional details on the EPA's approach to
modeling these sources of uncertainty.
L. 1.1 Total Organic Carbon Concentration Uncertainty
For the national cost analysis, TOC is an input to the technology selection and design equations
for granular activated carbon (GAC). Section 5.3.1.1 of the economic analysis provided a
description of how TOC affects the decision tree for technology selection. The process design
equations in Section 5.3.1.1.1 show the effect of TOC on the estimation of bed volumes for
GAC.
As noted in Section 4.3.3.2 of the economic analysis, there is no national dataset of TOC values
or ranges at PWSs. Some data are available at the system level in periodic data voluntarily
provided by primacy agencies. The EPA used the most recent data obtained in response to the
ICR for the fourth Six-Year Review of drinking water regulations. The EPA separated the
systems into two groups - those with ground water sources and those with surface water sources
- to reflect expected variations in TOC in different types of source water. Some of the systems
provided TOC values at different facilities. Facilities can include water intakes or wells,
treatment processes, and distribution system EPs. TOC levels at systems that have treatment may
differ pre- and post-treatment.
The EPA randomly assigned a TOC level to each EP from the corresponding ground water or
surface water distribution. The EPA retained that value for each of the 4,000 uncertainty
simulations. Thus, the EPA's estimates reflect TOC uncertainty across EPs, but not TOC
uncertainty interacted with PFAS uncertainty.
LI.2 Compliance Technology Unit Cost Curve Selection
Uncertainty
Each WBS model includes an input that determines whether the cost estimate generated is a low,
medium, or high cost estimate (U.S. EPA, 2024e). This input drives the selection of materials for
equipment that can be constructed of different materials. For example, a low cost system might
include fiberglass pressure vessels and PVC piping. A high cost system might include stainless
steel pressure vessels and stainless steel piping. This input also drives other model assumptions
that can affect the total cost including assumptions about building quality. High, medium, and
low quality settings affect building costs for substructure, superstructure, exterior enclosure,
interior finishes, and mechanical and electrical services.
For every technology, the EPA generated cost curves for low-, medium-, and high-cost options.
SafeWater MCBC randomly selects from these cost curves. The EPA assigned a triangular
Final PFAS Rule Economic Analysis
L-l
April 2024
-------
FINAL RULE
APRIL 2024
distribution to the cost curve selection: 25% probability for low-cost, 50% probability for
medium-cost, and 25% for high-cost.
L.2 Benefits Analysis Uncertainty Characterization
The EPA characterizes sources of uncertainty in its analysis of potential benefits resulting from
changes in PFAS levels in drinking water. The analysis reports uncertainty bounds for benefits
estimated in each category modeled for the final rule. Each lower (upper) bound value is the 5th
(95th) percentile of the category-specific benefits estimate distribution represented by 4,000
Monte Carlo draws. Table L-l provides the sources of uncertainty that the EPA quantified in the
benefits analysis that are specific to this analysis. In addition to these sources of uncertainty,
reported uncertainty bounds also reflect the following upstream sources of uncertainty: baseline
PFAS occurrence (Section 4.4 of the economic analysis), affected population size and
demographic composition (Section 4.3 of the economic analysis), and the magnitude of PFAS
concentration reduction (Section 4.4 of the economic analysis).
Table L-l: Quantified Sources of Uncertainty in Benefits Estimates
Source Description of Uncertainty
TC-serum
PFOA slope
factor; TC-
serum PFOS
slope factor3
The slope factors that express the effects of PFOA and PFOS on serum lipid markers are
based on 12 key studies with high-quality data and clearly defined PFAS-lipid level
relationships (see Appendix F). The EPA meta-analysis of these studies provides a central
estimate and a standard error estimate for the slope factors. The EPA uses a normal
distribution with a mean set at the central slope factor estimate and a standard deviation set at
the standard error estimate for the slope factor to characterize uncertainty surrounding these
parameters.
BP-serum PFOS
slope factor3
The slope factor that expresses the effects of serum PFOS on systolic BP is from Liao et al.
(2020) - a high confidence study conducted based on U.S. general population data from 2003-
2012 NHANES cycles. This study provides a central estimate and a standard error estimate
for the slope factor. The EPA uses a normal distribution with a mean set at the central slope
factor estimate and a standard deviation set at the standard error estimate for the slope factor
to characterize uncertainty surrounding this parameter.
BW-serum
PFOA slope
factor; BW-
serum PFOS
slope factor
The slope factors were obtained from meta-analyses of several studies on the subject:
Steenland et al. (2018) for PFOA and an the EPA reanalysis of Dzierlenga, Crawford, and
Longnecker (2020) for PFOS.b The meta-analyses provide a central estimate and a standard
error estimate for the slope factors. The EPA uses a normal distribution with a mean set at the
central slope factor estimate and a standard deviation set at the standard error estimate for the
slope factor to characterize uncertainty surrounding these parameters.
RCC-serum
PFOA slope
factor
The slope factor that expresses the effects of serum PFOA exposure on lifetime RCC risk is
from Shearer et al. (2021), which estimated a higher slope factor for the impact of PFOA on
RCC than previous estimates (Steenland & Woskie, 2012; Vieira et al., 2013).° This study
provides a central estimate and a standard error estimate for the slope factor. The EPA uses a
normal distribution with a mean set at the central slope factor estimate and a standard
deviation set at the standard error estimate for the slope factor to characterize uncertainty
surrounding this parameter.
Bladder cancer-
THM4 slope
factor
The slope factor that expresses the effect of co-occurring THM4 on bladder cancer is from
Regli et al. (2015), who estimated a linear slope factor relating the lifetime bladder cancer risk
associated with lifetime exposure to THM4 concentration in drinking water. This study
provides a central estimate for the slope factor. The EPA estimated a standard error for this
slope factor based on the data reported in Regli et al. (2015). The EPA uses a normal
distribution with a mean set at the central slope factor estimate and a standard deviation set at
the standard error estimate for the slope factor to characterize uncertainty surrounding this
parameter.
Final PFAS Rule Economic Analysis
L-2
April 2024
-------
FINAL RULE APRIL 2024
Table L-l: Quantified Sources of Uncertainty in Benefits Estimates
Source
Description of Uncertainty
RCC PAF to cap
risk reductions
for this endpoint
The EPA developed a central tendency estimate and an uncertainty distribution for the PAF
values to cap the relative risk estimates derived from the RCC exposure-response relationship.
Abbreviations: ASC VD -atherosclerotic cardiovascular disease; BW - birth weight; BP - blood pressure; C VD -
cardiovascular disease; PAF - population attributable fraction; PFAS - per- and polyfluoroalkyl substances; PFOA -
perfluorooctanoic acid; PFOS - perfluorooctane sulfonic acid; RCC - renal cell carcinoma; TC - total cholesterol; THM4-
four regulated trihalomethanes.
Notes:
aThe slope factors contributing to the CVD benefits analysis include the relationship between total cholesterol and PFOA
and PFOS, the relationship between high-density lipoprotein cholesterol and PFOA and PFOS, and the relationship between
blood pressure and PFOS.
bIn the original Dzierlenga, Crawford, and Longnecker (2020) estimate, the authors duplicated an estimate from Chen et al.
(2017) in the pooled estimate. The EPA reran the analysis excluding the duplicated estimate.
CA sensitivity analysis of the RCC slope factor based on alternate estimates from Vieira et al. (2013) and pooled estimates of
studies included in Shearer et al. (2021) and Vieira et al. (2013) is shown in Appendix K.
As described in Section 6.1 of the economic analysis, the EPA did not characterize the following
sources of potential uncertainty: U.S. population life tables (including standard and cause-
eliminated life tables; See Section 6.1.4 of the economic analysis), annual all-cause and health
outcome-specific mortality rates, CVD risk model (Goff et al., 2014) predictors (e.g., share of
smokers) estimated from health survey data, prevalence of CVD event history in the U.S.
population, distribution of CVD events by type, the estimated infant mortality-birth weight slope
factor (See Section 6.4.3.1 of the economic analysis), state-level distributions of infant births and
infant deaths over discrete birth weight ranges, the 200-g cap on birth weight changes estimated
under the rule, COI estimates for all modeled non-fatal health outcomes, the Value of Statistical
Life reference value, the Value of Statistical Life income elasticity value used for Value of
Statistical Life income growth adjustment, and the gross domestic product per capita projection
used to for Value of Statistical Life income growth adjustment (see Appendix J). The EPA
expects that the sources listed in Table L-l, in addition to uncertainty surrounding about
estimated PFAS occurrence, affected population size, and the magnitude of PFAS reduction,
account for the largest portion of uncertainty in the benefits analysis.
L.2.1 Exposure-Response Function Uncertainty
Table L-2 presents the central tendency estimates, 95% confidence interval bounds (2.5th and
97.5th quantile), and standard errors for the slope factors used in the EPA's assessment of
benefits resulting from the final PFAS NPDWR. This table also presents information on the
uncertainty distribution used by the EPA to characterize uncertainty for each slope factor.
Final PFAS Rule Economic Analysis
L-3
April 2024
-------
FINAL RULE APRIL 2024
Table L-2: Standard Errors and Distributions for Benefits Model Exposure-Response Slope Factors
Pollutant
Health
Benefits
Analysis
Category
Health
Outcome
Exposure-Response Slope Factor
Units
Uncertainty
Distribution
Data Source
Central
Estimate
LCB
UCB
Standard
Error
PFOA
CVD
TC
1.57
0.02
3.13
0.79
mg/dL per
ng/mL
Normal
EPA meta-analysis
based on 12 studies
(see Appendix F)
BW
BW
-10.5
-16.7
-4.4
3.14
g per ng/mL
Normal
Steenland et al.
(2018)
RCC
RCC
0.00178
0.00005
0.00352
0.00
per ng/mL
Normal
Shearer et al.
(2021)
PFOS
CVD
TC
0.08
-0.01
0.16
0.04
mg/dL per
ng/mL
Normal
EPA meta-analysis
based on 12 studies
(see Appendix F)
BP
0.044
0.006
0.083
0.02
mmHg per
ng/mL
Normal
Liao et al. (2020)
BW
BW
-3.0
-4.9
-1.1
0.97
g per ng/mL
Normal
EPA reanalysis of
Dzierlenga,
Crawford, and
Longnecker (2020)
THM4
Bladder
cancer
Bladder
cancer
0.00427
0.00331
0.00522
0.00
Per |ig/L
Normal
Reglietal. (2015)
Abbreviations: BW - birth weight; BP - blood pressure; CVD - cardiovascular disease; HDLC - high-density lipoprotein cholesterol; LCB - lower confidence bound,
2.5% quantile; PFAS - per- and polyfluoroalkyl substances; PFOA - perfluorooctanoic acid; PFOS - perfluorooctane sulfonic acid; RCC - renal cell carcinoma;
TC - total cholesterol; THM4- four regulated trihalomethanes; UCB - upper confidence bound, 97.5% quantile.
Final PFAS Rule Economic Analysis
L-4
April 2024
-------
FINAL RULE
APRIL 2024
L2.2 Population Attributable Fraction Uncertainty
As described in Section 6.6 of the economic analysis and ICF (2022), the EPA placed a PAF-
based cap on the estimated RCC risk reductions associated with changes in serum PFOA
exposure. The EPA used a log-uniform distribution (also known as reciprocal) to approximate
the distribution of PAF estimates given existing PAF estimates for other specific environmental
exposures and other specific cancers (i.e., nitrate exposure in drinking water and colon cancer).
The minimum of the distribution was set at the smallest identified PAF estimate (0.2%) and the
maximum was set at the largest identified estimated PAF (17.9%). The EPA used 3.94% (i.e., the
mean of this log-uniform distribution) as the central estimate of the PAF-based cap on the RCC
relative risk reductions.
Final PFAS Rule Economic Analysis
L-0
April 2024
-------
FINAL RULE
APRIL 2024
Appendix M. Environmental Justice
This appendix provides additional detail on the EPA's environmental justice (EJ) analysis. This
includes discussion of results from the EPA's EJ exposure analysis using the EJSCREENbatch R
package for PWS service areas in categories 4 and 5.
M.l Demographic Profile of Category 4 and 5 PWS Service
Areas
Table M-l summarizes the number of PWSs, size of PWSs, and population served for category 4
PWS service areas. There are 440 category 4 PWSs serving a population of 959,972, or 0.3% of
the overall U.S. population; 97% of category 4 PWSs are small systems, serving 883,187 people.
Table M-2 summarizes the demographic profile of category 5 PWS service areas. There are 296
category 5 PWSs serving a population of 1,104,891, or 0.3% of the overall U.S. population. 97%
percent of category 5 PWSs are small systems, serving 990,083 people.
Table M-3 summarizes the demographic profile for category 4 and 5 PWS service areas
combined and compares it to the demographic characteristics of the overall U.S. population.
Population served by category 4 and 5 PWS service areas account for 0.6% of the U.S.
population. Compared to the overall U.S. population, the population served by category 4 and 5
PWSs has lower percentages of non-Hispanic American Indian or Alaska Native, non-Hispanic
Asian, non-Hispanic Black, non-Hispanic Pacific Islander, and Hispanic populations. Category 4
and 5 PWS service areas also have a lower percentage of populations with income less than
twice the poverty level. Category 4 and 5 PWS service areas have relatively higher percentages
of non-Hispanic White populations and populations with income above twice the federal poverty
level. Among category 4 and 5 PWS service areas, there are no tribal-owned community water
systems.
Final PFAS Rule Economic Analysis
M-l
April 2024
-------
FINAL RULE APRIL 2024
Table M-l: Number of Category 4 PWSs and Population Served by Size and State
State
Number of Total
Service Areas
Number of Small
Service Areas
Total Population Population Served in
Served Small Systems3
Population Served in
Medium and Large
Systems
Missouri
37
37
88,025
88,025
New Jersey
361
347
618,244
554,259
63,985
New York
42
41
253,703
240,903
12,800
TOTAL
440
425
959,972
883,187
76,785
Abbreviation: PWS - public water system.
Note:
aSmall systems are defined as serving populations of 10,000 people or less.
Table M-2: Number of Category 5 PWSs and Population Served by Size and State
State
Number of Total Number of Small
Service Areas Service Areas
Total Population
Served
Population
Served in Small
Systems3
Population Served in
Medium and Large
Systems
Alabama
3
3
9,955
9,955
-
Colorado
24
23
94,604
83,737
10,867
Florida
1
1
25
25
-
Illinois
31
31
111,047
111,047
-
Indiana
16
16
67,129
67,129
-
Kentucky
8
8
45,099
45,099
-
Maine
14
14
43,954
43,954
-
Maryland
5
5
17,633
17,633
-
Massachusetts
23
20
127,048
93,072
33,976
Michigan
30
28
130,011
105,728
24,283
Missouri
5
5
12,599
12,599
-
New Hampshire
15
15
28,355
28,355
-
New Jersey
5
5
4,177
4,177
-
New York
45
45
104,808
104,808
-
Final PFAS Rule Economic Analysis
M-2
April 2024
-------
FINAL RULE APRIL 2024
Table M-2: Number of Category 5 PWSs and Population Served by Size and State
State
Number of Total
Service Areas
Number of Small
Service Areas
Total Population
Served
Population
Served in Small
Systems3
Population Served in
Medium and Large
Systems
North Dakota
3
3
17,035
17,035
-
Ohio
33
33
123,541
123,541
-
South Carolina
17
16
85,679
73,765
11,914
Vermont
8
8
26,784
26,784
-
Wisconsin
10
7
55,408
21,640
33,768
TOTAL
296
286
1,104,891
990,083
114,808
Abbreviation: PWS - public water system.
Note:
aSmall systems are defined as serving populations of 10,000 people or less.
Final PFAS Rule Economic Analysis
M-3
April 2024
-------
FINAL RULE
APRIL 2024
Table M-3: Population Served by Category 4 and 5 PWSs Compared to Percent of U.S. Population by Demographic Group
Race and Ethnicity
Income
Non-
Hispanic
American
Indian or
Alaska
Native
Non-Hispanic
Asian
Non-
Hispanic
Black
Non-
Hispanic
Pacific
Islander
Below Above
Non- Twice Twice
Hispanic Hispanic the the
White Poverty Poverty
Level Level
Total Population
Served
Population
Served
6,967
41,639
108,752
943
157,691
1,762,325
556,461
1,563,894
2,120,355
Percent of
Total
Population
Served
0.3%
2.0%
5.1%
0.0%
7.4%
83.1%
26.2%
73.8%
100.00%
U.S.
Population
Percent by
Demographic
Group
0.6%
5.6%
12.2%
0.2%
18.2%
60.1%
29.8%
70.2%
Percent
Difference
Between
Population
-
Served and
U.S.
Population
-0.3%
-3.6%
-7.1%
-0.2%
-10.8%
23.0%
-3.6%
3.6%
Final PFAS Rule Economic Analysis
M-4
April 2024
-------
FINAL RULE
APRIL 2024
M.2 Exposure Analysis Results
M.2.1 Baseline Scenario
Table M-4 summarizes the population served by category 4 and 5 PWS service areas with PFAS
occurrence above baseline thresholds based on a trigger level of 2 ppt for each PFAS analyte,
which is slightly above the Method 537.1 detection limits. The second set of rows in Table M-4
summarizes the percentage of the total population served by demographic group with PFAS
occurrence above these baseline thresholds. Percentages are bolded and italicized when the
percentage of the population in a specific demographic group exposed to modeled PFAS above
the baseline threshold is greater than the percentage of the total population served across all
demographic groups exposed to PFAS above this threshold (right-hand column). In Table M-4,
the highlighted numbers represent where percentages of the population served in a particular
demographic group are more than 1 percentage point greater than percentages of the total
population. Higher percentages indicate higher PFAS exposure for a given demographic group
compared to the percentage of the total population served across all demographic groups.
Notably, anticipated PFAS exposure above the baseline thresholds is higher for non-Hispanic
Asian populations across all PFAS analytes compared to the total population served across all
demographic groups. The difference in exposure is even greater when compared to non-Hispanic
White populations (28.6% vs. 14.7% for PFOS and 18.1% vs. 12.4% for PFOA). PFAS exposure
above baseline thresholds is higher for non-Hispanic Black populations for PFHxS and PFOA
and Hispanic populations for all PFAS analytes examined compared to the total population
served across all demographic groups. When compared to non-Hispanic White populations
instead of the total population served, Hispanic populations face even greater exposure (21.2%
vs. 14.7% for PFOS and 16.5% vs. 12.4% for PFOA). In addition, non-Hispanic Pacific Islander
populations have a greater percent of the population exposed to all PFAS analytes in comparison
to the total population served. The percent of non-Hispanic Pacific Islander populations exposed
to PFHpA and PFOA is at least two percentage points higher than the percent of non-Hispanic
White populations exposed to these analytes (5.2% vs. 2.9% for PFHpA and 14.8% vs. 12.4%
for PFOA). However, it should be noted that the sample size of the non-Hispanic Pacific Islander
population included in this analysis is relatively small at only 943 individuals. Exposure for non-
Hispanic American Indian or Alaska Native populations is less than or similar to exposure rates
for the total population served across all demographic groups for all PFAS analytes. PFAS
exposure above the baseline thresholds is generally lower for populations with income below
twice the Federal poverty level compared to exposure for the total population served across all
demographic groups. Populations with income above twice the Federal poverty level have
comparable but slightly higher PFAS exposure in comparison to the total population served
across all demographic groups.
Table M-5 expands on this analysis, showing average population-weighted PFAS concentrations
across demographic groups in category 4 and 5 PWSs. Cells are highlighted in yellow when the
average concentration for a given demographic group is higher than the average for the total
population served across all demographic groups. These results demonstrate again that non-
Hispanic Asian, non-Hispanic Pacific Islander, and Hispanic populations have higher average
exposure to all the PFAS analytes compared to the total population served in category 4 and 5
PWSs. Non-Hispanic American Indian or Alaska Native populations and populations with
income below twice the Federal poverty level have higher average exposures to PFHxS
Final PFAS Rule Economic Analysis
M-5
April 2024
-------
FINAL RULE APRIL 2024
compared to the total population served. Non-Hispanic Black populations have less than or
comparable average population-weighted PFAS concentrations across all four analytes in this
analysis.
Final PFAS Rule Economic Analysis
M-6
April 2024
-------
FINAL RULE
APRIL 2024
Table M-4: Baseline Scenario: Population Served by Category 4 and 5 PWS Service Areas Above Baseline Thresholds and
as a Percent of Total Population Served
Race and Ethnicity Income
PFAS
Non-Hispanic
American Indian or
Alaska Native
Non-
Hispanic
Asian
Non-
Hispanic
Black
Non-Hispanic
Pacific
Islander
Hispanic
Non-
Hispanic
White
Below Twice
the Poverty
Level
Above Twice
the Poverty
Level
population
Served
Population Served Above Baseline Threshold
PFOS
552
11,915
16,861
155
33,499
259,771
64,755
263,942
328,697
PFHxS
225
3,322
6,810
56
13,865
89,308
27,740
88,585
116,325
PFHpA
69
3,328
1,399
49
8,725
50,630
9,061
56,760
65,821
PFOA
590
7,545
14,455
140
25,948
217,734
63,857
207,811
271,668
Population Served Above Baseline Threshold as a Percent of Total Population Served
PFOS
7.9%
28.6%
15.5%
16.4%
21.2%
14.7%
11.6%
16.9%
15.5%
PFHxS
3.2%
8.0%
6.3%
5.9%
8.8%
5.1%
5.0%
5.7%
5.5%
PFHpA
1.0%
8.0%
1.3%
5.2%
5.5%
2.9%
1.6%
3.6%
3.1%
PFOA
8.5%
18.1%
13.3%
14.8%,
16.5%
12.4%
11.5%
13.3%
12.8%
Abbreviations: PFHpA - Perfluoroheptanoic acid; PFHxS - Perfluorohexanesulfonic acid; PFOA - Perfluorooctanoic Acid; PFOS - Perfluorooctanesulfonic Acid.
Final PFAS Rule Economic Analysis
M-7
April 2024
-------
FINAL RULE APRIL 2024
Table M-5: Average PFAS Concentrations (ppt) by Demographic Group in the Baseline, Category 4 and 5 PWS Service
Areas
Race and Ethnicity T
•7 I n en m p
PFAS
Non-
Hispanic
American
Indian or
Alaska
Native
Non-
Hispanic
Asian
Non-
Hispanic
Black
Non-
Hispanic
Pacific
Islander
Hispanic
Non-
Hispanic
White
Below
Twice the
Poverty
Level
Above
Twice the
Poverty
Level
Total
Population
Served
PFOS
0.44
2.32
1.07
1.34
1.53
1.04
0.77
1.22
1.10
PFHxS
0.79
0.79
0.53
2.58
1.45
0.51
0.66
0.60
0.62
PFHpA
0.14
0.52
0.16
0.44
0.41
0.26
0.20
0.30
0.28
PFOA
0.74
1.31
0.95
2.05
1.59
0.91
0.94
1.00
0.99
Abbreviations: PFHpA - periluoroheptaiioic acid; PFHxS - periluorohexanesulfonic acid; PFOA - periluorooctanoic acid; PFOS - perfluorooctanesulfonic acid.
Final PFAS Rule Economic Analysis
M-8
April 2024
-------
FINAL RULE
APRIL 2024
M.2.2 Hypothetical Regulatory Scenario #1: UCMR 5 MRLs
Table M-6 summarizes the results for populations served by category 4 and 5 PWS service areas
with PFAS occurrence above UCMR 5 MRL values. For this hypothetical regulatory scenario,
the EPA assumed that PWSs with PFAS system-level means above the MRL value will reduce
PFAS levels to comply with the final rule. The first set of rows in Table M-6 summarizes
population served by category 4 and 5 PWS service areas with PFAS occurrence above the
UCMR 5 MRLs. The second set of rows provides these estimates as a percentage of the total
population served by PWS service areas included in the EPA's analysis.
Percentages are bolded and italicized when the percentage of the population in a specific
demographic group with PFAS occurrence above the MRL is greater than the percentage of the
total population served across all demographic groups with PFAS occurrence above the MRL
(right-hand column). In Table M-6, the highlighted numbers represent where percentages of the
population served in a particular demographic group are more than 1 percentage point greater
than percentages of the total population. Under this hypothetical regulatory scenario, where
MCLs are assumed to be equal to UCMR 5 MRL values, these populations would be expected to
experience reductions in PFAS exposure to below the hypothetical regulatory thresholds.
The EPA's EJ exposure analysis shows that anticipated PFAS exposure above the UCMR 5
MRL values at category 4 and 5 systems is higher for non-Hispanic Asian, non-Hispanic Black,
and Hispanic populations for almost all PFAS analytes (the exception being exposure to PFHpA
for non-Hispanic Black populations) compared to occurrence over the MRL for the total
population served across all demographic groups. Exposure to PFOS and PFHpA is also higher
for non-Hispanic Pacific Islander populations in comparison to the total population served. PFAS
exposures above the UCMR 5 MRL values for non-Hispanic Asian populations are the highest of
any demographic group for several PFAS analytes, with PFOA, PFOS, and PFHpA exposure in
particular being roughly twice the exposure rate for the total population served across all
demographic groups. The percent of non-Hispanic American Indian or Alaska Native
populations with exposure above the UCMR 5 MRL values is generally somewhat lower in
comparison to the exposure rate for the total population served across all demographic groups.
Similarly, a lower percent of populations with income below twice the Federal poverty level
have PFAS exposure above the UCMR 5 MRLs values compared to the total population served
across all demographic groups.
Table M-7 presents average population-weighted PFAS reductions across demographic groups in
category 4 and 5 PWSs under a hypothetical regulatory scenario where system-level means are
reduced to UCMR 5 MRL values. Cells are highlighted when the average concentration for a
given demographic group is higher than the average for the total population served across all
demographic groups. Reductions in all PFAS analytes to UCMR 5 MRL values are larger for
Hispanic, non-Hispanic Asian, and non-Hispanic Pacific Islander populations than the total
population served across all demographic groups. Non-Hispanic American Indian or Native
Alaska populations see larger reductions in PFHxS, while populations with income below twice
the Federal poverty level see larger reductions of PFHxS and PFOA compared to the total
population served. Non-Hispanic Pacific Islander populations see the greatest reductions in
PFOA, PFHpA, and PFHxS of any demographic group.
Final PFAS Rule Economic Analysis
M-9
April 2024
-------
FINAL RULE
APRIL 2024
M.2.3 Hypothetical Regulatory Scenario #2:10.0 ppt
Table M-8 summarizes the results of the population served by category 4 and 5 PWS service
areas with PFAS occurrence above 10.0 ppt. For this hypothetical regulatory scenario, the EPA
assumed that PWSs with PFAS system-level means above 10.0 ppt will reduce PFAS levels to
comply with the final rule. Percentages are bolded and italicized when the percentage of the
population in a specific demographic group with PFAS occurrence above 10.0 ppt is greater than
the percentage of the total population served across all demographic groups with PFAS
occurrence above 10.0 ppt (right-hand column). In Table M-8, the highlighted numbers represent
where percentages of the population served in a particular demographic group are more than 1
percentage point greater than percentages of the total population. Under this hypothetical
regulatory scenario, where MCLs are assumed to be equal to 10.0 ppt, these populations would
be expected to experience reductions in PFAS exposure to below the hypothetical regulatory
thresholds.
The EPA's EJ exposure analysis shows that anticipated PFAS exposure above 10.0 ppt is higher
for non-Hispanic Asian, Hispanic, and non-Hispanic Pacific Islander populations for particular
PFAS analytes when compared to exposure for the total population served across all
demographic groups. Specifically, PFAS exposure above 10.0 ppt is higher for Hispanic
populations for PFOA, PFHxS, and PFOS compared to the total population served. Exceedances
of 10.0 ppt for non-Hispanic Asian populations are the highest of any demographic group, with
PFOS exposure in particular being roughly three times the exposure rate for the total population
served across all demographic groups. Exposure to PFOS, PFHxS, and PFOA over 10.0 ppt is
substantially higher for non-Hispanic Pacific Islander populations in comparison to the total
population, with PFHxS occurrence nearly six times the levels observed in the total population
served by category 4 and 5 systems. However, the sample size of Pacific Islander populations is
relatively small, and so these differences in population percentages reflect no more than 50
individuals. PFAS exposure above 10.0 ppt is similar or somewhat lower for other populations
compared to the exposure rate for the total population served across all demographic groups.
Table M-9 presents average population-weighted PFAS reductions across demographic groups in
category 4 and 5 PWSs under a hypothetical regulatory scenario where system-level means are
reduced to 10.0 ppt. Cells are highlighted when the average concentration for a given
demographic group is higher than the average for the total population served across all
demographic groups. Table M-9 shows that reductions are higher for non-Hispanic Pacific
Islander and Hispanic populations for all three PFAS analytes for which there are exposures
above 10.0 ppt in the sample of category 4 and 5 PWSs (PFOA, PFOS, and PFHxS). Reductions
for these population groups are highest for PFHxS and PFOA. For instance, reductions in PFHxS
and PFOA exposure for non-Hispanic Pacific Islander populations are roughly 10 and four times
the exposure rate for the total population served across all demographic groups, respectively
(2.09 vs. 0.23 for PFHxS and 1.12 vs. 0.25 for PFOA). Non-Hispanic American Indian or Alaska
Native populations served also see greater reductions in PFHxS and PFOA in comparison to the
total population served. Reductions PFAS exposure above 10.0 ppt are also higher for non-
Hispanic Asian populations for PFOS and for populations with income below twice the Federal
poverty level for PFHxS and PFOA.
Final PFAS Rule Economic Analysis
M-10
April 2024
-------
FINAL RULE
APRIL 2024
Table M-6: Hypothetical Regulatory Scenario #1: Demographic Breakdown of Population Served by Category 4 and 5 PWS
Service Areas Above UCMR 5 MRL and as a Percent of Total Population Served
Race and Ethnicity Income
PFAS
Non-Hispanic
American
Indian and
Alaska Native
Non-Hispanic
Asian
Non-Hispanic
Black
Non-Hispanic
Pacific
Islander
Non-Hispanic
Hispanic White
Below
Twice the
Poverty
Level
Above
Twice the
Poverty
Level
Population
Served
Population Served Above UCMR 5 MRL
PFOS
176
9067
13162
116
20464
169469
41499
175155
216,654
PFHxS
209
2814
6172
29
11266
58040
19574
61082
80,656
PFHpA
53
2314
676
29
6077
26539
5263
31685
36,948
PFOA
238
5703
10018
58
17925
113903
32348
118765
151,113
Population Served Above UCMR 5 MRL as a Percent of Total Population Served
PFOS
2.5%
21.8%
12.1%
12.3%
13.0%
9.6%
7.5%
11.2%
10.2%
PFHxS
3.0%
6.8%
5.7%
3.1%
7.1%
3.3%
3.5%
3.9%
3.8%
PFHpA
0.8%
5.6%
0.6%
3.1%
3.9%
1.5%
0.9%
2.0%
1.7%
PFOA
3.4%
13.7%
9.2%
6.2%
11.4%
6.5%
5.8%
7.6%
7.1%
Abbreviations: PFHpA - Perfluoroheptanoic acid; PFHxS - Perfluorohexanesulfonic acid; PFOA - Perfluorooctanoic Acid PFOS - Perfluorooctanesulfonic Acid.
Final PFAS Rule Economic Analysis
M-ll
April 2024
-------
FINAL RULE
APRIL 2024
Table M-7: Reductions in Average PFAS Concentrations (ppt) by Demographic Group in a Hypothetical Regulatory
Scenario with Maximum Contaminant Level at the UCMR 5 MRLs, Category 4 and 5 PWS Service Areas
Race and Ethnicity WnmP
PFAS
Non-
Hispanic
American
Indian or
Alaska
Native
Non-
Hispanic
Asian
Non-
Hispanic
Black
Non-
Hispanic
Pacific
Islander
Hispanic
Non-
Hispanic
White
Below
Twice the
Poverty
Level
Above
Twice the
Poverty
Level
Total
Population
Served
PFOS
0.13
1.14
0.27
0.71
0.61
0.38
0.23
0.47
0.41
PFHxS
0.59
0.36
0.21
2.27
1.01
0.23
0.40
0.28
0.32
PFHpA
0.03
0.08
0.02
0.13
0.09
0.03
0.03
0.04
0.03
PFOA
0.39
0.49
0.37
1.40
0.85
0.37
0.47
0.40
0.42
Abbreviations: PFHpA - periluoroheptanoic acid; PFHxS - periluorohexanesulfonic acid; PFOA - periluorooctanoic acid; PFOS - perfluorooctanesulfonic acid.
Final PFAS Rule Economic Analysis
M-12
April 2024
-------
FINAL RULE
APRIL 2024
Table M-8: Hypothetical Regulatory Scenario #2: Demographic Breakdown of Population Served by Category 4 and 5 PWS
Service Areas Above 10.0 ppt and as a Percent of Total Population Served
PFAS
Race and Ethnicity
Income
Non-Hispanic
American
Indian and
Alaska Native
Non-Hispanic
Asian
Non-Hispanic
Black
Non-Hispanic
Pacific
Islander
Hispanic
Non-Hispanic
White
Below
Twice the
Poverty
Level
Above
Twice the
Poverty
Level
Population
Served
Population Served Above 10.0 ppt
PFOS
59
2,397
891
49
5,396
32,632
5,465
37,385
42,850
PFHxS
59
162
494
29
1,997
7,662
4,306
6,879
11,185
PFHpA
0
0
0
0
0
0
0
0
0
PFOA
66
553
1,027
36
4,269
20,842
6,824
20,906
27,730
Population Served Above 10.0 ppt as a Percent of Total Population Served
PFOS
0.8%
5.8%
0.8%
5.2%
3.4%
1.9%
1.0%
2.4%
2.0%
PFHxS
0.8%
0.4%
0.5%
3.1%
1.3%
0.4%
0.8%
0.4%
0.5%
PFHpA
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
PFOA
0.9%
1.3%
0.9%
3.8%
2.7%
1.2%
1.2%
1.3%
1.3%
Abbreviations: PFHpA - Perfluoroheptanoic acid; PFHxS - Perfluorohexanesulfonic acid; PFOA - Periluorooctanoic Acid PFOS - Perfluorooctanesulfonic Acid.
Final PFAS Rule Economic Analysis
M-13
April 2024
-------
FINAL RULE
APRIL 2024
Table M-9: Reductions in Average PFAS Concentrations (ppt) by Demographic Group in a Hypothetical Regulatory
Scenario with Maximum Contaminant Level at 10.0 ppt, Category 4 and 5 PWS Service Areas
Race and Ethnicity Income
PFAS
Non-
Hispanic
American
Indian or
Alaska
Native
Non-
Hispanic
Asian
Non-
Hispanic
Black
Non-
Hispanic
Pacific
Islander
Hispanic
Non-
Hispanic
White
Below
Twice the
Poverty
Level
Above
Twice the
Poverty
Level
Total
Population
Served
PFOS
0.04
0.31
0.04
0.25
0.17
0.09
0.04
0.11
0.10
PFHxS
0.51
0.21
0.17
2.09
0.81
0.16
0.32
0.2
0.23
PFHpA
0
0
0
0
0
0
0
0
0
PFOA
0.32
0.2
0.2
1.12
0.56
0.21
0.32
0.22
0.25
Abbreviations: PFHpA - periluoroheptanoic acid; PFHxS - periluorohexanesulfonic acid; PFOA - periluorooctanoic acid; PFOS - perfluorooctanesulfonic acid.
Final PFAS Rule Economic Analysis
M-14
April 2024
-------
FINAL RULE
APRIL 2024
Appendix N. Supplemental Cost Analyses
Section N.l discusses the approach the EPA used to estimate the costs of the rule for PWSs
serving more than 1 million people. Section N.2 discusses the potential impact on national costs
if PWSs must dispose of treatment residuals as hazardous waste. Section N.3 explores the
potential impact of PFNA, perfluorobutane sulfonic acid (PFBS), HFPO-DA occurrence data on
national cost estimates.
N.l Cost Analysis for Very Large Systems
The EPA identified 25 PWS that serve more than one million people based on retail population
estimates in SDWIS/Fed. All of these systems are CWS with multiple EPs; most are surface
water systems (see Table N-l).
Table N-l: Characteristics of PWSs Serving a Retail Population Greater than One
Million
PWSID
Name
SDWIS/Fed
Water
Entry
Retail
Source
Points
Population
AZ0407025
Phoenix, City Of
1,579,000
SW
20
CAO110005
East Bay Municipal Utility District
1,405,000
SW
5
CA1910067
Los Angeles-City, Dept. Of Water & Power
4,041,284
SW
11
CA3710020
San Diego - City Of
1,394,515
SW
3
CA4310011
San Jose Water
1,007,514
SW
3
COOl 16001
Denver Water Board
1,362,071
SW
3
FL4130871
Miami-Dade Water and Sewer Department - Main System
2,300,000
GW
3
GA1210001
Atlanta
1,089,893
SW
2
IL0316000
Chicago
2,700,000
SW
2
MA6000000
Massachusetts Water Resources Authority
2,550,000
SW
2
MD0150005
Washington Suburban Sanitary Commission
1,800,000
SW
2
MD0300002
Baltimore City
1,600,000
SW
3
M06010716
Missouri American St Louis County St Charles County
1,100,000
SW
4
NC0160010
Charlotte Water
1,093,901
SW
2
NV0000090
Las Vegas Valley Water District
1,502,604
SW
10
NY5110526
Suffolk County Water Authority
1,100,000
GW
236
NY7003493
New York City System
8,271,000
SW
4
OH1801212
Cleveland Public Water System
1,308,955
SW
4
OH2504412
Columbus Public Water System
1,233,879
SW
3
PA1510001
Philadelphia Water Department
1,600,000
SW
3
TX0150018
San Antonio Water System
1,999,472
SW
38
TX0570004
Dallas Water Utility
1,286,380
SW
3
TX1010013
City of Houston
2,221,706
SW
41
TX2270001
City of Austin Water & Wastewater
1,044,405
SW
3
VA6059501
Fairfax County Water Authority
1,074,422
SW
2
Abbreviations: GW - ground water; PWS - public water system; PWSID - public water system identification; SDWIS/Fed -
Safe Drinking Water Information System Federal Data Warehouse; SW - surface water.
Rather than model treatment costs using the PFAS occurrence values simulated from the MCMC
model, the EPA reviewed UCMR3 data and recent system consumer confidence reports to obtain
EP PFAS values. Given the type of sources used there were not enough data to confidently
estimate running annual averages (RAA). As a result, the EPA used these values to determine
which EPs at these systems exceed the MCLs and/or HI for the final rule and alternative options.
Final PFAS Rule Economic Analysis
N-l
April 2024
-------
FINAL RULE
APRIL 2024
Any value reported above relevant limit was interpreted as an exceedance. This approach likely
overestimates treatment costs at a national level, since maximum individual reported values are
typically higher than RAAs. For example, if a system were to observe four values with one
PFOA result at 6 ppt and three PFOA results at 3 ppt, for purposes of calculating an RAA, this
EP would be below the PFOA MCL and would not be compelled to take action. However, using
the methodology here, because the single value is above the MCL, the EPA treated that EP as
needing to take action such as installing treatment.
PFOA and PFOS levels at multiple EPs for two systems exceeded one or more MCLs for the
final rule and alternative options (no HI exceedances occurred). The EPA used these reported
PFAS values as the baseline occurrence estimates for the cost analysis. The EPA applied the cost
estimating methods described in Chapter 5 to these systems to derive estimates of the costs to
meet each MCL.
N.2 Hazardous Waste Disposal Cost Impacts
The national cost analysis reflects the assumption that PFAS-contaminated wastes are not
considered RCRA regulatory or characteristic hazardous wastes. Stakeholders have expressed
concern to the EPA that a hazardous substance designation for certain PFAS may limit their
disposal options for drinking water treatment residuals (e.g., spent media, concentrated waste
streams) and/or potentially increase costs. Designation of PFOA and PFOS as CERCLA
hazardous substances would not require waste (e.g., biosolids, treatment residuals, etc.) to be
treated in any particular fashion, nor disposed of at any specific particular type of landfill. The
designation also would not restrict, change, or recommend any specific activity or type of waste
at landfills. Although designating chemicals as hazardous substances under CERCLA would not
result in new requirements for disposal of PFAS drinking water treatment residuals, to address
stakeholder concerns, including those raised during the SBREFA process, the EPA conducted a
sensitivity analysis with an assumption of hazardous waste disposal for illustrative purposes
only. As part of this analysis, the EPA generated a second full set of unit cost curves that are
identical to the curves used for the national cost analysis with the exception that spent GAC and
spent IX resin are considered hazardous. The EPA acknowledges that if PF AS-contaminated
wastes are required to be handled as hazardous wastes, the residuals management costs are
expected to be higher.
For GAC, the national cost analysis assumes the spent media is reactivated off-site under current
RCRA non-hazardous waste regulations. Under this scenario, the WBS model uses a unit cost for
reactivation that includes transportation to the reactivation facility and back to the treatment
plant. To account for losses in the reactivation and replacement process, it also adds the cost of
replacing 30 percent of the spent GAC with virgin media. The hazardous waste sensitivity
analysis assumes spent GAC is disposed off-site as a hazardous waste in a RCRA Subtitle C
landfill and replaced with virgin GAC (i.e., single use operation). Under this scenario, the WBS
model incorporates the cost of hazardous waste disposal, transportation to a hazardous waste
facility 200 miles away, a minimum charge per hazardous waste shipment, and replacement of
100 percent of the spent GAC with virgin media. This scenario provides an upper bound on other
options that might emerge under future air quality regulations that prevent reactivation of PF AS-
contaminated GAC (i.e., spent GAC must be disposed off-site as a non-hazardous waste and
replaced with virgin GAC) or RCRA hazardous waste regulations (i.e., off-site reactivation
remains feasible, but process wastes require hazardous waste disposal).
Final PFAS Rule Economic Analysis
N-2
April 2024
-------
FINAL RULE
APRIL 2024
For IX, the national cost analysis assumes the spent resin is incinerated off-site under current
RCRA non-hazardous waste regulations. Under this scenario, the WBS model uses a unit cost for
non-hazardous incineration that includes transportation to the incineration facility. The
hazardous waste sensitivity analysis assumes spent resin is incinerated off-site as a hazardous
waste and replaced with virgin resin. Under this scenario, the WBS model incorporates the cost
of hazardous waste incineration, transportation to a hazardous waste facility 200 miles away, and
a minimum charge per hazardous waste shipment. Both scenarios incorporate the cost of
replacing the spent resin with virgin resin. Because hazardous waste incineration costs more than
disposal of spent resin in a hazardous waste landfill this hazardous waste scenario provides an
upper bound on other options that might emerge under future air quality regulations (e.g., off-site
disposal in a non-hazardous waste landfill) or RCRA hazardous waste regulations (e.g., off-site
disposal in a hazardous waste landfill).
The potential impact on PWS treatment costs is shown in Table N-2 for the final rule. At a 2
percent discount rate, the annualized cost would be $98.90 million (7%) higher if hazardous
waste disposal is required. Note that these estimated costs do not include the costs associated
with the storage, transportation and underground injection of the brine concentrate residuals from
the RO/NF process that could possibly be required under a PFAS hazardous waste scenario.
Table N-2: Annualized PWS Treatment Cost Associated with Non-Hazardous and
Hazardous Residual Management Requirements, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, and HFPO-DA MCLs of 10 ppt each and HI of 1) (Million
$2022)
2% Discount Rate
5th Percentile
Mean
95th Percentile
Non-Hazardous Disposal
$1,395.23
$1,506.44
$1,627.65
Hazardous Disposal
$1,487.73
$1,605.34
$1,731.75
Increase due to Hazardous Disposal
$98.90
Note: Percentiles cannot be subtracted. See Appendix P for results presented at 3 and 7 percent discount rates.
Final PFAS Rule Economic Analysis
N-3
April 2024
-------
FINAL RULE
APRIL 2024
N.3 National Level Sensitivity Analysis of Incremental Treatment
Cost of PFNA, PFBS and HFPO-DA
The EPA has estimated the national level costs of the final rule using occurrence data for PFOA,
PFOS and PFHxS. As discussed in Chapter 4 of the EA, there are limitations with nationally
representative occurrence information for the other compounds in the final rule (PFNA, HFPO-
DA and PFBS), therefore the additional treatment costs associated with the occurrence of PFNA,
HFPO-DA, PFBS, are not reported in the national cost estimates. Instead, quantified cost
estimates for PFNA, HFPO-DA, and PFBS are considered here as part of this sensitivity
analysis. When available, nationally representative occurrence information is preferable for an
economic analysis of the national level costs and benefits. However, this does not mean that non-
nationally representative occurrence data cannot be used to meaningfully inform regulatory
development of drinking water standards and they often represent the best available science and
information.
In the case of PFOA, PFOS, and PFHxS, the EPA has a sufficiently robust nationally
representative dataset from UCMR3. UCMR3 required all large community and non-transient
non-community water systems serving more than 10,000 people to monitor and also required
monitoring by a nationally representative sample of small systems (i.e., those serving 10,000 or
fewer people). The survey sample design for small systems uses a statistically-derived set of
systems for the nationally representative sample that is population-weighted within each system
size and source water category so that any PWS within a category has an equivalent likelihood of
selection (77 FR 26072). The EPA used additional state data that were available at systems that
were part of this UCMR3 set of systems to fit the MCMC occurrence model that informed cost
estimates for PFOA, PFOS, and PFHxS. When incorporating the additional state data, the EPA
used QC measures to ensure that the data represented finished drinking water, to verify that the
majority of data were analyzed using EPA approved drinking water methods60 and that the set of
systems used to inform the model maintained the nationally representative structure. Further
details on the MCMC model are available in Cadwallader et al. (2022). For more information on
the application of the model in this analysis, see Section 4.4 and Appendix A. For more
information on the data and analyses that the EPA used to develop national estimates of PFAS
occurrence in public drinking water systems see U.S. EPA (2024a).
In the case of PFNA, HFPO-DA, and PFBS, EPA lacks the same level of precision as described
above. While PFNA and PFBS were included in UCMR3, the amount of results above the
UCMR3 MRLs was insufficient for incorporation into the MCMC occurrence model and
prevented direct quantification through model extrapolation. However, a substantial amount of
data (about 36,000 samples from 10,000 systems or more per contaminant) were collected from
states. These state data also underwent QC measures to ensure that the data represented finished
drinking water and to verify that the majority of data were analyzed using EPA approved
drinking water methods.60 While the state-led data collection efforts provided valuable
information about occurrence for PFNA, HFPO-DA, and PFBS, they did not provide the
nationally representative foundation provided by UCMR 3 for PFOA, PFOS, and PFHxS to be
incorporated into the MCMC model. Therefore, because there is somewhat greater uncertainty in
o0 The EPA was able to verify that approximately 97% of the state data were analyzed using EPA approved methods.
Final PFAS Rule Economic Analysis
N-4
April 2024
-------
FINAL RULE
APRIL 2024
the number of systems that are likely to exceed the MCLs, the quantified cost estimates for
PFNA, HFPO-DA, and PFBS are discussed in the context of this sensitivity analysis. For HFPO-
DA, PFBS, and PFNA, the EPA extrapolated system level maximums from non-targeted state
datasets as part of a conservative approach to estimate occurrence for PFAS without a nationally
representative dataset. EPA presents these cost results separately from the results for PFOA,
PFOS, and PFHxS to recognize the higher level of uncertainty associated with the occurrence of
PFNA, HFPO-DA and PFBS and the different approaches taken to derive occurrence estimates.
In the EA for the proposed PFAS NPDWR, the EPA used a model system approach to illustrate
the potential incremental costs for removing PFAS not included in the national economic model.
After considering public comments on the incremental cost analysis, the EPA decided to further
explore the incremental costs associated with the HI and MCLs with a national level sensitivity
analysis in the final rule.
To inform this sensitivity analysis, the EPA estimated the occurrence of HFPO-DA, PFBS, and
PFNA, using available state-level data. The EPA then used these estimates to determine the
potential impact of exceedance of the HI (mixtures of two or more of PFHxS, PFNA, HFPO-DA,
and PFBS) and individual PFNA and HFPO-DA MCLs in addition to exceedances of the PFOA,
PFOS and PFHxS MCLs. For more information on the occurrence model output used in this
sensitivity analysis, including its development and results, See Section 10.3.2. of Per- and
Polyfluoroalkyl Substances (PFAS) Occurrence & Contaminant Background Support Document
(U.S. EPA, 2024d).
This sensitivity analysis has two major limitations that are important to note. They are:
1. The occurrence data for HFPO-DA, PFBS, and PFNA are modeled using limited available
aggregated state-level data that is extrapolated to the nation. Specifically, HFPO-DA does
not currently have a completed nationally representative dataset while PFNA and PFBS were
not included in the national occurrence model because of the limited reported values above
the minimum reporting levels in UCMR 3. As described in the Technical Support Document
for PFAS Occurrence and Contaminant Background Chapter 10.3, non-targeted state
monitoring datasets were used for extrapolation of PFNA, HFPO-DA, and PFBS in lieu of a
nationally representative dataset.
2. The EPA has insufficient quantitative data to include HFPO-DA in the linear equations used
to estimate bed life for IX. In this analysis, the EPA assumes the bed life is the same as
PFHxA, the contaminant for which quantitative data are available that is the most difficult to
remove by IX. The EPA has insufficient quantitative data to include PFNA in the linear
equations used to estimate bed life for GAC and IX. For GAC, the EPA assumes the bed life
for PFNA is the same as PFOS. For IX, the EPA assumes the bed life is the same as PFOA.
Given the chain length of PFNA, these assumptions likely underestimate the actual bed life
and will result in the EPA estimating higher costs than will actually be realized for this part
of the estimate.
When the modeled occurrence data for PFNA, HFPO-DA, PFBS is incorporated into the
SafeWater MCBC model, the estimated number of EPs exceeding one or more MCLs, and
therefore required to treat or use a different water source, increases to 9,471 from 9,043. This
results in an increase in the expected national costs. Under the primary analyses (see Chapter 5)
the expected total national cost at a 2 percent discount rate is $1,548.64 million. Under the
Final PFAS Rule Economic Analysis
N-5
April 2024
-------
FINAL RULE
APRIL 2024
sensitivity analysis, the expected national costs increase to $1,631.05 million, or approximately a
5 percent increase in national costs. Broken out by system size, expected national rule costs
increase from $275.84 million to $293.09 million (6 percent increase) and $1,272.83 million to
$1,337.93 million (5 percent increase) for small and large systems, respectively. This small
increase in costs would not change the Administrator's determination at proposal that the EPA is
reaffirming for the final rule that the benefits of the rule justify its costs.
N.4 National Level Sensitivity Analysis Considering PFNA and
HFPO-DA MCLs
The final rule consists of PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA
MCLs of 10 ppt each and an HI MCL of 1 (unitless). To evaluate the costs of the rulemaking in
the absence of the HI MCL, the EPA estimated the cost of an MCL only scenario which included
only the PFOA and PFOS MCLs of 4.0 ppt each, and the PFHxS, PFNA, HFPO-DA MCLs of 10
ppt each. The EPA then examined the marginal costs of two individual contaminant MCLs (i.e.,
PFNA, and HFPO-DA) using the MCL only scenario as the base cost. As discussed in Section
N.3 above and Section 10.3 of Per- and Polyfluoroalkyl Substances (PFAS) Occurrence &
Contaminant Background Support Document (U.S. EPA, 2024d) the total estimated annualized
cost for the MCL only rule was estimated using combined information from the national
occurrence model (PFOA, PFOS, and PFHxS) and state level occurrence information for HFPO-
DA and PFNA. For the HFPO-DA and PFNA MCLs the EPA estimated system level maximums
using several methods due to uncertainty as part of a conservative approach to estimate
occurrence for PFAS without a nationally representative dataset. Results presented below use the
method that selects equal percentages of systems among systems that a) are already exceeding an
MCL for PFOA or PFOS and b) are not exceeding an MCL for PFOA or PFOS. Within a group,
the probability of being selected is proportionate to the system's maximum sum of modeled
PFAS. Therefore, these cost estimates can be considered cost conservative. In Section 5.1.3, the
EPA discusses the marginal costs associated with PFHxS MCL exceedances. Computationally
the EPA first modeled the costs of the MCL only scenario considering the individual MCLs for
PFOA (4.0 ppt), PFOS (4.0 ppt), PFNA (10 ppt), HFPO-DA (10 ppt), and PFHxS (10 ppt). The
estimated mean total annualized MCL only scenario cost is $1,545.35 ($2022, 2 percent discount
rate). The EPA then modeled the costs of the rule without the MCLs for PFNA and HFPO-DA
one at a time. The difference between the cost for all analyte MCLs and the cost for all analyte
MCLs except the one removed from the model is the marginal costs of the removed MCL. Table
N-3 shows the marginal costs of the rulemaking associated with MCLs for PFNA and HFPO-
DA.
Table N-3. Marginal Mean Annualized Rule Costs Associated with Individual MCLs of
10 ppt each for PFNA, HFPO-DA (Million $2022)
PFNA $40.45
HFPO-DA $14.87
The PFNA MCL is estimated to affect 208 PWSs (393 EPs), 191 PWSs (346 EPs) of which need
to take corrective action for PFNA alone and 17 PWSs (46 EPs) will take corrective action due
to more than one PFAS MCL. The HFPO-DA MCL is estimated to affect 44 PWSs (84 EPs), 40
Final PFAS Rule Economic Analysis
N-6
April 2024
-------
FINAL RULE
APRIL 2024
PWSs (73 EPs) of which need to take corrective action for HFPO-DA alone, and 4 PWSs (11
EPs) will take corrective action due to more than one PFAS MCL. As demonstrated by these
results, the EPA expects that the more unique systems triggered into corrective action by a given
MCL, the higher its marginal cost will be.
Considering the MCL only scenario, total annualized costs of $1,545.35 million ($2022, 2
percent discount rate), the PFNA MCL contributes 2.6 percent of the overall costs and the
HFPO-DA MCL contributes 0.9 percent of the overall costs.
Final PFAS Rule Economic Analysis
N-7
April 2024
-------
FINAL RULE
APRIL 2024
Appendix O. Supplemental Benefits Analyses
O.lSupplemental Liver Cancer Analysis
This section presents an analysis that considers potential changes in liver cancer cases and deaths
associated with reduced exposures to PFOS considered under the final rule. This analysis is
presented as a supplemental analysis for the final rule to respond to public comments received on
the proposed rule requesting that the EPA quantify additional health benefits.
0.1.1.Overview of the Liver Cancer Risk Reduction Analysis
Figure 0-1 illustrates the approach used to quantify and value the changes in liver cancer risk
associated with decreased serum PFOS levels from reductions in drinking water PFOS
concentrations under the regulatory alternatives. Section 4.4 and Section 6.3 detail the PWS EP-
specific PFOS drinking water occurrence estimation and modeling of serum PFOS
concentrations, respectively. PWS EP-specific time series of the differences between serum
PFOS concentrations under baseline and regulatory alternatives are inputs into this analysis. For
each PWS EP, evaluation of the changes in liver cancer impacts involves the following key
steps:
1. Estimating the changes in liver cancer risk based on modeled changes in serum PFOS levels
and the exposure-response function for the effect of serum PFOS on liver cancer;
2. Estimating the annual incidence of liver cancer cases and excess mortality among those with
liver cancer in all populations corresponding to baseline and regulatory alternative liver
cancer risk levels, as well as estimating the regulatory alternative-specific reduction in cases
relative to the baseline; and
3. Estimating the economic value of reducing liver cancer mortality and morbidity from
baseline to regulatory alternative levels, using the Value of Statistical Life and willingness to
pay measures, respectively.
Section 0.1.2 discusses the exposure-response modeling for liver cancer. Section 0.1.3
summarizes the life table-based approach for estimation of liver cancer risk reductions. Section
0.1.4 discusses the EPA's valuation methodology for liver cancer mortality and morbidity.
Section 0.1.5 presents the results of the analysis.
Final PFAS Rule Economic Analysis
0-1
April 2024
-------
FINAL RULE
APRIL 2024
Legend:
Result of
upstream analysis
Data/Inputs
Model
Serum PFOS
concentration difference
between baseline and
regulatory alternative
# Analysis step
Change in cumulative liver
cancer risk
Change in the
number of liver
cancer cases
Location-specific
population size
National Cancer
Institute SEER
program data
Annual cause-specific
mortality rates and life
table information3
Change in excess
liver cancer
population
mortality
Willingness to pay to
avoid liver cancer
Value of reduced
liver cancer cases
Value of avoided
liver cancer
mortality
Total value of reduced liver cancer
Abbreviations: PFOS - perfluorooctane sulfonic acid, SEER - Surveillance, Epidemiology, and End Results program
Notes:
''Data from the Centers for Disease Control (CDC) and Prevention.
Figure O-l. Overview of Analysis of Reduced Liver Cancer Risk
0.1.2 Liver Cancer Exposure-Response Modeling
Evidence of the association between PFOA and PFOS exposure and liver cancer in humans was
considered inconclusive based on occupational and general population epidemiology studies
(U.S. EPA, 2024b; U.S. EPA, 2024c). However, the EPA found evidence of a positive
association between PFOS exposure and hepatocellular tumors in animal studies. Butenhoff et al.
(2012)/Thomford (2002) reported a statistically significant increase in combined hepatocellular
adenomas and carcinomas tumor incidence in female Sprague-Dawley rats exposed to high doses
of PFOS. The study reported a statisti cally significant trend of increased incidence with
Final PFAS Rule Economic Analysis
0-2
April 2024
-------
FINAL RULE
APRIL 2024
increasing PFOS concentrations across dose groups. The EPA reviewed the weight of the
evidence and determined that PFOS is Likely to Be Carcinogenic to Humans, as "the evidence is
adequate to demonstrate carcinogenic potential to humans but does not reach the weight of
evidence for the descriptor Carcinogenic to Humans." The EPA evaluated the effects of the final
rule on liver cancer using the relationships between PFOS exposure and hepatocellular adenomas
and carcinomas in female rats.
To evaluate changes between baseline and regulatory alternative liver cancer risk resulting from
reduced exposure to PFOS, the EPA relied on the estimated time series of changes in serum
PFOS concentrations (Section 6.3) and the cancer slope factor calculated based on the EPA's
benchmark dose (BMD) modeling results for hepatocellular adenomas and carcinomas in female
rats following exposure to PFOS. The EPA carried forward the animal BMD of 37.2 mg/L
(Table E-46 of the PFOS MCLG Appendix; U.S. EPA, 2024a), which corresponds to the internal
human BMD. This value represents the internal human BMD because the animal BMD was
based on area under the curve (AUC) normalized per day (AUCavg), equivalent to mean serum
concentration during the duration of the study, which as selected for this model; the AUC
accounts for the accumulation of effects expected to precede the increased incidence of
adenomas and/or carcinomas (U.S. EPA, 2005a). The EPA then applied the linear extrapolation
approach to calculate the human cancer risk factor by dividing the benchmark response (BMR)
of 10% by the human BMD, which resulted in 2.69* 10"6 per ng/mL. This linear slope factor
enables estimation of the changes in lifetime and relative liver cancer risk associated with
reduced lifetime serum PFOS levels, as described in Equations 15, 16, and 27 of Section 6.6.2.
0.1.3 Estimation of Liver Cancer Risk Reductions
The EPA relies on the life table approach to estimate liver cancer risk reductions because:
• Changes in serum PFOS in response to changes in drinking water PFOS occur over multiple
years;
• Annual risk of new liver cancer should be quantified only among those not already
experiencing this chronic condition; and
• Liver cancer has elevated mortality implications.
The EPA used recurrent life table calculations to estimate PWS EP-specific time series of liver
cancer incidence for a population cohort characterized by sex, race/ethnicity, birth year, and age
at the beginning of the evaluation period (i.e., 2024) under the baseline scenario and the
regulatory alternatives. The life table analysis accounts for the gradual changes in lifetime
exposures to PFOS following implementation of treatment under the regulatory alternatives
compared to the baseline. Details of the life table calculations are provided in Appendix H. The
outputs of the life table calculations are the PWS EP-specific estimates of the annual change in
the number of liver cancer cases and the annual change in liver cancer population mortality.
Although the change in PFOS exposure likely affects the risk of developing liver cancer beyond
the end of the analysis period (the majority of liver cancer cases manifest during the latter half of
the average individual lifespan; see Appendix H), the EPA does not capture effects after the end
of the period of analysis, 2105. Individuals alive after the end of the period of analysis likely
benefit from lower lifetime exposure to PFOS. Lifetime health risk model data sources include
SDWIS/Fed; age-, sex-, and race/ethnicity-specific population estimates from the U.S. Census
Final PFAS Rule Economic Analysis
0-3
April 2024
-------
FINAL RULE
APRIL 2024
Bureau (U.S. Census Bureau, 2020); the SEER program database (National Cancer Institute),
and the CDC National Center for Health Statistics. Appendix H provides additional detail on the
data sources and information used in this analysis as well as baseline liver cancer statistics.
Appendix B describes estimation of the affected population.
0.1.4 Valuation of Liver Cancer Risk Reductions
The EPA uses the Value of Statistical Life to estimate the benefits of reducing mortality
associated with liver cancer in the population exposed to PFOS in drinking water. Appendix J
provides information on updating Value of Statistical Life for inflation and income growth. The
EPA uses the willingness to pay estimates per statistical non-site specific nonfatal cancer
avoided from Bosworth et al. (2009), which was identified in a literature review from Abt
Associates (2022), Estimated Values of Avoiding Cancer Risks by Cancer Site and Population,
to value liver cancer morbidity. Bosworth et al. (2009) elicited willingness to pay to avoid
illnesses and premature death using a national survey in a choice experiment format. The
valuation scenarios presented to survey respondents described a proposed public policy that will
reduce community-level risk of both illness and death for these diseases by improving air
pollution, drinking water contamination, and the levels of pesticides in foods. Survey participants
were asked to choose between the two offered policies based on the private cost of the policy and
the number of avoided illnesses and deaths.
To obtain a willingness to pay value suitable for valuation of liver cancer morbidity risk
reductions during 2024-2105, the EPA relies on the base value estimate of $245,000 ($2009,
2009 income year) from Bosworth et al. (2009). The EPA followed the methodology used to
adjust the base Value of Statistical Life for inflation and income growth for adjusting willingness
to pay estimates (see Appendix J for details). Unlike the Value of Statistical Life, which is
adjusted for income growth based on an assumed elasticity of 0.4, willingness to pay values are
adjusted based on an assumed elasticity of 0.45, which represents the central elasticity estimate
for severe and chronic health effects (U.S. EPA, 2023b). Like Value of Statistical Life,
willingness to pay estimates are approximated using the CAGR from 2024 to 2050 (the final year
that income growth projections are available) to estimate willingness to pay values for the entire
period of analysis, 2024 to 2105. The estimates of willingness to pay per statistical non-site
specific cancer morbidity avoided range from $364,060 ($2022) in 2024 to $643,142 ($2022) in
2105. Table O-l summarizes the projected willingness to pay estimates through 2050 and the
approximated willingness to pay estimates through 2105.
Table O-l. Estimated Liver Cancer Willingness to Pay Series
Year
Historical
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Projected
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Income Growth
Factor (Ratio of
Projected
PDYPP to
Historical 2009
PDYPP to the
Power of 0.45)
Projected
Willingness to Pay
($2022)
Approximated
Willingness to Pay
($2022)
2009
30,327
-
1
328,021
-
2024
-
47,987
1.109868225
364,060
364,060
2025
-
48,917
1.119502162
367,220
366,627
2026
-
49,760
1.128145153
370,055
369,212
Final PFAS Rule Economic Analysis
0-4
April 2024
-------
FINAL RULE
APRIL 2024
Table O-l. Estimated Liver Cancer Willingness to Pay Series
Year
Historical
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Projected
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Income Growth
Factor (Ratio of
Projected
PDYPP to
Historical 2009
PDYPP to the
Power of 0.45)
Projected
Willingness to Pay
($2022)
Approximated
Willingness to Pay
($2022)
2027
-
50,616
1.136833419
372,905
371,815
2028
-
51,496
1.145687172
375,810
374,436
2029
-
52,407
1.154763474
378,787
377,076
2030
-
53,393
1.164489967
381,977
379,734
2031
-
54,326
1.17359746
384,965
382,411
2032
-
55,258
1.182614732
387,923
385,107
2033
-
56,207
1.191716687
390,908
387,822
2034
-
57,145
1.200625062
393,830
390,556
2035
-
58,072
1.209352808
396,693
393,310
2036
-
58,985
1.217865528
399,486
396,082
2037
-
59,874
1.226099557
402,187
398,875
2038
-
60,753
1.234164989
404,832
401,687
2039
-
61,643
1.242269802
407,491
404,519
2040
-
62,513
1.250121412
410,066
407,371
2041
-
63,408
1.25815293
412,701
410,243
2042
-
64,346
1.266493129
415,437
413,135
2043
-
65,282
1.274745794
418,144
416,047
2044
-
66,210
1.282866035
420,807
418,981
2045
-
67,148
1.291015411
423,480
421,934
2046
-
68,095
1.299174972
426,157
424,909
2047
-
69,069
1.307509356
428,891
427,905
2048
-
70,076
1.316055374
431,694
430,921
2049
-
71,066
1.324386799
434,427
433,959
2050
-
72,024
1.332395223
437,054
437,019
2051
-
-
-
-
440,100
2052
-
-
-
-
443,202
2053
-
-
-
-
446,327
2054
-
-
-
-
449,474
2055
-
-
-
-
452,642
2056
-
-
-
-
455,834
2057
-
-
-
-
459,047
2058
-
-
-
-
462,283
2059
-
-
-
-
465,543
2060
-
-
-
-
468,825
2061
-
-
-
-
472,130
Final PFAS Rule Economic Analysis
0-5
April 2024
-------
FINAL RULE
APRIL 2024
Table O-l. Estimated Liver Cancer Willingness to Pay Series
Historical Projected Income Growth
Personal Personal Factor (Ratio of
Disposable Disposable Projected Projected
Approximated
Year
Income Per Income Per PDYPP to Willingness to Pay
Willingness to Pay
Capita Capita Historical 2009 ($2022)
($2022)
(PDYPP, (PDYPP, PDYPP to the
$2012) $2012) Power of 0.45)
2062
.
475,458
2063
-
478,810
2064
-
482,186
2065
-
485,585
2066
-
489,009
2067
-
492,456
2068
-
495,928
2069
-
499,424
2070
-
502,945
2071
-
506,491
2072
-
510,062
2073
-
513,658
2074
-
517,279
2075
-
520,926
2076
-
524,598
2077
-
528,297
2078
-
532,021
2079
-
535,772
2080
-
539,549
2081
-
543,353
2082
-
547,184
2083
-
551,041
2084
-
554,926
2085
-
558,838
2086
-
562,778
2087
-
566,746
2088
-
570,741
2089
-
574,765
2090
-
578,817
2091
-
582,898
2092
-
587,007
2093
-
591,146
2094
-
595,313
2095
-
599,510
2096
-
603,737
Final PFAS Rule Economic Analysis
0-6
April 2024
-------
FINAL RULE APRIL 2024
Table O-l. Estimated Liver Cancer Willingness to Pay Series
Historical Projected Income Growth
Personal Personal Factor (Ratio of
Disposable Disposable Projected Projected Approximated
Year Income Per Income Per PDYPP to Willingness to Pay Willingness to Pay
Capita Capita Historical 2009 ($2022) ($2022)
(PDYPP, (PDYPP, PDYPP to the
$2012) $2012) Power of 0.45)
2097 ... . 607,993
2098 ... . 612,280
2099 ... . 616,596
2100 ... . 620,943
2101 ... . 625,321
2102 ... . 629,729
2103 ... . 634,169
2104 ... . 638,640
210 5 - - - - 643,142
Acronym: PDYPP- personal disposable income per capita.
0.1.5 Results
Table 0-2 provides the health effects avoided and valuation associated with liver cancer under
the final rule MCL and HI assumptions. Modeled uncertainty includes uncertainty regarding the
PAF estimation and occurrence estimates. Annualized liver cancer benefits are $4.79 million.
Table 0-2. National Liver Cancer Benefits, Final Rule (PFOA and PFOS MCLs of 4.0
ppt each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1)
2% Discount Rate
Benefits Category
5th Percentile3
Expected Value
95th Percentile3
Number of Non-Fatal Liver
Cancer Cases Avoided
13.30
14.17
15.08
Number of Liver Cancer-
Related Deaths Avoided
29.36
31.25
33.29
Total Annualized Liver
Cancer Benefits (Million
$2022)b
$4.50
$4.79
$5.10
Notes: Detail may not add exactly to total due to independent rounding. See Appendix P for results presented at 3 and 7
percent discount rates.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty related to PAF and occurrence. This range
does not include the uncertainty described in Table 0-3.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
0-7
April 2024
-------
FINAL RULE
APRIL 2024
0.1.6Limitations and Uncertainties
Table 0-3 describes limitations and uncertainties of the supplemental liver cancer benefits
analysis. Limitations and uncertainties that apply to all health benefits analyses are summarized
in Table 6-48.
Table 0-3. Limitations and Uncertainties in the Analysis of Liver Cancer Benefits
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
Characterizing the Exposed Population
The analysis uses national-level
estimates of liver cancer incidence,
prevalence, stage distribution, and
relative survival data, as well as
national-level life tables.
Uncertain
Using national-level baseline health data may
over- or underestimate the effects of regulatory
alternatives on liver cancer morbidity and
mortality in specific PWSs and well as overall.
Liver cancer risks are estimated for
populations for which reductions in
PFOS exposures relative to
baseline exposures start at different
ages, including children.
Uncertain
The relative cancer potency of PFOS in children
is unknown which may bias benefits estimates
either upward or downward. Because liver
cancer incidence in children is very small, we
assess any bias to be negligible.
Modeling Changes in Health Risks
The analysis relies on associations
between PFOS exposure and
hepatocellular adenomas and
carcinomas in animals.
Uncertain
The cancer slope factor is based on associations
between PFOS exposure and hepatocellular
adenomas and carcinomas observed in female
rats. This relationship may not accurately reflect
association be PFOS exposure and risk of liver
cancer in humans. The effect of using a cancer
slope factor specific to animals to evaluate
changes in the incidence of liver cancer in
humans is uncertain.
Final PFAS Rule Economic Analysis
0-8
April 2024
-------
FINAL RULE APRIL 2024
Table 0-3. Limitations and Uncertainties in the Analysis of Liver Cancer Benefits
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
The analysis does not explicitly
model variability of baseline liver
cancer risk by cirrhosis and
hepatitis B infection status.
Uncertain
In humans, 95 percent of primary liver tumors
are malignant, with hepatocellular carcinoma
comprising 90 percent of malignancies in adults
(B. B. Anderson et al., 1992). The risk of
hepatocellular carcinoma is 33 to 200 times
higher in populations with cirrhosis of the liver
and populations with hepatitis B infection.
While each population represents approximately
1 percent of the U.S. population overall, 75
percent of hepatocellular carcinoma incidence
occurs in those affected by cirrhosis/hepatitis B
(B. B. Anderson et al., 1992). The cancer slope
factor used in the analysis represents an additive
change in liver cancer risk and the extent to
which it may be modified in the
cirrhosis/hepatitis B populations is uncertain.
The available association between PFOS
exposure and liver cancer risk is linear, implying
that the estimated lifetime risk reductions do not
depend on the baseline liver cancer risk level.
Therefore, modeling cirrhosis/hepatitis B
population in this analysis will not generate
additional insights.
The analysis assumes that the
magnitude of liver cancer risk
reductions resulting from
reductions in serum PFOA levels
will not exceed a PAF of 3.94
percent.
Uncertain
The EPA placed a cap of 3.94 percent on the
magnitude of the estimated cumulative liver
cancer risk reduction resulting from reductions
in serum PFOS levels, based on its analysis of
PAF values found in the literature on
environmental contaminants and cancers (ICF,
2022). This review found that changes in
environmental exposures result in relatively
modest PAFs (between 0.2 percent and 17.9%);
however, few of the studies provided PAFs
related specifically to liver cancer. The EPA
characterized the uncertainty surrounding this
parameter using a log-uniform distribution with
a minimum of 0.2 percent and a maximum of
17.9 percent. For the central estimate of liver
cancer benefits, the EPA used a PAF of 3.94
percent, which is the mean of the PAF
uncertainty distribution. As such, the EPA
assumed that liver cancer risk reduction
estimates in excess of the PAF are unreasonable
even as a result of large changes in serum PFOS
concentrations. Because this PAF cap is not
based on liver cancer studies specifically, it is
uncertain whether the liver cancer impacts are
under- or overestimated.
Final PFAS Rule Economic Analysis
0-9
April 2024
-------
FINAL RULE APRIL 2024
Table 0-3. Limitations and Uncertainties in the Analysis of Liver Cancer Benefits
Uncertainty/Assumption
Effect on Benefits
Estimate
Notes
The analysis assumes that there is
no lag between changes in serum
PFOS concentrations and changes
in liver cancer incidence.
Overestimate
The studies estimating the association between
serum PFOS and liver cancer are not dynamic,
and hence do not provide insights into whether
liver cancer incidence may respond gradually to
changes in serum PFOS. The PK model
estimates daily serum levels, which are averaged
annually for the purposes of modeling gradual
serum changes for the liver cancer risk reduction
analysis. The liver cancer risk reduction analysis
assumes immediate liver cancer incidence
adjustment within each year, which may
overestimate impacts to the exposed population.
The analysis relies on public-access
SEER 20 10-year relative liver
cancer survival data to model
mortality patterns in the liver
cancer population.
Uncertain
Reliance on these data generates both a
downward and an upward bias. The downward
bias is due to the short, 10-year excess mortality
follow-up window. Survival rates beyond 10
years following the initial diagnosis are likely to
be lower. The upward bias comes from the
inability to determine how many of the excess
deaths were deaths from liver cancer.
The analysis models the 85+ year
old group jointly and applies the
average mortality rate for those
aged 85+ in this age group.
Uncertain
The effect of this modeling approximation on
the liver cancer benefits is not certain because
integer age-specific mortality rates may be
above or below the average mortality rate.
The analysis models the 85+ year
old group jointly and uses serum
PFOS estimates for those aged 85
to initiate calculations in this age
group.
Underestimate
Because the impacts of changes in PFOS
drinking water concentrations on serum PFOS
levels increase over time, the use of serum
PFOS concentrations at 85 years to model the
85+ age group will underestimate the liver
cancer risk impacts in this group.
Economic Valuation of Changes in Health Risk
The analysis relies on willingness
to pay estimates per statistical non-
site specific nonfatal cancer
avoided to estimate benefits from
avoided liver cancer cases.
Uncertain
Primary liver cancer is most treatable if detected
early which is not a common situation.
Moreover, people who develop liver cancer
usually already have an unhealthy liver and
would require liver transplant rather than partial
hepatectomy. Given the complexity of this organ
and a low rate of cure for this type of cancer, the
use of willingness to pay for avoiding non-site
specific cancer may underestimate the value of
avoiding non-fatal liver cancer. On the other
hand, Bosworth et al. (2009) found "little
heterogeneity of preferences according to the
type of illness".
Abbreviations: PFOS - perfluorooctane sulfonic acid; PK - pharmacokinetic.
Final PFAS Rule Economic Analysis
O-IO
April 2024
-------
FINAL RULE
APRIL 2024
0.25upplemental Analysis Using Willingness to Pay for Cancer
Morbidity Risk Reductions
Table 0-5 and Table 0-6 present results for supplemental national-lev el estimates of RCC
benefits and bladder cancer co-benefits, respectfully, considering willingness to pay metrics for
monetization of non-fatal cancer cases. The EPA relies on base willingness to pay estimates from
Bosworth et al. (2009) for unspecified cancer in monetizing RCC benefits and for colon/bladder
cancer in monetizing bladder cancer co-benefits.61 The base estimates of willingness to pay per
illness avoided based on an affected population of 50,000 for a duration of ten years are
$245,000 for unspecified cancer and $400,000 for colon/bladder cancer (reported in $2009). The
EPA relied on the approach described in Appendix J to adjust these estimates for inflation and
income growth from 2009 to 2024-2050 and calculated the compound annual growth to
approximate willingness to pay values during the analysis period (2024 to 2105; see Equations J-
1, J-2, and J-3). As described in Section 0.1.4, willingness to pay estimates were adjusted for
income growth using an assumed elasticity of 0.45, the central elasticity estimate for severe and
chronic health effects (U.S. EPA, 2023b). Unspecified cancer willingness to pay estimates range
from $364,060 ($2022) in 2024 to $643,142 ($2022), as reported in Table O-l. Colon/bladder
cancer willingness to pay estimates range from $594,384 in 2024 to $1,050,028 in 2105 ($2022),
as reported below in Table 0-4. When using willingness to pay instead of cost of illness values
to monetize cancer morbidity impacts, annualized RCC benefits are $360.97 million, whereas
annualized bladder cancer benefits are $456.28 million.
Table 0-4. Estimated Bladder Cancer Willingness to Pay Series
Year
Historical
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Projected
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Income Growth
Factor (Ratio of
Projected
PDYPP to
Historical 2009
PDYPP to the
Power of 0.45)
Projected
Willingness to Pay
($2022)
Approximated
Willingness to Pay
($2022)
2009
38,064
-
1
535,545
2024
-
47,987
1.109868225
594,384
594,384
2025
-
48,917
1.119502162
599,543
598,574
2026
-
49,760
1.128145153
604,172
602,794
2027
-
50,616
1.136833419
608,825
607,044
2028
-
51,496
1.145687172
613,567
611,324
2029
-
52,407
1.154763474
618,427
615,634
2030
-
53,393
1.164489967
623,636
619,974
2031
-
54,326
1.17359746
628,514
624,345
2032
-
55,258
1.182614732
633,343
628,746
01 The EPA did not identify a willingness to pay estimate specific to kidney cancer in the available literature. Estimates from
Bosworth et al. (2009) were implemented in the EPA's Economic Analysis of the Proposed Regulation of Methylene Chloride
Under TSCA Section 6(a) (U.S. EPA, 2023a).
Final PFAS Rule Economic Analysis
O-ll
April 2024
-------
FINAL RULE
APRIL 2024
Table 0-4. Estimated Bladder Cancer Willingness to Pay Series
Year
Historical
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Projected
Personal
Disposable
Income Per
Capita
(PDYPP,
$2012)
Income Growth
Factor (Ratio of
Projected
PDYPP to
Historical 2009
PDYPP to the
Power of 0.45)
Projected
Willingness to Pay
($2022)
Approximated
Willingness to Pay
($2022)
2033
-
56,207
1.191716687
638,218
633,179
2034
-
57,145
1.200625062
642,988
637,643
2035
-
58,072
1.209352808
647,662
642,138
2036
-
58,985
1.217865528
652,221
646,665
2037
-
59,874
1.226099557
656,631
651,224
2038
-
60,753
1.234164989
660,951
655,815
2039
-
61,643
1.242269802
665,291
660,439
2040
-
62,513
1.250121412
669,496
665,095
2041
-
63,408
1.25815293
673,797
669,784
2042
-
64,346
1.266493129
678,264
674,506
2043
-
65,282
1.274745794
682,683
679,261
2044
-
66,210
1.282866035
687,032
684,050
2045
-
67,148
1.291015411
691,396
688,873
2046
-
68,095
1.299174972
695,766
693,729
2047
-
69,069
1.307509356
700,230
698,620
2048
-
70,076
1.316055374
704,806
703,545
2049
-
71,066
1.324386799
709,268
708,505
2050
-
72,024
1.332395223
713,557
713,500
2051
-
-
-
-
718,530
2052
-
-
-
-
723,596
2053
-
-
-
-
728,697
2054
-
-
-
-
733,835
2055
-
-
-
-
739,008
2056
-
-
-
-
744,218
2057
-
-
-
-
749,465
2058
-
-
-
-
754,749
2059
-
-
-
-
760,070
2060
-
-
-
-
765,428
2061
-
-
-
-
770,824
2062
-
-
-
-
776,259
2063
-
-
-
-
781,731
2064
-
-
-
-
787,242
2065
-
-
-
-
792,792
2066
-
-
-
-
798,382
2067
-
-
-
-
804,010
Final PFAS Rule Economic Analysis
0-12
April 2024
-------
FINAL RULE
APRIL 2024
Table 0-4. Estimated Bladder Cancer Willingness to Pay Series
Historical Projected Income Growth
Personal Personal Factor (Ratio of
Disposable Disposable Projected Projected Approximated
Year Income Per Income Per PDYPP to Willingness to Pay Willingness to Pay
Capita Capita Historical 2009 ($2022) ($2022)
(PDYPP, (PDYPP, PDYPP to the
$2012) $2012) Power of 0.45)
2068 ... . 809,679
2069 ... . 815,387
2070 ... . 821,135
2071 - - - - 826,924
2072 ... . 832,754
2073 ... . 838,625
2074 ... . 844,537
2075 ... . 850,491
2076 ... . 856,487
2077 ... . 862,525
2078 ... . 868,606
2079 ... . 874,730
2080 ... . 880,897
2081 ... . 887,107
2082 ... . 893,361
2083 ... . 899,659
2084 ... . 906,002
2085 ... . 912,389
2086 ... . 918,822
2087 ... . 925,299
2088 ... . 931,823
2089 ... . 938,392
2090 ... . 945,008
2091 ... . 951,670
2092 ... . 958,379
2093 ... . 965,136
2094 ... . 971,940
2095 ... . 978,792
2096 ... . 985,693
2097 ... . 992,642
2098 ... . 999,640
2099 ... . 1,006,688
2100 ... . 1,013,785
2101 ... . 1,020,932
2102 ... . 1,028,129
Final PFAS Rule Economic Analysis
0-13
April 2024
-------
FINAL RULE
APRIL 2024
Table 0-4. Estimated Bladder Cancer Willingness to Pay Series
Historical
Projected
Income Growth
Personal
Personal
Factor (Ratio of
Disposable
Disposable
Projected
Projected
Approximated
Year
Income Per
Income Per
PDYPP to
Willingness to Pay
Willingness to Pay
Capita
Capita
Historical 2009
($2022)
($2022)
(PDYPP,
(PDYPP,
PDYPP to the
$2012)
$2012)
Power of 0.45)
2103 ... . 1,035,378
2104
1,042,677
2105
1,050,028
Acronym: PDYPP- personal disposable income per capita.
Table 0-5. National Willingness to Pay-Based RCC Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1)
2% Discount Rate
Benefits Category 5th Percentile3 Expected Value
95th Percentile3
Number of Non-Fatal RCC 1,091.50 6,964.20
Cases Avoided
17,937.00
Number of RCC-Related 320.36 2,028.80
Deaths Avoided
5,206.50
Total Annualized RCC $62.07 $360.97
Benefits (Million $2022)b
$901.91
Notes: Detail may not add exactly to total due to independent rounding. See Appendix P for results presented at 3 and 7
percent discount rates.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
0-14
April 2024
-------
FINAL RULE
APRIL 2024
Table 0-6. National Willingness to Pay-Based Bladder Cancer Benefits, Final Rule
(PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA of 10 ppt each and
HI of 1)
2% Discount Rate
Benefits Category
5th Percentile3
Expected Value
95th Percentile3
Number of Non-Fatal
5,781.00
7,313.00
8,912.70
Bladder Cancer Cases
Avoided
Number of Bladder Cancer-
2,029.60
2,567.80
3,129.90
Related Deaths Avoided
Total Annualized Bladder
$360.61
$456.28
$556.21
Cancer Benefits (Million
$2022)b
Notes: Detail may not add exactly to total due to independent rounding. See Appendix P for results presented at 3 and 7
percent discount rates.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
0-15
April 2024
-------
FINAL RULE
APRIL 2024
Appendix P. Additional Model Outputs
In the tables below, the EPA reports additional costs and benefits model outputs. Sections P. 1
through P.7 report results at 3 percent and 7 percent discount rates. The EA for the proposed
PFAS NPDWR presented costs and benefits consistent with the OMB Circular A-4 guidance at
the time of proposal. OMB guidance at the time of proposal indicated that the 3 percent discount
rate reflects society's valuation of differences in the timing of consumption; the 7 percent
discount rate reflects the opportunity cost of capital to society. In the 2003 Circular A-4, the
OMB recommended that 3 percent be used when a regulation affects private consumption, and 7
percent be used when evaluating a regulation that would mainly displace or alter the use of
capital in the private sector (OMB, 2003; updated 2009). In this appendix, the EPA presents
costs and benefits at both 3 and 7 percent discount rates to allow for a direct comparison for the
final quantified cost and benefits to the quantified costs and benefits presented for the proposed
rule. The EPA notes that given the updated default social discount rate of 2 percent prescribed in
the finalized OMB Circular A-4 (OMB, 2023) and also public input received on the discount
rates considered by the EPA in the proposed NPDWR, for this final rule, the EPA estimated
national benefits and costs at the 2 percent discount rate for the final rule and incorporated those
results into the final economic analysis. The Administrator reaffirms his determination that the
benefits of the rule justify the costs. The EPA's determination is based on its analysis under in
SDWA Section 1412(b)(3)(C) of the quantifiable benefits and costs at the 2 percent discount
rate, in addition to at the 3 and 7 percent discount rate, as well as the nonquantifiable benefits
and costs. The EPA found that significant nonquantifiable benefits are likely to occur from the
final PFAS NPDWR.
Section P.8 presents undiscounted benefits and costs.
Final PFAS Rule Economic Analysis
P-l
April 2024
-------
FINAL RULE
APRIL 2024
P.l Total Estimated Benefits and Costs
Table P-l: Quantified Total National Annualized Benefits, All Options (Million $2022)
Option
3% Discount Rate
,a
7% Discount Rate
.a
5th
Percentile13
Expected
Value
95th
Percentile13
5th
Percentile13
Expected
Value
95th
Percentile13
Final rule0
$821.07
$1,393.56
$2,053.30
$536.67
$916.49
$1,328.90
Option lad
$815.03
$1,387.48
$2,043.00
$534.22
$912.35
$1,321.70
Option lbe
$688.91
$1,167.15
$1,722.70
$450.77
$769.28
$1,117.10
Option lcf
$356.37
$598.63
$872.69
$233.73
$396.05
$572.67
Notes: Detail may not add exactly to total due to independent rounding. Quantified total national annualized benefits do not
include quantified sensitivity analysis results for PFNA effects on birth weight and PFOS effects on liver cancer, and as such,
the quantified total national annualized benefits may be underestimated. See appendices K and O for PFNA birth weight and
PFOS liver cancer sensitivity analysis results, respectively.
aSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
bThe 5tli and 95tli percentile range is based on modeled variability and uncertainty described in Section 6.1.2 and Table 6-1
for benefits. This range does not include the uncertainty described in 6-48 for benefits.
The final rule sets PFOA and PFOS MCLs of 4.0 ppt each, an HI of 1, and MCLs for HFPO-DA, PFNA, and PFHxS of 10
ppt each.
Option la sets PFOA and PFOS MCLs only, at 4.0 ppt each.
eOption lb sets PFOA and PFOS MCLs only, at 5.0 ppt each.
fOption lc sets PFOA and PFOS MCLs only, at 10.0 ppt each.
Final PFAS Rule Economic Analysis
P-2
April 2024
-------
FINAL RULE APRIL 2024
Table P-2: Quantified Total National Annualized Costs, All Options (Million $2022)
3% Discount Rate'
i,b
7% Discount Rate'
i,b
Option
5th
Percentile0
Mean
95th
Percentile0
5th
Percentile0
Mean
95th
Percentile0
Final ruled'e
$1,431.50
$1,545.61
$1,670.10
$1,437.00
$1,553.98
$1,688.00
Option laf
$1,420.30
$1,534.03
$1,658.20
$1,425.50
$1,542.57
$1,676.70
Option lbg
$1,100.10
$1,189.99
$1,290.30
$1,103.90
$1,197.32
$1,304.10
Option lch
$461.72
$498.64
$540.36
$464.77
$503.02
$547.76
Notes: Detail may not add exactly to total due to independent rounding.
aSee Table 7-6 for a list of the nonquantifiable costs, and the potential direction of impact these costs would have on the
estimated monetized total annualized costs in this table.
bPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore
total costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To
address stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be
regulated as such in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for
illustrative purposes only. See Appendix N and Section N.2 for additional detail.
The 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1 for
costs. This range does not include the uncertainty described in Table 5-22 for costs.
Quantified national costs do not include quantified sensitivity analysis results for PFNA, PFBS, and HFPO-DA. Including the
costs of treating for these compounds increases total annualized cost of the final rule to $1,630.46 million at a 3 percent
discount rate and $1,634.56 million at a 7 percent discount rate. These benefits and costs are considered quantitatively in the
sensitivity analysis. See Section N.3for more information.
eThe final rule sets PFOA and PFOS MCLs of 4.0 ppt each, an HI of 1 and MCLs for HFPO-DA, PFNA, and PFHxS of 10 ppt
each.
fOption la sets PFOA and PFOS MCLs of 4.0 ppt each.
gOption lb sets PFOA and PFOS MCLs of 5.0 ppt each.
hOption lc sets PFOA and PFOS MCLs of 10.0 ppt each.
Final PFAS Rule Economic Analysis
P-3
April 2024
-------
FINAL RULE
APRIL 2024
P.2 National Annualized Costs
Table P-3: National Annualized Costs, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each,
PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) (Million $2022)
3% Discount Rate 7% Discount Rate
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Annualized PWS Sampling
$34.45
$37.14
$40.06
$37.71
$40.80
$44.13
Costs
Annualized PWS
$1.72
$1.72
$1.72
$3.41
$3.41
$3.41
Implementation and
Administration Costs
Annualized PWS Treatment
$1,391.16
$1,501.68
$1,624.89
$1,388.69
$1,503.01
$1,634.84
Costs
Total Annualized PWS
$1,426.60
$1,540.54
$1,665.10
$1,431.30
$1,547.22
$1,680.60
Costs
Primacy Agency Rule
$4.73
$5.07
$5.45
$6.26
$6.76
$7.32
Implementation and
Administration Cost
Total Annualized Rule
$1,431.50
$1,545.61
$1,670.10
$1,437.00
$1,553.98
$1,688.00
Costsb'c'd
Abbreviations: PWS - public water system.
Notes: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule costs are not
additive across cost category as the categories are not completely correlated.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1. This
range does not include the uncertainty described in 5-22.
bSee Table 7-6 for a list of the nonquantifiable costs, and the potential direction of impact these costs would have on the estimated
monetized total annualized costs in this table.
The national level cost estimates for PFHxS are reflective of both the total national cost for PFHxS individual MCL exceedances,
and HI MCL exceedances where PFHxS is present above its HBWC while one or more other HI PFAS is also present in that same
mixture. Total quantified national cost values do not include the incremental treatment costs associated with the co-occurrence of
HFPO-DA, PFBS, and PFNA. EPA has considered the additional national costs of the HI and individual MCLs associated with
HFPO-DA, PFNA, and PFBS occurrence in a quantified sensitivity analysis; See Appendix N and Section N.3 for the analysis and
more information..
dPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore
total costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To
address stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be
regulated as such in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for
illustrative purposes only. See Appendix N and Section N.2 for additional detail.
Final PFAS Rule Economic Analysis
P-4
April 2024
-------
FINAL RULE APRIL 2024
Table P-4: National Annualized Costs, Option la (PFOA and PFOS MCLs of 4.0 ppt) (Million
$2022)
3% Discount Rate 7% Discount Rate
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Annualized PWS Sampling
$34.17
$36.88
$39.80
$37.42
$40.51
$43.84
Costs
Annualized PWS
$1.72
$1.72
$1.72
$3.41
$3.41
$3.41
Implementation and
Administration Costs
Annualized PWS Treatment
$1,379.26
$1,490.37
$1,612.89
$1,377.38
$1,491.91
$1,623.54
Costs
Total Annualized PWS
$1,415.40
$1,528.98
$1,653.10
$1,419.30
$1,535.83
$1,669.60
Costs
Primacy Agency Rule
$4.71
$5.05
$5.42
$6.24
$6.73
$7.29
Implementation and
Administration Cost
Total Annualized Rule
$1,420.30
$1,534.03
$1,658.20
$1,425.50
$1,542.57
$1,676.70
Costsb'c
Abbreviations: PWS - public water system.
Notes: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule costs are not
additive across cost category as the categories are not completely correlated.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1. This
range does not include the uncertainty described in Table 5-22.
bSee Table 7-6 for a list of the nonquantifiable costs, and the potential direction of impact these costs would have on the estimated
monetized total annualized costs in this table.
cPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore total
costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To address
stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be regulated as such in
the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for illustrative purposes only. See
Appendix N and Section N.2 for additional detail.
Final PFAS Rule Economic Analysis
P-5
April 2024
-------
FINAL RULE
APRIL 2024
Table P-5: National Annualized Costs, Option lb (PFOA and PFOS MCLs of 5.0 ppt)
(Million $2022)
3% Discount Rate 7% Discount Rate
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Annualized PWS Sampling
$31.75
$34.07
$36.60
$34.62
$37.25
$40.15
Costs
Annualized PWS
$1.72
$1.72
$1.72
$3.41
$3.41
$3.41
Implementation and
Administration Costs
Annualized PWS Treatment
$1,061.02
$1,149.63
$1,248.65
$1,059.22
$1,150.64
$1,254.96
Costs
Total Annualized PWS
$1,095.90
$1,185.42
$1,285.60
$1,098.40
$1,191.30
$1,298.00
Costs
Primacy Agency Rule
$4.31
$4.57
$4.87
$5.63
$6.02
$6.46
Implementation and
Administration Cost
Total Annualized Rule
$1,100.10
$1,189.99
$1,290.30
$1,103.90
$1,197.32
$1,304.10
Costsb'c
Abbreviations: PWS - public water system.
Notes: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule costs are not
additive across cost category as the categories are not completely correlated.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1. This
range does not include the uncertainty described in Table 5-22.
bSee Table 7-6 for a list of the nonquantifiable costs, and the potential direction of impact these costs would have on the estimated
monetized total annualized costs in this table.
cPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore
total costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To
address stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be
regulated as such in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for
illustrative purposes only. See Appendix N and Section N.2 for additional detail.
Final PFAS Rule Economic Analysis
P-6
April 2024
-------
FINAL RULE
APRIL 2024
Table P-6: National Annualized Costs, Option lc (PFOA and PFOS MCLs of 10.0 ppt)
(Million $2022)
3% Discount Rate 7% Discount Rate
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Annualized PWS Sampling
$26.57
$27.99
$29.53
$28.62
$30.22
$31.96
Costs
Annualized PWS
$1.72
$1.72
$1.72
$3.41
$3.41
$3.41
Implementation and
Administration Costs
Annualized PWS Treatment
$429.35
$465.33
$506.21
$427.86
$464.79
$508.64
Costs
Total Annualized PWS
$458.15
$495.04
$536.59
$460.46
$498.42
$543.00
Costs
Primacy Agency Rule
$3.50
$3.60
$3.73
$4.46
$4.61
$4.79
Implementation and
Administration Cost
Total Annualized Rule
$461.72
$498.64
$540.36
$464.77
$503.02
$547.76
Costsb'c
Abbreviations: PWS - public water system.
Notes: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule costs are not
additive across cost category as the categories are not completely correlated.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1. This
range does not include the uncertainty described in Table 5-22.
bSee Table 7-6 for a list of the nonquantifiable costs, and the potential direction of impact these costs would have on the estimated
monetized total annualized costs in this table.
cPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore total
costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To address
stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be regulated as such
in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for illustrative purposes
only. See Appendix N and Section N.2 for additional detail.
Final PFAS Rule Economic Analysis
P-7
April 2024
-------
FINAL RULE
APRIL 2024
P.3 National Annualized Benefits
Table P-7: National Annualized Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt
each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1) (Million $2022)
3% Discount Rate 7% Discount Rate
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Annualized CVD Benefits
$129.38
$557.78
$984.00
$89.33
$392.35
$691.87
Annualized Birth Weight
Benefits
$114.45
$191.42
$268.19
$80.26
$134.65
$188.51
Annualized RCC Benefits
$58.61
$317.71
$777.42
$44.40
$206.04
$469.78
Annualized Bladder
Cancer Benefits
$258.13
$326.65
$398.24
$144.92
$183.45
$223.73
Total Annualized Rule
Benefitsb
$821.07
$1,393.56
$2,053.30
$536.67
$916.49
$1,328.90
Abbreviations: C VD - cardiovascular disease; RCC - renal cell carcinoma.
Note: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule benefits are
not additive across benefit category as the categories are not completely correlated. Quantifiable benefits are increased under
final rule table results relative to the other options presented because of modeled PFHxS occurrence, which results in
additional benefits from co-removed PFOA and PFOS.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-8: National Annualized Benefits, Option la (PFOA and PFOS MCLs of 4.0 ppt)
(Million $2022)
3% Discount Rate 7% Discount Rate
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Annualized CVD Benefits
$128.88
$554.68
$979.99
$88.85
$390.18
$688.72
Annualized Birth Weight
Benefits
$113.38
$190.33
$266.56
$80.00
$133.89
$187.59
Annualized RCC Benefits
$58.40
$315.82
$771.62
$44.19
$204.83
$466.90
Annualized Bladder
Cancer Benefits
$258.48
$326.65
$397.24
$145.11
$183.45
$223.24
Total Annualized Rule
Benefitsb
$815.03
$1,387.48
$2,043.00
$534.22
$912.35
$1,321.70
Abbreviations: C VD - cardiovascular disease; RCC - renal cell carcinoma.
Note: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule benefits are
not additive across benefit category as the categories are not completely correlated.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-8
April 2024
-------
FINAL RULE
APRIL 2024
Table P-9: National Annualized Benefits, Option lb (PFOA and PFOS MCLs of 5.0 ppt)
(Million $2022)
3% Discount Rate 7% Discount Rate
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Annualized CVD
$109.42
$472.36
$828.37
$76.25
$332.29
$583.00
Benefits
Annualized Birth Weight
$98.27
$163.90
$229.43
$68.86
$115.27
$161.46
Benefits
Annualized RCC
$46.81
$261.37
$645.73
$36.03
$170.35
$391.04
Benefits
Annualized Bladder
$211.62
$269.52
$329.18
$118.81
$151.37
$184.69
Cancer Benefits
Total Annualized Rule
$688.91
$1,167.15
$1,722.70
$450.77
$769.28
$1,117.10
Benefitsb
Abbreviations: C VD - cardiovascular disease; RCC - renal cell carcinoma.
Note: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule benefits are
not additive across benefit category as the categories are not completely correlated.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-10: National Annualized Benefits, Option lc (PFOA and PFOS MCLs of 10.0
ppt) (Million $2022)
3% Discount Rate 7% Discount Rate
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Annualized CVD Benefits
$61.50
$246.21
$431.85
$42.83
$173.14
$302.88
Annualized Birth Weight
Benefits
$55.10
$90.63
$126.17
$38.59
$63.70
$88.71
Annualized RCC Benefits
$20.71
$123.87
$310.93
$16.70
$81.75
$189.76
Annualized Bladder
Cancer Benefits
$103.85
$137.92
$173.58
$58.35
$77.46
$97.51
Total Annualized Rule
Benefitsb
$356.37
$598.63
$872.69
$233.73
$396.05
$572.67
Abbreviations: C VD - cardiovascular disease; RCC - renal cell carcinoma.
Note: Detail may not add exactly to total due to independent rounding. 5tli and 95th percentile values for total rule benefits are
not additive across benefit category as the categories are not completely correlated.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-9
April 2024
-------
FINAL RULE
APRIL 2024
P.3.1 National Birth Weight Benefits
Table P-ll: National Birth Weight Benefits, Final Rule (PFOA and PFOS MCLs of 4.0
ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Increase in Birth Weight
129.6
216.8
304.1
129.6
216.8
304.1
(millions of grams)
Number of Birth Weight-
781.9
1,301.7
1,823.6
781.9
1,301.7
1,823.6
Related Deaths Avoided
Total Annualized Birth
$114.45
$191.42
$268.19
$80.26
$134.65
$188.51
Weight Benefits (Million
$2022)b
Note: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-12: National Birth Weight Benefits, Option la (PFOA and PFOS MCLs of 4.0
ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Increase in Birth Weight
128.8
215.6
302.1
128.8
215.6
302.1
(millions of grams)
Number of Birth Weight-
777.4
1,294.4
1,812.9
777.4
1,294.4
1,812.9
Related Deaths Avoided
Total Annualized Birth
$113.38
$190.33
$266.56
$80.00
$133.89
$187.59
Weight Benefits (Million
$2022)b
Note: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-10
April 2024
-------
FINAL RULE APRIL 2024
Table P-13: National Birth Weight Benefits, Option lb (PFOA and PFOS MCLs of 5.0
ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Increase in Birth Weight
111.3
185.6
260.3
111.3
185.6
260.3
(millions of grams)
Number of Birth Weight-
668.9
1,114.7
1,561.2
668.9
1,114.7
1,561.2
Related Deaths Avoided
Total Annualized Birth
$98.27
$163.90
$229.43
$68.86
$115.27
$161.46
Weight Benefits (Million
$2022)b
Note: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-14: National Birth Weight Benefits, Option lc (PFOA and PFOS MCLs of 10.0
ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Increase in Birth Weight
62.1
102.0
142.4
62.1
102.0
142.4
(millions of grams)
Number of Birth Weight-
375.8
616.6
859.1
375.8
616.6
859.1
Related Deaths Avoided
Total Annualized Birth
$55.10
$90.63
$126.17
$38.59
$63.70
$88.71
Weight Benefits (Million
$2022)b
Note: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-ll
April 2024
-------
FINAL RULE
APRIL 2024
P.3.2 National CVD Benefits
Table P-15: National CVD Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each,
PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
1,407.7
6,333.1
11,189.0
1,407.7
6,333.1
11,189.0
MI Cases Avoided
Number of Non-Fatal
2,074.8
9,247.6
16,279.0
2,074.8
9,247.6
16,279.0
IS Cases Avoided
Number of CVD
845.5
3,715.8
6,555.6
845.5
3,715.8
6,555.6
Deaths Avoided
Total Annualized
$129.38
$557.78
$984.00
$89.33
$392.35
$691.87
CVD Benefits
(Million $2022)b
Abbreviations: CVD - cardiovascular disease, MI - myocardial infarction, IS - Ischemic Stroke.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-16: National CVD Benefits, Option la (PFOA and PFOS MCLs of 4.0 ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
1,400.8
6,296.0
11,115.0
1,400.8
6,296.0
11,115.0
MI Cases Avoided
Number of Non-Fatal
2,065.0
9,194.8
16,203.0
2,065.0
9,194.8
16,203.0
IS Cases Avoided
Number of CVD
839.9
3,695.1
6,484.4
839.9
3,695.1
6,484.4
Deaths Avoided
Total Annualized
$128.88
$554.68
$979.99
$88.85
$390.18
$688.72
CVD Benefits
(Million $2022)b
Abbreviations: CVD - cardiovascular disease, MI - myocardial infarction, IS - Ischemic Stroke.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-12
April 2024
-------
FINAL RULE APRIL 2024
Table P-17: National CVD Benefits, Option lb (PFOA and PFOS MCLs of 5.0 ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
1,209.2
5,352.0
9,417.5
1,209.2
5,352.0
9,417.5
MI Cases Avoided
Number of Non-Fatal
1,778.3
7,826.9
13,778.0
1,778.3
7,826.9
13,778.0
IS Cases Avoided
Number of CVD
733.1
3,146.8
5,518.0
733.1
3,146.8
5,518.0
Deaths Avoided
Total Annualized
$109.42
$472.36
$828.37
$76.25
$332.29
$583.00
CVD Benefits
(Million $2022)b
Abbreviations: CVD - cardiovascular disease, MI - myocardial infarction, IS - Ischemic Stroke.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable costs, and the potential direction of impact these costs would have on the
estimated monetized total annualized costs in this table.
Table P-18: National CVD Benefits, Option lc (PFOA and PFOS MCLs of 10.0 ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
673.7
2,776.5
4,872.8
673.7
2,776.5
4,872.8
MI Cases Avoided
Number of Non-Fatal
987.0
4,079.2
7,145.6
987.0
4,079.2
7,145.6
IS Cases Avoided
Number of CVD
411.6
1,640.9
2,878.1
411.6
1,640.9
2,878.1
Deaths Avoided
Total Annualized
$61.50
$246.21
$431.85
$42.83
$173.14
$302.88
CVD Benefits
(Million $2022)b
Abbreviations: CVD - cardiovascular disease, MI - myocardial infarction, IS - Ischemic Stroke.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-13
April 2024
-------
FINAL RULE
APRIL 2024
P.3.3 National RCC Benefits
Table P-19: National RCC Benefits, Final Rule (PFOA and PFOS MCLs of 4.0 ppt each,
PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Number of Non-Fatal RCC
Cases Avoided
1,091.5
6,964.2
17,937.0
1,091.5
6,964.2
17,937.0
Number of RCC-Related
Deaths Avoided
320.4
2,028.8
5,206.5
320.4
2,028.8
5,206.5
Total Annualized RCC
Benefits (Million $2022)b c
$58.61
$317.71
$777.42
$44.40
$206.04
$469.78
Abbreviations: RCC - renal cell carcinoma.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
cWhen using willingness to pay metrics to monetize morbidity benefits, total annualized RCC benefits are increased by $5.7
million at a 3 percent discount rate and by $2.6 million at a 7 percent discount rate (see Appendix O).
Table P-20: National RCC Benefits, Option la (PFOA and PFOS MCLs of 4.0 ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Number of Non-Fatal RCC
Cases Avoided
1,082.0
6,922.4
17,870.0
1,082.0
6,922.4
17,870.0
Number of RCC-Related
Deaths Avoided
319.1
2,016.7
5,190.9
319.1
2,016.7
5,190.9
Total Annualized RCC
Benefits (Million $2022)b
$58.40
$315.82
$771.62
$44.19
$204.83
$466.90
Abbreviations: RCC - renal cell carcinoma.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-14
April 2024
-------
FINAL RULE APRIL 2024
Table P-21: National RCC Benefits, Option lb (PFOA and PFOS MCLs of 5.0 ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Number of Non-Fatal RCC
Cases Avoided
851.9
5,696.1
14,906.0
851.9
5,696.1
14,906.0
Number of RCC-Related
Deaths Avoided
251.6
1,663.8
4,328.4
251.6
1,663.8
4,328.4
Total Annualized RCC
Benefits (Million $2022)b
$46.81
$261.37
$645.73
$36.03
$170.35
$391.04
Abbreviations: RCC - renal cell carcinoma.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-22: National RCC Benefits, Option lc (PFOA and PFOS MCLs of 10.0 ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Number of Non-Fatal RCC
Cases Avoided
372.1
2,648.1
6,967.4
372.1
2,648.1
6,967.4
Number of RCC-Related
Deaths Avoided
111.5
782.8
2,057.3
111.5
782.8
2,057.3
Total Annualized RCC
Benefits (Million $2022)b
$20.71
$123.87
$310.93
$16.70
$81.75
$189.76
Abbreviations: RCC - renal cell carcinoma.
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-15
April 2024
-------
FINAL RULE
APRIL 2024
P.3.4 National Bladder Cancer Benefits
Table P-23: National Bladder Cancer Benefits, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
5,781.0
7,313.0
8,912.7
5,781.0
7,313.0
8,912.7
Bladder Cancer Cases
Avoided
Number of Bladder Cancer-
2,029.6
2,567.8
3,129.9
2,029.6
2,567.8
3,129.9
Related Deaths Avoided
Total Annualized Bladder
$258.13
$326.65
$398.24
$144.92
$183.45
$223.73
Cancer Benefits (Million
$2022)b'c
Notes:
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
cWhen using willingness to pay metrics to monetize morbidity benefits, total annualized bladder cancer benefits are increased
by $65.7 million at a 3 percent discount rate and by $38.6 million at a 7 percent discount rate (see Appendix O).
Table P-24: National Bladder Cancer Benefits, Option la (PFOA and PFOS MCLs of 4.0
ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
5,789.3
7,312.9
8,896.0
5,789.3
7,312.9
8,896.0
Bladder Cancer Cases
Avoided
Number of Bladder Cancer-
2,032.5
2,567.8
3,123.2
2,032.5
2,567.8
3,123.2
Related Deaths Avoided
Total Annualized Bladder
$258.48
$326.65
$397.24
$145.11
$183.45
$223.24
Cancer Benefits (Million
$2022)b
Notes:
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-16
April 2024
-------
FINAL RULE APRIL 2024
Table P-25: National Bladder Cancer Benefits, Option lb (PFOA and PFOS MCLs of 5.0
ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
4,739.4
6,034.0
7,367.1
4,739.4
6,034.0
7,367.1
Bladder Cancer Cases
Avoided
Number of Bladder Cancer-
1,664.0
2,118.7
2,587.1
1,664.0
2,118.7
2,587.1
Related Deaths Avoided
Total Annualized Bladder
$211.62
$269.52
$329.18
$118.81
$151.37
$184.69
Cancer Benefits (Million
$2022)b
Notes:
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-26: National Bladder Cancer Benefits, Option lc (PFOA and PFOS MCLs of
10.0 ppt)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
2,326.9
3,087.9
3,885.3
2,326.9
3,087.9
3,885.3
Bladder Cancer Cases
Avoided
Number of Bladder Cancer-
816.8
1,084.3
1,364.3
816.8
1,084.3
1,364.3
Related Deaths Avoided
Total Annualized Bladder
$103.85
$137.92
$173.58
$58.35
$77.46
$97.51
Cancer Benefits (Million
$2022)b
Notes:
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized costs in this table.
Final PFAS Rule Economic Analysis
P-17
April 2024
-------
FINAL RULE
APRIL 2024
P.4 Comparison of Costs and Benefits
Table P-27: Annualized Quantified National Costs and Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of
1) (Million $2022)
3% Discount Rate 7% Discount Rate
5th
Mean
95th
5th
Mean
95th
Percentile3
Percentile3
Percentile3
Percentile3
Total Annualized Rule
$1,431.50
$1,545.61
$1,670.10
$1,437.00
$1,553.98
$1,688.00
Costs
Total Annualized Rule
$821.07
$1,393.56
$2,053.30
$536.67
$916.49
$1,328.90
Benefits
Total Net Benefitsb c d
-$717.96
-$152.05
$494.34
-$1,022.20
-$637.49
-$224.87
Notes: Detail may not add exactly to total due to independent rounding. Quantifiable benefits are increased under final rule
table results relative to the other options presented because of modeled PFHxS occurrence, which results in additional
benefits from co-removed PFOA and PFOS.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1
for costs and Section 6.1.2 and Table 6-1 for benefits. This range does not include the uncertainty described in 5-22 for costs
and Table 6-48 for benefits.
bSee Table 7-6 for a list of the nonquantifiable benefits and costs, and the potential direction of impact these benefits and
costs would have on the estimated monetized total annualized benefits and costs in this table.
The national level cost estimates for PFFLxS are reflective of both the total national cost for PFHxS individual MCL
exceedances, and HI MCL exceedances where PFHxS is present above its HBWC while one or more other HI PFAS is also
present in that same mixture. Total quantified national cost values do not include the incremental treatment costs associated
with the co-occurrence of HFPO-DA, PFBS, and PFNA. EPA has considered the additional national costs of the HI and
individual MCLs associated with HFPO-DA, PFNA, and PFBS occurrence in a quantified sensitivity analysis; see Appendix
N and Section N.3 for the analysis and more information.
dPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and
therefore total costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration
materials. To address stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous
should they be regulated as such in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous
waste disposal for illustrative purposes only. See Appendix N and Section N.2 for additional detail.
Final PFAS Rule Economic Analysis
P-18
April 2024
-------
FINAL RULE
APRIL 2024
/
/
Prob>0: 31.5%
-1500
-1000
-500
0
500
1000
Net Benefits (Million $)
Figure P-l: Distribution of Estimated Net Quantified Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1;
3 percent Discount Rate; Million $2022)
Final PFAS Rule Economic Analysis
P-19
April 2024
-------
FINAL RULE
APRIL 2024
Figure P-2: Distribution of Estimated Net Quantified Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA MCLs of 10 ppt each and HI of 1;
7 percent Discount Rate; Million $2022)
Final PFAS Rule Economic Analysis
P-20
April 2024
-------
FINAL RULE
APRIL 2024
Table P-28: Annualized Quantified National Costs and Benefits, Option la (PFOA and
PFOS MCLs of 4.0 ppt) (Million $2022)
3% Discount Rate 7% Discount Rate
5th
Mean
95th
5th
Mean
95th
Percentile3
Percentile3
Percentile3
Percentile3
Total Annualized Rule
$1,420.30
$1,534.03
$1,658.20
$1,425.50
$1,542.57
$1,676.70
Costs
Total Annualized Rule
$815.03
$1,387.48
$2,043.00
$534.22
$912.35
$1,321.70
Benefits
Total Net Benefitsb c
-$709.19
-$146.55
$498.73
-$1,013.40
-$630.22
-$219.47
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1
for costs and Section 6.1.2 and Table 6-1 for benefits. This range does not include the uncertainty described in Table 5-22 for
costs and Table 6-48 for benefits.
bSee Table 7-6 for a list of the nonquantifiable benefits and costs, and the potential direction of impact these benefits and
costs would have on the estimated monetized total annualized benefits and costs in this table.
cPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore
total costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To
address stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be
regulated as such in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for
illustrative purposes only. See Appendix N and Section N.2 for additional detail.
Table P-29: Annualized Quantified National Costs and Benefits, Option lb (PFOA and
PFOS MCLs of 5.0 ppt) (Million $2022)
3% Discount Rate 7% Discount Rate
5th
Mean
95th
5th
Mean
95th
Percentile3
Percentile3
Percentile3
Percentile3
Total Annualized Rule
$1,100.10
$1,189.99
$1,290.30
$1,103.90
$1,197.32
$1,304.10
Costs
Total Annualized Rule
$688.91
$1,167.15
$1,722.70
$450.77
$769.28
$1,117.10
Benefits
Total Net Benefitsb c
-$496.16
-$22.84
$517.44
-$748.65
-$428.04
-$79.59
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1
for costs and Section 6.1.2 and 6-1 for benefits. This range does not include the uncertainty described in Table 5-22 for costs
and Table 6-48 for benefits.
bSee Table 7-6 for a list of the nonquantifiable benefits and costs, and the potential direction of impact these benefits and
costs would have on the estimated monetized total annualized benefits and costs in this table.
cPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore
total costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To
address stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be
regulated as such in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for
illustrative purposes only. See Appendix N and Section N.2 for additional detail.
Final PFAS Rule Economic Analysis
P-21
April 2024
-------
FINAL RULE
APRIL 2024
Table P-30: Annualized Quantified National Costs and Benefits, Option lc (PFOA and
PFOS MCLs of 10.0 ppt) (Million $2022)
3% Discount Rate 7% Discount Rate
5th
Percentile3
Mean
95th
Percentile3
5th
Percentile3
Mean
95th
Percentile3
Total Annualized Rule
$461.72
$498.64
$540.36
$464.77
$503.02
$547.76
Costs
Total Annualized Rule
$356.37
$598.63
$872.69
$233.73
$396.05
$572.67
Benefits
Total Net Benefitsb c
-$136.94
$99.99
$370.06
-$270.13
-$106.98
$68.02
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty described in Section 5.1.2 and Table 5-1
for costs and Section 6.1.2 and Table 6-1 for benefits. This range does not include the uncertainty described in Table 5-22 for
costs and Table 6-48 for benefits.
bSee Table 7-6 for a list of the nonquantifiable benefits and costs, and the potential direction of impact these benefits and
costs would have on the estimated monetized total annualized benefits and costs in this table.
cPFAS-contaminated wastes are not considered RCRA regulatory or characteristic hazardous wastes at this time and therefore
total costs reported in this table do not include costs associated with hazardous waste disposal of spent filtration materials. To
address stakeholder concerns about potential costs for disposing PFAS-contaminated wastes as hazardous should they be
regulated as such in the future, the EPA conducted a sensitivity analysis with an assumption of hazardous waste disposal for
illustrative purposes only. See Appendix N and Section N.2 for additional detail.
Final PFAS Rule Economic Analysis
P-22
April 2024
-------
FINAL RULE
APRIL 2024
P.5 Benefits Sensitivity Analyses
Table P-31: Summary of CVD Sensitivity Analysis for Hypothetical Exposure Reduction 1 (PFOA+PFOS)
Exposure-Response Scenario13
Result Description3
W
ex
s
o
o
s
5
u
-j
o
n
<
w
ex U
S J
® M
u
-j
a
n
+
a.
as
<
w
a.
CO
ex
s
o
Q
a.
CO
s
5
Sa
^ £5
W I
1 X
o +
&
e O
£5 W
¦ +
.5 Q
fN +
0.014 -0.002
0.004 0.004
Average reduction in serum 0.091 0.091 0.091 0.091 0.091
PFOA concentration (ng/mL)
Average reduction in serum 0.084 0.084 0.084 0.084 0.084
PFOS concentration (ng/mL)
Average reduction in TC 0.150 0.168 0.160 0.150 0.168
concentration (mg/dL)
Average reduction in HDLC 0.000 0.000 0.000
concentration (mg/dL)
Average reduction in BP 0.004 0.004 0.004
(mmHg)
Non-fatal first MI (total cases 2.745 3.084 2.920 1.973 3.187
avoided)d
Non-fatal first IS (total cases 3.965 4.455 4.218 3.005 4.583
avoided)d
CVD deaths (total cases 0.778 0.875 0.828 0.641 0.893
avoided)d
PDV, non-fatal first MI (3% 0.104 0.117 0.111 0.074 0.121
discount rate, millions
$2022)
PDV, non-fatal first IS (3% 0.043 0.048 0.046 0.032 0.050
discount rate, millions
$2022)
PDV, CVD deaths (3% 5.090 5.707 5.410 3.989 5.854
discount rate, millions
$2022)
0.091
0.084
0.160
-0.014
0.004
3.654
5.130
0.958
0.139
0.056
6.458
0.091
0.084
0.150
0.000
0.000
2.708
3.909
0.755
0.103
0.042
4.973
0.091
0.084
0.168
0.000
0.000
3.048
4.399
0.852
0.115
0.048
5.590
0.091
0.084
0.160
0.000
0.000
2.883
4.161
0.804
0.109
0.045
5.294
0.091
0.084
0.150
0.014
0.000
1.936
2.948
0.618
0.073
0.032
3.872
0.091
0.084
0.168
-0.002
0.000
3.150
4.526
0.870
0.119
0.049
5.737
0.091
0.084
0.160
-0.014
0.000
3.618
5.073
0.935
0.138
0.056
6.341
Final PFAS Rule Economic Analysis
P-23
April 2024
-------
FINAL RULE APRIL 2024
Table P-31: Summary of CVD Sensitivity Analysis for Hypothetical Exposure Reduction 1 (PFOA+PFOS)
Exposure-Response Scenario13 0
Result Description3
•< S -5 u i ^
^ p ^ + fig + r g* ^ ^ogo.so
^ i w ^ 5 -5 W M a M J M
< .a ^ i? 2? ^
^ ^ ^ si ^ S -H
^ s£
PDV, total CVD benefits (3% 5.237 5.872 5.567 4.095 6.024 6.653 5.119 5.753 5.449 3.977 5.906 6.535
discount rate, millions
$2022)
Annualized CVD benefits 0.172 0.193 0.183 0.135 0.198 0.219 0.168 0.189 0.179 0.131 0.194 0.215
(3% discount rate, millions
$2022)
PDV, non-fatal first MI (7% 0.040 0.045 0.043 0.029 0.046 0.053 0.039 0.044 0.042 0.028 0.046 0.053
discount rate, millions
$2022)
PDV, non-fatal first IS (7% 0.017 0.019 0.018 0.013 0.020 0.022 0.017 0.019 0.018 0.012 0.019 0.022
discount rate, millions
$2022)
PDV, CVD deaths (7% 2.286 2.553 2.429 1.733 2.629 2.957 2.245 2.512 2.388 1.692 2.588 2.916
discount rate, millions
$2022)
PDV, total CVD benefits (7% 2.343 2.617 2.489 1.774 2.695 3.033 2.301 2.575 2.447 1.732 2.653 2.991
discount rate, millions
$2022)
Annualized CVD benefits 0.165 0.184 0.175 0.125 0.189 0.213 0.162 0.181 0.172 0.122 0.186 0.210
(7% discount rate, millions
$2022)
Abbreviations: PFOA - perfluorooctanoic acid; PFOS - perfluorooctane sulfonic acid; TC - total cholesterol; HDLC - high-density lipoprotein cholesterol; BP -
systolic blood pressure; CVD - cardiovascular disease; EA - economic analysis; SAB - Science Advisory Board; MI - myocardial infarction; IS - ischemic stroke;
PDV - present discounted value.
Notes:
aSee Table K-l
bSee Table K-3
cNegative values refer to increases in a particular result (e.g., the HDLC reduction of -0.002 mg/dL in Scenario 2-Dong refers to an increase in HDLC).
dTotal over the period of analysis.
Final PFAS Rule Economic Analysis
P-24
April 2024
-------
FINAL RULE APRIL 2024
Table P-32: Summary of Birth Weight Sensitivity Analysis
Hypothetical Exposure Reduction3 /
Exposure-Response Scenario13
Result Description (PFOA+PFOS) (PFOA+PFOS+PFNA)
1-EA 2-First 3-EA+Lenters 4-EA+Valvi
Trimester
Average reduction in serum PFOA
0.089
0.089
0.089
0.089
concentration (ng/mL)
Average reduction in serum PFOS
0.081
0.081
0.081
0.081
concentration (ng/mL)
Average reduction in serum PFNA
0.000
0.000
0.136
0.136
concentration (ng/mL)
Total increase in birth weight (g)
1.180
0.404
6.654
9.320
Total number of births affected0
102,268
102,268
102,268
102,268
Total number of surviving births affected0
101,804
101,803
101,806
101,808
Birth weight-related deaths (total cases
0.616
0.211
4.841
4.841
avoided)0
PDV, birth weight-related deaths (3% discount
2.724
0.932
15.133
21.144
rate, millions $2022)
PDV, birth weight-related morbidity (3%
0.083
0.028
0.462
0.646
discount rate, millions $2022)
PDV, total birth weight benefits (3% discount
2.807
0.960
15.595
21.791
rate, millions $2022)
Annualized birth weight benefits (3%
0.092
0.032
0.513
0.717
discount rate, millions $2022)
PDV, birth weight-related deaths (7% discount
0.882
0.301
4.804
6.704
rate, millions $2022)
PDV, birth weight-related morbidity (7%
0.029
0.010
0.157
0.219
discount rate, millions $2022)
PDV, total birth weight benefits (7% discount
0.910
0.311
4.961
6.923
rate, millions $2022)
Annualized birth weight benefits (7%
0.064
0.022
0.349
0.487
discount rate, millions $2022)
Abbreviations: PDV - present discounted value; PFNA - perfluorononanoic acid; PFOA - perfluorooctanoic acid; PFOS -
perfluorooctane sulfonic acid.
Notes:
aSee Table K-l
bSee Table K-5
cTotal over the period of analysis.
Table P-33: Summary of RCC Sensitivity Analysis
Exposure-Response Scenario3
Result Description
1-EA 2-Vieira 3- VieiraExciudeHigh
Average reduction in serum PFOA
concentration (ng/mL)
Non-fatal RCC (cases avoided)
0.085
9.329
0.085
0.365
0.085
1.295
Final PFAS Rule Economic Analysis
P-25
April 2024
-------
FINAL RULE
APRIL 2024
Table P-33: Summary of RCC Sensitivity Analysis
3.762
0.147
0.522
RCC-related deaths (cases avoided)13
PDV, Non-fatal RCC (3% discount
1.530
0.060
0.212
rate, millions $2022)
PDV, RCC-related deaths (3%
14.696
0.574
2.039
discount rate, millions $2022)
PDV, total RCC benefits (3%
16.226
0.634
2.251
discount rate, millions $2022)
Annualized RCC benefits (3%
0.534
0.021
0.074
discount rate, millions $2022)
PDV, Non-fatal RCC (7% discount
0.444
0.017
0.062
rate, millions $2022)
PDV, RCC-related deaths (7%
3.834
0.150
0.532
discount rate, millions $2022)
PDV, total RCC benefits (7%
4.278
0.167
0.593
discount rate, millions $2022)
Annualized RCC benefits (7%
0.301
0.012
0.042
discount rate, millions $2022)
Abbreviations: PDV - present discounted value; PFOA - pertluorooctanoic acid; RCC - renal cell carcinoma.
Notes:
aSee Table K-8.
bTotal over the period of analysis.
P.6 Supplemental Cost Analyses
Table P-34: Annualized PWS Treatment Cost Associated with Non-Hazardous and
Hazardous Residual Management Requirements, Final Rule (PFOA and PFOS MCLs of
4.0 ppt each, PFHxS, PFNA, and HFPO-DA MCLs of 10 ppt each and HI of 1) (Million
$2022)
3% Discount Rate
7% Discount Rate
5th Mean 95th 5th
Percentile Percentile Percentile
Mean
95th
Percentile
Non-Hazardous $1,391.16 $1,501.68 $1,624.89 $1,388.69 $1,503.01 $1,634.84
Disposal
Hazardous $1,480.76 $1,598.08 $1,725.79 $1,470.69 $1,590.41 $1,725.64
Disposal
Increase due to
Hazardous
Disposal
$96.40
$87.40
Note: Percentiles cannot be subtracted.
Final PFAS Rule Economic Analysis
P-26
April 2024
-------
FINAL RULE
APRIL 2024
P.7 Supplemental Benefits Analyses
Table P-35. National Liver Cancer Benefits, Final Rule (PFOA and PFOS MCLs of 4.0
ppt each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Number of Non-Fatal Liver
Cancer Cases Avoided
13.3
14.2
15.1
13.3
14.2
15.1
Number of Liver Cancer-
Related Deaths Avoided
29.4
31.3
33.3
29.4
31.2
33.3
Total Annualized Liver
Cancer Benefits (Million
$2022)b
$3.81
$4.05
$4.31
$1.97
$2.10
$2.23
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty related to PAF and occurrence. This range
does not include the uncertainty described in Table 0-3.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Table P-36. National Willingness to Pay-Based RCC Benefits, Final Rule (PFOA and
PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA, of 10 ppt each and HI of 1)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Percentile3
Expected
Value
95th
Percentile3
5th
Percentile3
Expected
Value
95th
Percentile3
Number of Non-Fatal RCC
Cases Avoided
1091.50
6964.20
17937.00
1,091.50
6,964.20
17,937.00
Number of RCC-Related
Deaths Avoided
320.36
2028.80
5206.50
320.36
2,028.80
5,206.50
Total Annualized RCC
Benefits (Million $2022)b
$59.08
$323.40
$793.48
$44.80
$208.56
$477.05
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-27
April 2024
-------
FINAL RULE
APRIL 2024
Table P-37. National Willingness to Pay-Based Bladder Cancer Benefits, Final Rule
(PFOA and PFOS MCLs of 4.0 ppt each, PFHxS, PFNA, HFPO-DA of 10 ppt each and
HI of 1)
3% Discount Rate 7% Discount Rate
Benefits Category
5th
Expected
95th
5th
Expected
95th
Percentile3
Value
Percentile3
Percentile3
Value
Percentile3
Number of Non-Fatal
5,781.00
7,313.00
8,912.70
5,781.00
7,313.00
8,912.70
Bladder Cancer Cases
Avoided
Number of Bladder Cancer-
2,029.60
2,567.80
3,129.90
2,029.60
2,567.80
3,129.90
Related Deaths Avoided
Total Annualized Bladder
$310.08
$392.38
$478.37
$175.42
$222.06
$270.81
Cancer Benefits (Million
$2022)b
Notes: Detail may not add exactly to total due to independent rounding.
aThe 5th and 95th percentile range is based on modeled variability and uncertainty. This range does not include the uncertainty
described in Table 6-48.
bSee Table 7-6 for a list of the nonquantifiable benefits, and the potential direction of impact these benefits would have on the
estimated monetized total annualized benefits in this table.
Final PFAS Rule Economic Analysis
P-28
April 2024
-------
FINAL RULE
APRIL 2024
P.8 Undiscounted Benefits and Costs
Table P-38. Quantified Total National Annual Costs, Final Rule (Undiscounted, Million $2022)
Year
Primacy
Agency
Administration
Primacy
Agency
Sampling
Review
Primacy
Agency
Treatment
Plan
Review
PWS
Treatment
Capital
PWS
Treatment
Operations
and
Maintenance
PWS
Administration
PWS
Sampling
PWS
Treatment
Plan
Submittal
Primacy
Agency
Total
PWS Total
Rule Total
2024
$6.48
$9.59
$0.00
$0.00
$0.00
$18.49
$92.76
$0.00
$16.08
$111.25
$127.32
2025
$6.48
$0.00
$0.00
$0.00
$0.00
$18.49
$0.00
$0.00
$6.48
$18.49
$24.97
2026
$6.48
$0.00
$0.00
$0.00
$0.00
$18.49
$0.00
$0.00
$6.48
$18.49
$24.97
2027
$0.00
$10.44
$0.00
$0.00
$0.00
$0.00
$113.71
$0.00
$10.44
$113.71
$124.15
2028
$0.00
$5.00
$35.25
$0.00
$0.00
$0.00
$63.55
$6.21
$40.24
$69.76
$110.01
2029
$0.00
$5.00
$0.00
$14,378.00
$1,023.30
$0.00
$63.55
$0.00
$5.00
$15,464.85
$15,469.85
2030
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2031
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2032
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2033
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2034
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2035
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2036
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2037
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2038
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2039
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2040
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2041
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2042
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2043
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2044
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2045
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2046
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2047
$0.00
$1.25
$0.00
$363.67
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,402.86
$1,404.11
2048
$0.00
$6.70
$0.00
$405.22
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,494.57
$1,501.26
2049
$0.00
$1.25
$0.00
$183.95
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,223.14
$1,224.39
2050
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2051
$0.00
$6.70
$0.00
$2,095.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$3,184.35
$3,191.04
2052
$0.00
$1.25
$0.00
$126.49
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,165.68
$1,166.93
Final PFAS Rule Economic Analysis
P-29
April 2024
-------
FINAL RULE APRIL 2024
Table P-38. Quantified Total National Annual Costs, Final Rule (Undiscounted, Million $2022)
Year
Primacy
Agency
Administration
Primacy
Agency
Sampling
Review
Primacy
Agency
Treatment
Plan
Review
PWS
Treatment
Capital
PWS
Treatment
Operations
and
Maintenance
PWS
Administration
PWS
Sampling
PWS
Treatment
Plan
Submittal
Primacy
Agency
Total
PWS Total
Rule Total
2053
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2054
$0.00
$6.70
$0.00
$101.26
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,190.61
$1,197.30
2055
$0.00
$1.25
$0.00
$27.17
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,066.35
$1,067.60
2056
$0.00
$1.25
$0.00
$1.95
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,041.14
$1,042.38
2057
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2058
$0.00
$1.25
$0.00
$27.31
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,066.50
$1,067.75
2059
$0.00
$1.25
$0.00
$40.26
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,079.45
$1,080.69
2060
$0.00
$6.70
$0.00
$23.34
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,112.68
$1,119.38
2061
$0.00
$1.25
$0.00
$1.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,040.19
$1,041.43
2062
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2063
$0.00
$6.70
$0.00
$1,960.20
$1,023.30
$0.00
$66.05
$0.00
$6.70
$3,049.55
$3,056.24
2064
$0.00
$1.25
$0.00
$5,851.40
$1,023.30
$0.00
$15.89
$0.00
$1.25
$6,890.59
$6,891.84
2065
$0.00
$1.25
$0.00
$2,373.40
$1,023.30
$0.00
$15.89
$0.00
$1.25
$3,412.59
$3,413.84
2066
$0.00
$6.70
$0.00
$1,151.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$2,240.35
$2,247.04
2067
$0.00
$1.25
$0.00
$405.22
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,444.41
$1,445.66
2068
$0.00
$1.25
$0.00
$8.97
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,048.16
$1,049.41
2069
$0.00
$6.70
$0.00
$183.95
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,273.30
$1,279.99
2070
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2071
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2072
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2073
$0.00
$1.25
$0.00
$2,095.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$3,134.19
$3,135.44
2074
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2075
$0.00
$6.70
$0.00
$126.49
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,215.84
$1,222.53
2076
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2077
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2078
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2079
$0.00
$1.25
$0.00
$101.26
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,140.45
$1,141.70
2080
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2081
$0.00
$6.70
$0.00
$27.17
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,116.51
$1,123.21
2082
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2083
$0.00
$1.25
$0.00
$365.62
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,404.81
$1,406.06
2084
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2085
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
Final PFAS Rule Economic Analysis
P-30
April 2024
-------
FINAL RULE APRIL 2024
Table P-38. Quantified Total National Annual Costs, Final Rule (Undiscounted, Million $2022)
Year
Primacy
Agency
Administration
Primacy
Agency
Sampling
Review
Primacy
Agency
Treatment
Plan
Review
PWS
Treatment
Capital
PWS
Treatment
Operations
and
Maintenance
PWS
Administration
PWS
Sampling
PWS
Treatment
Plan
Submittal
Primacy
Agency
Total
PWS Total
Rule Total
2086
$0.00
$1.25
$0.00
$405.22
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,444.41
$1,445.66
2087
$0.00
$6.70
$0.00
$27.31
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,116.66
$1,123.36
2088
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2089
$0.00
$1.25
$0.00
$224.21
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,263.40
$1,264.65
2090
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2091
$0.00
$1.25
$0.00
$23.34
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,062.52
$1,063.77
2092
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2093
$0.00
$6.70
$0.00
$1.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,090.34
$1,097.04
2094
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2095
$0.00
$1.25
$0.00
$2,095.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$3,134.19
$3,135.44
2096
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2097
$0.00
$1.25
$0.00
$1,960.20
$1,023.30
$0.00
$15.89
$0.00
$1.25
$2,999.39
$3,000.64
2098
$0.00
$1.25
$0.00
$126.49
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,165.68
$1,166.93
2099
$0.00
$6.70
$0.00
$5,851.40
$1,023.30
$0.00
$66.05
$0.00
$6.70
$6,940.75
$6,947.44
2100
$0.00
$1.25
$0.00
$0.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,039.19
$1,040.44
2101
$0.00
$1.25
$0.00
$2,373.40
$1,023.30
$0.00
$15.89
$0.00
$1.25
$3,412.59
$3,413.84
2102
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
2103
$0.00
$1.25
$0.00
$1,151.00
$1,023.30
$0.00
$15.89
$0.00
$1.25
$2,190.19
$2,191.44
2104
$0.00
$1.25
$0.00
$101.26
$1,023.30
$0.00
$15.89
$0.00
$1.25
$1,140.45
$1,141.70
2105
$0.00
$6.70
$0.00
$0.00
$1,023.30
$0.00
$66.05
$0.00
$6.70
$1,089.35
$1,096.04
Final PFAS Rule Economic Analysis
P-31
April 2024
-------
FINAL RULE
APRIL 2024
Table P-39. Quantified Total National Annual Benefits, Final Rule (Undiscounted,
Million $2022)
Year
Birth Weight
CVD
RCC
Bladder
Cancer
Rule Total
2024
$0.00
$0.00
$0.00
$0.00
$0.00
2025
$0.00
$0.00
$0.00
$0.00
$0.00
2026
$0.00
$0.00
$0.00
$0.00
$0.00
2027
$0.00
$0.00
$0.00
$0.00
$0.00
2028
$0.00
$0.00
$0.00
$0.00
$0.00
2029
$29.20
$62.83
$57.07
$61.84
$210.93
2030
$72.24
$156.58
$132.60
$84.70
$446.12
2031
$105.92
$245.95
$191.16
$100.26
$643.29
2032
$132.31
$330.69
$234.03
$115.73
$812.76
2033
$153.04
$410.55
$265.09
$130.92
$959.60
2034
$169.35
$481.69
$288.48
$145.13
$1,084.65
2035
$182.24
$540.26
$307.52
$160.02
$1,190.04
2036
$192.48
$587.20
$322.59
$174.56
$1,276.83
2037
$200.68
$624.79
$335.02
$189.25
$1,349.74
2038
$207.30
$654.85
$345.66
$203.90
$1,411.71
2039
$212.70
$678.67
$355.08
$216.60
$1,463.05
2040
$217.17
$697.45
$364.16
$229.47
$1,508.25
2041
$220.93
$712.55
$372.19
$242.81
$1,548.48
2042
$224.14
$724.65
$379.50
$256.45
$1,584.74
2043
$226.94
$734.33
$386.34
$270.26
$1,617.87
2044
$229.42
$741.94
$392.88
$284.17
$1,648.41
2045
$231.65
$747.55
$399.25
$297.63
$1,676.08
2046
$233.69
$751.89
$405.12
$311.21
$1,701.91
2047
$235.60
$755.25
$410.67
$324.80
$1,726.32
2048
$237.41
$757.81
$416.01
$338.32
$1,749.55
2049
$239.15
$759.72
$421.23
$351.73
$1,771.83
2050
$240.83
$760.91
$426.18
$364.15
$1,792.07
2051
$242.47
$761.71
$430.90
$376.48
$1,811.56
2052
$244.10
$762.35
$435.46
$388.69
$1,830.60
2053
$245.70
$762.94
$439.97
$400.77
$1,849.38
2054
$247.29
$763.58
$444.46
$412.69
$1,868.02
2055
$248.89
$764.30
$448.74
$423.48
$1,885.41
2056
$250.47
$765.15
$452.94
$434.20
$1,902.76
2057
$252.07
$766.25
$457.12
$444.87
$1,920.31
2058
$253.66
$767.67
$461.32
$455.52
$1,938.17
2059
$255.26
$769.49
$465.54
$466.16
$1,956.45
2060
$256.87
$771.70
$469.66
$476.17
$1,974.40
2061
$258.49
$774.19
$473.85
$486.30
$1,992.83
2062
$260.11
$776.97
$478.08
$496.58
$2,011.74
2063
$261.74
$780.07
$482.38
$507.02
$2,031.21
2064
$263.39
$783.51
$486.71
$517.63
$2,051.24
2065
$265.04
$785.85
$490.67
$528.25
$2,069.81
2066
$266.70
$787.99
$494.68
$539.14
$2,088.51
2067
$268.37
$789.88
$498.75
$550.29
$2,107.29
2068
$270.06
$791.52
$502.85
$561.68
$2,126.11
2069
$271.75
$792.96
$506.96
$573.30
$2,144.97
2070
$273.45
$793.98
$510.52
$585.25
$2,163.20
2071
$275.17
$794.94
$514.11
$597.39
$2,181.61
2072
$276.89
$795.80
$517.72
$609.68
$2,200.09
2073
$278.63
$796.58
$521.34
$622.11
$2,218.66
Final PFAS Rule Economic Analysis
P-32
April 2024
-------
FINAL RULE
APRIL 2024
Table P-39. Quantified Total National Annual Benefits, Final Rule (Undiscounted,
Million $2022)
Year
Birth Weight
CVD
RCC
Bladder
Cancer
Rule Total
2074
$280.37
$797.24
$524.98
$634.64
$2,237.23
2075
$282.13
$797.63
$527.76
$647.13
$2,254.65
2076
$283.90
$797.89
$530.56
$659.50
$2,271.85
2077
$285.68
$798.07
$533.39
$671.79
$2,288.93
2078
$287.47
$798.17
$536.25
$684.00
$2,305.89
2079
$289.27
$798.20
$539.16
$696.13
$2,322.76
2080
$291.09
$798.11
$540.87
$707.46
$2,337.53
2081
$292.91
$798.00
$542.68
$718.38
$2,351.97
2082
$294.75
$797.97
$544.59
$729.05
$2,366.36
2083
$296.60
$798.04
$546.63
$739.55
$2,380.82
2084
$298.46
$798.30
$548.75
$749.90
$2,395.41
2085
$300.33
$798.60
$549.73
$759.15
$2,407.81
2086
$302.21
$798.99
$550.95
$767.94
$2,420.09
2087
$304.10
$799.55
$552.41
$776.54
$2,432.60
2088
$306.01
$800.29
$554.06
$785.06
$2,445.42
2089
$307.93
$801.27
$555.89
$793.57
$2,458.66
2090
$309.86
$802.03
$556.61
$800.58
$2,469.08
2091
$311.80
$803.04
$557.67
$807.24
$2,479.75
2092
$313.76
$804.29
$559.02
$813.78
$2,490.85
2093
$315.72
$805.75
$560.63
$820.25
$2,502.35
2094
$317.70
$807.44
$562.49
$826.80
$2,514.43
2095
$319.70
$809.23
$564.20
$832.04
$2,525.17
2096
$321.70
$811.09
$566.42
$836.97
$2,536.18
2097
$323.72
$813.13
$558.61
$841.87
$2,537.33
2098
$325.75
$815.33
$552.43
$846.86
$2,540.37
2099
$327.79
$817.72
$544.46
$852.01
$2,541.98
2100
$329.84
$820.26
$529.89
$853.24
$2,533.23
2101
$331.91
$822.92
$514.44
$854.25
$2,523.52
2102
$333.99
$825.67
$493.80
$855.62
$2,509.08
2103
$336.09
$828.47
$466.19
$857.60
$2,488.35
2104
$338.19
$831.38
$426.39
$860.23
$2,456.19
2105
$340.31
$834.39
$357.62
$862.30
$2,394.62
Final PFAS Rule Economic Analysis
P-33
April 2024
-------
FINAL RULE
APRIL 2024
Appendix Q. Appendix References
Abt Associates. (2022). Estimated Values of Avoiding Cancer Risks by Cancer Site and
Population. Rockville, MD
Agency for Healthcare Research and Quality. (2011). MEPSHC-128: 2009 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.isp?cboPuf
Numb er=HC -12 8
Agency for Healthcare Research and Quality. (2012a). MEPSHC-137: 2010 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.isp?cboPuf
Number=HC-137
Agency for Healthcare Research and Quality. (2012b). MEPS HC-139: Panel 14 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Numb er=HC -13 9
Agency for Healthcare Research and Quality. (2013a). MEPSHC-146: 2011 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-146
Agency for Healthcare Research and Quality. (2013b). MEPS HC-148: Panel 15 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Numb er=HC -14 8
Agency for Healthcare Research and Quality. (2014a). MEPSHC-154: 2012 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-154
Agency for Healthcare Research and Quality. (2014b). MEPSHC-156: Panel 16 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Numb er=HC -15 6
Agency for Healthcare Research and Quality. (2015a). MEPS HC-162: 2013 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-162
Agency for Healthcare Research and Quality. (2015b). MEPS HC-164: Panel 17 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-164
Agency for Healthcare Research and Quality. (2016a). MEPS HC-170: 2014 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Numb er=HC -170
Agency for Healthcare Research and Quality. (2016b). MEPS HC-172: Panel 18 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
Final PFAS Rule Economic Analysis
Q-l
April 2024
-------
FINAL RULE
APRIL 2024
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.isp?cboPuf
Number=HC-172
Agency for Healthcare Research and Quality. (2017a). Healthcare Cost and Utilization Project.
Retrieved from: https://hcupnet.ahrq.gov
Agency for Healthcare Research and Quality. (2017b). MEPSHC-180: 2015 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.isp?cboPuf
Number=HC-180
Agency for Healthcare Research and Quality. (2017c). MEPSHC-183: Panel 19 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-183
Agency for Healthcare Research and Quality. (2018). MEPSHC-193: Panel 20 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-193
Agency for Healthcare Research and Quality. (2019a). MEPS HC-190: 2016Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-190
Agency for Healthcare Research and Quality. (2019b). MEPS HC-199: 2017 Medical Conditions
File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-199
Agency for Healthcare Research and Quality. (2019c). MEPSHC-202: Panel 21 Longitudinal
Data File. Medical Expenditure Panel Survey. Retrieved from:
https://www.meps.ahrq.gov/mepsweb/data stats/download data files detail.!sp?cboPuf
Number=HC-202
Agency for Toxic Substances and Disease Registry. (2018). Toxicological Profile for
Perflaoroalkyls.
Ahrens, K. A., Rossen, L. M., Thoma, M. E., Warner, M., & Simon, A. E. (2017). Birth order
and injury-related infant mortality in the U.S. American Journal of Preventive Medicine,
55(4), 412-420.
Alio, A. P., Mbah, A. K., Kornosky, J. L., Wathington, D., Marty, P. J., & Salihu, H. M. (2011).
Assessing the impact of paternal involvement on racial/ethnic disparities in infant
mortality rates. Journal of Community Health, 36, 63-68. doi:10.1007/sl0900-010-9280-
3
Almond, D., Chay, K. Y., & Lee, D. S. (2005). The costs of low birth weight. The Quarterly
journal of economics, 120(3), 1031-1083. doi: 10.1093/qj e/120.3.1031
Ambavane, A., Yang, S., Atkins, M. B., Rao, S., Shah, A., Regan, M. M., . . . Michaelson, M. D.
(2020). Clinical and economic outcomes of treatment sequences for intermediate- to
poor-risk advanced renal cell carcinoma. Immunotherapy, 12(1), 37-51. doi:10.2217/imt-
2019-0199
Anderson, B. B., Ukah, F., Tette, A., Villaflor, S.-G., Koh, D., & Seton, P. (1992). Primary
tumors of the liver. Journal of the National Medical Association, 84(2), 129.
Final PFAS Rule Economic Analysis
Q-2
April 2024
-------
FINAL RULE
APRIL 2024
Anderson, R. N. (1999). Methodfor constructing complete annual US life tables (Vol. 129):
National Ctr for Health Statistics.
Apelberg, B. J., Witter, F. R., Herbstman, J. B., Calafat, A. M., Halden, R. U., Needham, L. L.,
& Goldman, L. R. (2007). Cord serum concentrations of perfluorooctane sulfonate
(PFOS) and perfluorooctanoate (PFOA) in relation to weight and size at birth. Children's
Health, 775(11), 1670-1676.
Arias, E., Heron, M., & Tejada-Vera, B. (2013). United States life tables eliminating certain
causes of death, 1999-2001.
Arias, E., & Xu, J. (2019). United States Life Tables, 2017.
Arizona Department of Environmental Quality. (2021). Public Water System PFAS Data.
Retrieved from https://static.azdeq.gov/wqd/pfas/libertv results.pdf
Arizona Department of Environmental Quality. (2023). PFAS Interactive Map. Retrieved from
https://experience.arcgis.com/experience/9a4b9734d7134b5e8e4820a996eb3191
Ashley-Martin, J., Dodds, L., Arbuckle, T. E., Bouchard, M. F., Fisher, M., Morriset, A. S., . . .
Piatt, R. W. (2017). Maternal concentrations of perfluoroalkyl substances and fetal
markers of metabolic function and birth weight: the Maternal-Infant Research on
Environmental Chemicals (MIREC) study. American journal of epidemiology, 185(3),
185-193. doi: 10.1093/aje/kww213
Bach, C. C., Bech, B. H., Nohr, E. A., Olsen, J., Matthiesen, N. B., Bonefeld-Jorgensen E.C., . . .
Henriksen, T. B. (2016). Perfluoroalkyl acids in maternal serum and indices of fetal
growth: the Aarhus Birth Cohort. Environmental health perspectives, 124(6), 848-854.
doi:10.1289/ehp,1510046.
Backes, C. H., Markham, K., Moorehead, P., Cordero, L., Nankervis, C. A., & Giannone, P. J.
(2011). Maternal preeclampsia and neonatal outcomes. Journal of pregnancy.
Bang, H., & Tsiatis, A. A. (2000). Estimating medical costs with censored data. Biometrika,
87(2), 329-343.
Bartell, S. M. (2003). Statistical Methods for Nonsteady State Exposure Inference using
Biomarkers. University of California-Davis, Davis, CA.
Bartell, S. M. (2017). Online serum PFOA calculator for adults. Environmental health
perspectives, 725(10), 104502. doi:10.1289/EHP2820.
Bartell, S. M., Calafat, A. M., Lyu, C., Kato, K., Ryan, P. B., & Steenland, K. (2010). Rate of
decline in serum PFOA concentrations after granular activated carbon filtration at two
public water systems in Ohio and West Virginia. Environ Health Perspect, 118(2), 222-
228. doi: 10.1289/ehp.0901252
Beatty, A. L., Ku, I. A., Bibbins-Domingo, K., Christenson, R. H., DeFillippi, C. R., Ganz, P., . .
. Whooley, M. A. (2015). Traditional risk factors versus biomarkers for prediction of
secondary events in patients with stable Coronary Heart Disease: from the Heart and Soul
Study. Journal of the American Heart Association, 4. doi: 10.1161/JAHA. 114.001646
Begg, C. B., & Mazumdar, M. (1994). Operating characteristics of a rank correlation test for
publication bias. Biometrics, 50(4), 1088-1101. doi:https://doi.org/10.2307/2533446
Bell, E. M., Yeung, E. H., Ma, W., Kannan, K., Sundaram, R., Smarr, M. M., & Louis, G. M. B.
(2018). Concentrations of endocrine disrupting chemicals in newborn blood spots and
infant outcomes in the upstate KIDS study. Environment international, 121, 232-239.
Bhattachaijee, S., Gharaibeh, M., Kamal, M. U., & Riaz, I. B. (2017). Influence of Health and
Functional Status and Co-occurring Chronic Conditions on Healthcare Expenditures
Among Community-dwelling Adults With Kidney Cancer in the United States: A
Final PFAS Rule Economic Analysis
Q-3
April 2024
-------
FINAL RULE
APRIL 2024
Propensity-score-matched Analysis. Clinical Genitourinary Cancer, 75(3), e357-e368.
doi:https://doi.org/10.1016/i.clgc.2016.12.017
Bosworth, R., Cameron, T. A., & DeShazo, J. (2009). Demand for environmental policies to
improve health: Evaluating community-level policy scenarios. Journal of Environmental
Economics and Management, 57(3), 293-308.
Butenhoff, J. L., Chang, S. C., Olsen, G. W., & Thomford, P. J. (2012). Chronic dietary toxicity
and carcinogenicity study with potassium perfluorooctanesulfonate in Sprague Dawley
rats. Toxicology, 293, 1-15. doi:10.1016/j.tox.2012.01.003
Cadwallader, A., A. Greene, H. Holsinger, A. Lan, M. Messner, M. Simic, & Albert, R. (2022).
A Bayesian hierarchical model for estimating national PFAS drinking water occurrence.
A WW A Water Science, el284.
California Division of Drinking Water. (2023). EDT Library and Water Quality Analyses Data
and Download Page. Retrieved from
https://www.waterboards.ca.gov/drinking water/certlic/drinkingwater/EDTlibrary.html.
Callaghan, W. M., MacDorman, M. F., Shapiro-Mendoza, C. K., & Barfield, W. D. (2017).
Explaining the recent decrease in US infant mortality rate, 2007-2013. American journal
of obstetrics and gynecology, 216(1), 73. e71-73. e78.
Callan, A. C., Rotander, A., Thompson, K., Heyworth, J., Meueller, J. F., Odland, J. O., &
Hinwood, A. L. (2016). Maternal exposure to perfluoroalkyl acids measured in whole
blood and birth outcomes in offspring. Science of the total environment, 569-570, 1107-
1113. doi:10.1016/j.scitotenv.2016.06.177
Cao, W., Liu, X., Liu, X., Zhou, Y., Zhang, X., Tian, H., . . . Bhatti, P. (2018). Perfluoroalkyl
substances in umbilical cord serum and gestational and postnatal growth in a Chinese
birth cohort. Environment international, 116, 197-205.
Centers for Disease Control and Prevention. (2013a). 2011-2012 Data Documentation,
Codebook, and Frequencies. Cholesterol - HDL (HDL G). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2011-2012/HDL G.XPT
Centers for Disease Control and Prevention. (2013b). 2011-2012 Data Documentation,
Codebook, and Frequencies. Cholesterol - Total (TCHOL G). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2011-2012/TDIQ G.XPT
Centers for Disease Control and Prevention. (2013c). 2011-2012 Data Documentation,
Codebook, and Frequencies. Diabetes (DIO G). National Health and Nutrition
Examination Survey.
Centers for Disease Control and Prevention. (2013d). 2011-2012 Data Documentation,
Codebook, and Frequencies. Smoking - Cigarette Use (SMO G). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2011-2012/SMQ G.XPT
Centers for Disease Control and Prevention. (2015a). 2011-2012 Data Documentation,
Codebook, and Frequencies. Demographic Variables & Sample Weights (DEMO G).
National Health and Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2011-2012/DEMQ G.XPT
Centers for Disease Control and Prevention. (2015b). 2013-2014 Data Documentation,
Codebook, and Frequencies. Demographic Variables & Sample Weights (DEMO H).
Final PFAS Rule Economic Analysis
Q-4
April 2024
-------
FINAL RULE
APRIL 2024
National Health and Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2013 -2014/DEMO H.XPT
Centers for Disease Control and Prevention. (2015c). 2013-2014 Data Documentation,
Codebook, and Frequencies. Diabetes (DIO H). National Health and Nutrition
Examination Survey. Retrieved from: https://wwwn.cdc.gov/Nchs/Nhanes/2011-
2012/DIQ H.XPT
Centers for Disease Control and Prevention. (2015d). 2013-2014 Data Documentation,
Codebook, and Frequencies. Smoking - Cigarette Use (SMOH). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2011-2012/SMQ H.XPT
Centers for Disease Control and Prevention. (2016a). 2013-2014 Data Documentation,
Codebook, and Frequencies. Cholesterol - HDL (HDL H). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2013-2014/HDL H.XPT
Centers for Disease Control and Prevention. (2016b). 2013-2014 Data Documentation,
Codebook, and Frequencies. Cholesterol - Total (TCHOL H). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2013-2014/TCHOL H.XPT
Centers for Disease Control and Prevention. (2016c, February 11). STDs During Pregnancy -
CDC Detailed Fact Sheet. Retrieved from https://www.cdc.gov/std/pregnancv/stdfact-
pregnancv-detailed.htm#r5
Centers for Disease Control and Prevention. (2017a). 2015-2016 Data Documentation,
Codebook, and Frequencies. Cholesterol - HDL (HDL I). National Health and Nutrition
Examination Survey. Retrieved from: https://wwwn.cdc.gov/Nchs/Nhanes/2015-
2016/HDL I.XPT
Centers for Disease Control and Prevention. (2017b). 2015-2016 Data Documentation,
Codebook, and Frequencies. Cholesterol - Total (TCHOL I). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2015-2016/TCHQL I.XPT
Centers for Disease Control and Prevention. (2017c). 2015-2016 Data Documentation,
Codebook, and Frequencies. Demographic Variables & Sample Weights (DEMO I).
National Health and Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2015-2016/DEMQ I.XPT
Centers for Disease Control and Prevention. (2017d). 2015-2016 Data Documentation,
Codebook, and Frequencies. Diabetes (DIO I). National Health and Nutrition
Examination Survey. Retrieved from: https://wwwn.cdc.gov/Nchs/Nhanes/2015-
2016/DIQ I.XPT
Centers for Disease Control and Prevention. (2017e). 2015-2016 Data Documentation,
Codebook, and Frequencies. Smoking - Cigarette Use (SMO I). National Health and
Nutrition Examination Survey. Retrieved from:
https://wwwn.cdc.gov/Nchs/Nhanes/2015-2016/SMQ I.XPT
Centers for Disease Control and Prevention. (2017f). User Guide to the 2017 Period 2016
Cohort Linked Birth Infant Death Public Use File.
Centers for Disease Control and Prevention. (2018). User Guide to the 2018 Period 2017 Cohort
Linked Birth Infant Death Public Use File.
Final PFAS Rule Economic Analysis
Q-5
April 2024
-------
FINAL RULE
APRIL 2024
Centers for Disease Control and Prevention. (2020a). Commit to Healthy Choices to Prevent
Birth Defects. Retrieved from https://www.cdc.gov/ncbddd/birthdefects/prevention.html
Centers for Disease Control and Prevention. (2020b). Pregnancy Complications.
Centers for Disease Control and Prevention. (2020c). Underlying Cause of Death, 1999-2020 on
CDC WONDER Online Database, released in 2020. Data are from the Multiple Cause of
Death Files, 1999-2019, as compiledfrom data provided by the 57 vital statistics
jurisdictions through the Vital Statistics Cooperative Program. Retrieved from:
http://wonder.cdc.gov/ucd-icdlO.html
Centers for Disease Control and Prevention. (2022). National Report on Human Exposure to
Environmental Chemicals. Updated March 2022. Retrieved from
https://www.cdc.gov/exposurereport/
Chang, H. Y., Nuyten, D. S. A., Sneddon, J. B., Hastie, T., Tibshirani, R., Sorlie, T., . . . van de
Vijver, M. J. (2005). Robustness, scalability, and integration of a wound-response gene
expression signature in predicting breast cancer survival. Proc. Natl. Acad. Sci. U. S. A.,
702(10), 3738-3743. doi:10.1073/pnas.0409462102
Chateau-Degat, M.-L., Pereg, D., Dallaire, R., Ayotte, P., Dery, S., & Dewailly, E. (2010).
Effects of perfluorooctanesulfonate exposure on plasma lipid levels in the Inuit
population of Nunavik (Northern Quebec). Environmental research, 110(7), 710-717.
Chen, M. H., Ha, E. H., Wen, T. W., Su, Y. N, Lien, G. W., Chen, C. Y., . . . Hsieh, W. S.
(2012). Perfluorinated compounds in umbilical cord blood and adverse birth outcomes.
PLoS One, 7(8), e42474.
Chen, M. H., Ng, S., Hsieh, C. J., Lin, C. C., Hsieh, W. S., & Chen, P. C. (2017). The impact of
prenatal perfluoroalkyl substances exposure on neonatal and child growth. Science of the
total environment, 607-608, 669-675. doi:10.1016/j.scitotenv.2017.06.273
Chu, C., Zhou, Y., Li, Q.-Q., Bloom, M. S., Lin, S., Yu, Y.-J., . . . Yang, B.-Y. (2020). Are
perfluorooctane sulfonate alternatives safer? New insights from a birth cohort study.
Environment international, 135, 105365.
Cochran, W. G. (1954). The combination of estimates from different experiments. Biometrics,
70(1), 101-129. doi:https://doi.org/10.2307/3001666
Cole, T. J., Hey, E., & Richmond, S. (2010). The PREM score: A graphical tool for predicting
survival in very preterm births. Archives of Disease in Childhood-Fetal and Neonatal
Edition, 95(1), F14-F19.
Collins Jr, J. W., & David, R. J. (2009). Racial disparity in low birth weight and infant mortality.
Clinics in perinatology, 5(5(1), 63-73.
Colorado Department of Public Health and Environment. (2020). 2020 PFAS Sampling Project.
Convertino, M., Church, T. R., Olsen, G. W., Liu, Y., Doyle, E., Elcombe, C. R., . . . Evans, T.
R. J. (2018). Stochastic Pharmacokinetic-Pharmacodynamic Modeling for Assessing the
Systemic Health Risk of Perfluorooctanoate (PFOA). Toxicological Sciences, 163(1),
293 -3 06. doi: 10.1093/toxsci/kfy03 5
Cooke, R. W. I. (2010). Does Neonatal and Infant Neurodevelopmental Morbidity of Multiples
and Singletons Differ? Paper presented at the Seminars in Fetal and Neonatal Medicine.
Crainiceanu, C. M., Stedinger, J. R., Ruppert, D., & Behr, C. T. (2003). Modeling the U.S.
national distribution of waterborne pathogen concentrations with application to
Cryptosporidium parvum. Water Resources Research, 39(9).
doi :https://doi.org/10.1029/2002WR001664
Final PFAS Rule Economic Analysis
Q-6
April 2024
-------
FINAL RULE
APRIL 2024
D'Agostino, R. B., Grundy, S., Sullivan, L. M., Wilson, P., & Group, C. H. D. R. P. (2001).
Validation of the Framingham coronary heart disease prediction scores: results of a
multiple ethnic groups investigation. Jama, 286(2), 180-187. doi:10.1001/jama.286.2.180
D'Agostino, R. B., Russell, M. W., Huse, D. M., Ellison, R. C., Silbershatz, H., Wilson, P. W., &
Hartz, S. C. (2000). Primary and subsequent coronary risk appraisal: new results from the
Framingham study. Am. Heart J., 139(2 Pt 1), 272-281. doi: 10.1067/mhj.2000.96469
D'Agostino, R. B., Vasan, R. S., Pencina, M. J., Wolf, P. A., Cobain, M., Massaro, J. M., &
Kannel, W. B. (2008). General cardiovascular risk profile for use in primary care.
Circulation, 117(6), 743-753.
D'Agostino, R. B., Wolf, P. A., Belanger, A. J., & Kannel, W. B. (1994). Stroke risk profile:
adjustment for antihypertensive medication. The Framingham Study. Stroke, 25(1), 40-
43. doi: 10.1161/01 str.25.1.40
Darrow, L. A., Stein, C. R., & Steenland, K. (2013). Serum Perfluorooctanoic Acid and
Perfluorooctane Sulfonate Concentrations in Relation to Birth Outcomes in the Mid-Ohio
Valley, 2005-2010.Environ. Health Per sped, 727(10), 1207-1213.
doi: 10.1289/ehp. 1206372
Deeks, J. J. (2002). Issues in the selection of a summary statistic for meta-analysis of clinical
trials with binary outcomes. Statistics in medicine, 27(11), 1575-1600.
Delaware Office of Drinking Water. (2021). Delaware Drinking Water Watch Data. Retrieved
from: https://drinkingwater.dhss.delaware.gov/
DerSimonian, R., & Laird, N. (1986). Meta-analysis in clinical trials. Controlled clinical trials,
7(3), 177-188.
Donahue, S. M., Kleinman, K. P., Gillman, M. W., & Oken, E. (2010). Trends in birth weight
and gestational length among singleton term births in the United States: 1990-2005.
Obstetrics and gynecology, 115(2 Pt 1), 357.
Dong, Z., Wang, H., Yu, Y. Y., Li, Y. B., Naidu, R, & Liu, Y. (2019). Using 2003-2014 US
NHANES data to determine the associations between per-and polyfluoroalkyl substances
and cholesterol: Trend and implications. Ecotoxicology and environmental safety, 173,
461-468.
Duval, S., & Tweedie, R. (2000a). A nonparametric "trim and fill" method of accounting for
publication bias in meta-analysis. J Am Stat Assoc, 95(449), 89-98.
doi:10.1080/01621459.2000.10473905
Duval, S., & Tweedie, R. (2000b). Trim and fill: A simple funnel-plot-based method of testing
and adjusting for publication bias in meta-analysis. Biometrics, 56(2), 455-463.
doi: 10.1111/j .0006-341X.2000.00455.X
Dzierlenga, M., Allen, B., Clewell, H., & Longnecker, M. (2020). Pharmacokinetic bias analysis
of an association between clinical thyroid disease and two perfluoroalkyl substances.
Environ Int, 141: 105784.
Dzierlenga, M., Crawford, L., & Longnecker, M. (2020). Birth weight and perfluorooctane
sulfonic acid: a random-effects meta-regression analysis. Environmental Epidemiology,
4(3).
Egger, M., Smith, G., Schneider, M., & Minder, C. (Eds.). (2008). Systematic reviews in health
care: Meta-analysis in context. London, UK: BMJ Publishing Group.
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by
a simple, graphical test. BMJ, 575(7109), 629-634.
doi :https://doi.org/10.1136/bmi. 315.7109.629
Final PFAS Rule Economic Analysis
Q-7
April 2024
-------
FINAL RULE
APRIL 2024
Ely, D. M., & Driscoll, A. K. (2020). Infant Mortality in the United States, 2018: Data From the
Period Linked Birth Infant Death File. Retrieved from
https://www.cdc.gov/nchs/data/nvsr/nvsr69/NVSR-69-7-508.pdf
Engels, E. A., Schmid, C. H., Terrin, N., Olkin, I., & Lau, J. (2000). Heterogeneity and statistical
significance in meta-analysis: an empirical study of 125 meta-analyses. Statistics in
medicine, 79(13), 1707-1728.
Eriksen, K. T., Raaschou-Nielsen, O., McLaughlin, J. K., Lipworth, L., Tjonneland, A., Overvad,
K., & Sorensen, M. (2013). Association between plasma PFOA and PFOS levels and
total cholesterol in a middle-aged Danish population. PLoS One, 5(2), e56969.
Fan, Y., Li, X., Xu, Q., Zhang, Y., Yang, X., Han, X., . . . Lu, C. (2020). Serum albumin
mediates the effect of multiple per-and polyfluoroalkyl substances on serum lipid levels.
Environmental pollution, 266, 115138.
Fei, C., McLaughlin, J. K., Tarone, R. E., & Olsen, J. (2007). Perfluorinated Chemicals and Fetal
Growth: A Study within the DanishNational Birth Cohort. Environmental health
perspectives, 775(11), 1677-1682.
Fisher, M., Arbuckle, T. E., Wade, M., & Haines, D. A. (2013). Do perfluoroalkyl substances
affect metabolic function and plasma lipids?—Analysis of the 2007-2009, Canadian
Health Measures Survey (CHMS) Cycle 1. Environmental research, 121, 95-103.
Fitz-Simon, N., Fletcher, T., Luster, M. I., Steenland, K., Calafat, A. M., Kato, K., & Armstrong,
B. (2013). Reductions in serum lipids with a 4-year decline in serum perfluorooctanoic
acid and perfluorooctanesulfonic acid. Epidemiology (Cambridge, Mass.), 24(4), 569.
Fragki, S., Dirven, H., Fletcher, T., Grasl-Kraupp, B., Bjerve Giitzkow, K., Hoogenboom, R., . . .
Luijten, M. (2021). Systemic PFOS and PFOA exposure and disturbed lipid homeostasis
in humans: what do we know and what not? CritRev Toxicol', 141-164.
Fried, L. P., Borhani, N. O., Enright, P., Furberg, C. D., Gardin, J. M., Kronmal, R. A., . . .
Newman, A. (1991). The cardiovascular health study: design and rationale. Annals of
epidemiology, 7(3), 263-276.
Friedman, G. D., Cutter, G. R., Donahue, R. P., Hughes, G. H., Hulley, S. B., Jacobs Jr, D. R., . .
. Savage, P. J. (1988). CARDIA: study design, recruitment, and some characteristics of
the examined subjects. Journal of clinical epidemiology, 41(11), 1105-1116.
Fromme, H., Mosch, C., Morovitz, M., Alba-Alejandre, I., Boehmer, S., Kiranoglu, M., . . .
Koletzko, B. (2010). Pre-and postnatal exposure to perfluorinated compounds (PFCs).
Environmental Science & Technology, 44(1%), 7123-7129.
Fu, Y., Wang, T., Fu, Q., Wang, P., & Lu, Y. (2014). Associations between serum concentrations
of perfluoroalkyl acids and serum lipid levels in a Chinese population. Ecotoxicology and
environmental safety, 106, 246-252.
Gabry, J., & Mahr, T. (2020). bayesplot: Plotting for Bayesian Models. Retrieved from
https://mc-stan.org/bavesplot/
Gabry, J., Veen, D., Stan Development Team, Andreae, M., Betancourt, M., Carpenter, B., . . .
Trangucci, R. (2018). shinystan: Interactive Visual and Numerical Diagnostics and
Posterior Analysis for Bayesian Models. Retrieved from: https://cran.r-
proiect.org/web/packages/shinystan/index.html
Gao, K., Zhuang, T., Liu, X., Fu, J., Zhang, J., Fu, J., . . . Song, M. (2019). Prenatal exposure to
per-and polyfluoroalkyl substances (PFASs) and association between the placental
transfer efficiencies and dissociation constant of serum proteins-PFAS complexes.
Environmental Science & Technology, 55(11), 6529-6538.
Final PFAS Rule Economic Analysis
Q-8
April 2024
-------
FINAL RULE
APRIL 2024
Georgia Environmnetal Protection Division. (2020). EPA Sampling Report for Summerville.
Retrieved from https://epd.georgia.gov/pfoa-and-pfos-information
Goff, D. C., Lloyd-Jones, D. M., Bennett, G., Coady, S., D'Agostino, R. B., Gibbons, R., . . .
Wilson, P. W. F. (2014). 2013 ACC/AHA guideline on the assessment of cardiovascular
risk: A report of the American College of Cardiology/American Heart Association task
force on practice guidelines. Circulation, 129(25 suppl 2), 49-73.
Govarts, E., Remy, S., Bruckners, L., Hond, E. D., Sioen, I., Nelen, V., . . . Schoeters, G. (2016).
Combined Effects of Prenatal Exposures to Environmental Chemicals on Birth Weight.
International journal of environmental research and public health, 75(495), 1-19.
doi:10.3390/ijerphl3050495
Hamm, M. P., Cherry, N. M., Chan, E., Martin, J. W., & Burstyn, I. (2010). Maternal exposure
to perfluorinated acids and fetal growth. Journal of Exposure Science and Environmental
Epidemiology, 20, 589-597. doi: 1559-0631/10
He, X., Liu, Y., Xu, B., Gu, L., & Tang, W. (2018). PFOA is associated with diabetes and
metabolic alteration in US men: National Health and Nutrition Examination Survey
2003-2012. Science of the total environment, 625, 566-574.
Higgins, J., Chandler, J., Cumpston, M., Page, M., & Welch, V. E. (2021). Cochrane Handbook
for Systematic Reviews of Interventions version 6.2 (updated February 2021). Cochrane.
Higgins, J., Thomas, J., Chandler, J., Cumpston, M., Li, T., Page, M. J., & Welch, V. A. (2019).
Cochrane handbook for systematic reviews of interventions: John Wiley & Sons.
Higgins, J., Thompson, S., Deeks, J., & Altman, D. (2003). Measuring inconsistency in meta-
analyses. BMJ, 527(7414), 557-560. doi:https://doi.org/10.1136/bmi.327.7414.557
Hjermitslev, M. H., Long, M., Wielsoe, M., & Bonefeld-Jorgensen, E. C. (2020). Persistent
organic pollutants in Greenlandic pregnant women and indices of foetal growth: The
ACCEPT study. Science of the total environment, 698, 134118.
Hoffman, K., Webster, T. F., Bartell, S. M., Weisskopf, M. G., Fletcher, T., & Vieira, V. M.
(2011). Private drinking water wells as a source of exposure to perfluorooctanoic acid
(PFOA) in communities surrounding a fluoropolymer production facility. Environmental
health perspectives, 779(1), 92-97.
Hollenbeak, C. S., Nikkei, L. E., Schaefer, E. W., Alemao, E., Ghahramani, N, & Raman, J. D.
(2011). Determinants of Medicare All-Cause Costs Among Elderly Patients with Renal
Cell Carcinoma. Journal of Managed Care Pharmacy, 77(8), 610-620.
doi:10.18553/jmcp.2011.17.8.610
Horon, I., Munson, K.M. (2020). Discussion of the Merits of Using Singleton Data from the
CDC to Develop a Relationship Between Birthweight and Infant Mortality. Email
Correspondence 11/10/2020.
Huang, M., Jiao, J., Zhuang, P., Chen, X., Wang, J., & Zhang, Y. (2018). Serum polyfluoroalkyl
chemicals are associated with risk of cardiovascular diseases in national US population.
Environment international, 119, 37-46.
ICF. (2020). Memorandum re: Trends in Infant Mortality by Birth Weight Categories, EPA
Contract No. 68HE0C18D0001, TO 68HERC20F0107. Submitted to U.S. Environmental
Protection Agency Office of Groundwater and Drinking Water July 30, 2020. Retrieved
from https://www.regulations.gov/docket/EPA-HQ-OW-2022-Q114
ICF. (2021). Memorandum re: Review ofPFAS Lipids Study Quality. Submitted to U.S.
Environmental Protection Agency Office of Groundwater and Drinking Water March 23,
2021. Retrieved from https://www.regulations.gov/docket/EPA-HQ-OW-2022-0114
Final PFAS Rule Economic Analysis
Q-9
April 2024
-------
FINAL RULE
APRIL 2024
ICF. (2022). Memorandum re: Literature Review of Population Attributable Fraction Estimates
and Next Steps for Renal Cell Carcinoma PFAS Benefits Modeling, EPA Contract No.
68HE0C18D0001, TO 68HERC22F0262. Submitted to U.S. Environmental Protection
Agency Office of Groundwater and Drinking Water July 28, 2022. Retrieved from
https ://www. regulations. gov/docket/EP A-HQ-0 W-2022-0114
Idaho Department of Environmental Quality. (2023). Idaho Drinking Water Watch. Retrieved
from http ://dww. deq. i daho. gov/IDPDWW/
Indiana Department of Environmental Management. (2023). PFAS Sampling Project for
Community Public Water Systems. Retrieved from
https://www.in.gov/idem/resources/nonrule-policies/per-and-polvfluoroalkyl-substances-
pfas/#sampling
Institute of Medicine. (1985). Preventing Low Birthweight. Washington, DC: The National
Academies Press.
Jacob, J. A. (2016). U.S. infant mortality rate declines but still exceeds other developed
countries. Journal of the American Medical Association, 315(5), 451-452.
Jain, R. B., & Ducatman, A. (2019). Roles of gender and obesity in defining correlations
between perfluoroalkyl substances and lipid/lipoproteins. Science of the total
environment, 653, 74-81.
Johnson, P. I., Sutton, P., Atchley, D. S., Koustas, E., Lam, J., Sen, S., . . . Woodruff, T. J.
(2014). The Navigation Guide—Evidence-Based Medicine Meets Environmental Health:
Systematic Review of Human Evidence for PFOA Effects on Fetal Growth.
Environmental health perspectives, 722(10), 1028-1039. doi: 10.1289/ehp. 1307893
Joyce, C., Goodman-Bryan, M., & Hardin, A. (2012). Preterm Birth and Low Birth Weight.
Retrieved from The Urban Child Institute:
http ://www.urbanchildinstitute. org/ sites/all/files/2010-10-01 -PTB-and-LBW.pdf
Kannel, W. B., D'Agostino, R. B., Silbershatz, H., Belanger, A. J., Wilson, P. W., & Levy, D.
(1999). Profile for estimating risk of heart failure. Arch. Intern. Med., 759(11), 1197-
1204. doi:10.1001/archinte,159.11.1197
Kashino, I., Sasaki, S., Okada, E., Matsuura, H., Goudarzi, H., Miyashita, C., . . . Kishi, R.
(2020). Prenatal exposure to 11 perfluoroalkyl substances and fetal growth: A large-scale,
prospective birth cohort study. Environment international, 136, 105355.
Kassambara, A. (2020). ggpubr: 'ggplot2' Based Publication Ready Plots. Retrieved from:
https://cran.r-proiect.org/web/packages/ggpubr/index.html
Kentucky Department for Environmental Protection. (2019). Evaluation of Kentucky Community
Drinking Water for Per- & Poly-Fluoroalkyl Substances. Retrieved from
https://eec.kv.gov/Documents%20for%20URLs/PFAS%20Drinking%20Water%20Repor
t%20Final.pdf
Kiely, J. L., Brett, K. M., Yu, S., & Rowley, D. L. (1994). Low birth weight and intrauterine
growth retardation. Birth Outcomes, 185-202.
Kim, S., Choi, K, Ji, K, Seo, J., Kho, Y., Park, J., . . . Giesy, J. P. (2011). Trans-Placental
Transfer of Thirteen Perfluorinated Compounds and Relations with Fetal Thyroid
Hormones. Environmental Science & Technology, 45(11), 7465-7472.
doi:10.1021/es202408a
Kochanek, K. D., Murphy, S. L., Xu, J., & Arias, E. (2019). Deaths: Final data for 2017.
National vital statistics reports : from the Centers for Disease Control and Prevention,
National Center for Health Statistics, National Vital Statistics System, 68(9), 1-77.
Final PFAS Rule Economic Analysis
Q-10
April 2024
-------
FINAL RULE
APRIL 2024
Kwon, E. J., Shin, J. S., Kim, B. M., Shah-Kulkarni, S., Park, H., Kho, Y. L., . . . Ha, E. H.
(2016). Prenatal Exposure to Perfluorinated Compounds Affects Birth Weight Through
GSTM1 Polymorphism. JOEM, 55(6), el98-e205.
Lau, C., Ambalavanan, N., Chakraborty, H., Wingate, M. S., & Carlo, W. A. (2013). Extremely
low birth weight and infant mortality rates in the United States. Pediatrics, 131(5), 855-
860.
Lauritzen, H. B., Larose, T. L., 0ien, T., Sandanger, T. M., Odland, J. O., van der Bor, M., &
Jacobsen, G. W. (2017). Maternal serum levels of perfluoroalkyl substances and
organochlorines and indices of fetal growth: a Scandinavian case-cohort study. Pediatric
research, 57(1), 33-42. doi:10.1038/pr.2016.187
Lee, E. S., Han, S., & Oh, J. E. (2016). Association between perfluorinated compound
concentrations in cord serum and birth weight using multiple regression model.
Reproductive Toxicology, 59, 53-59. doi:10.1016/j.reprotox.2015.10.020
Lenters, V., Portengen, L., Rignell-Hydbom, A., Jonsson, B. A. G., Lindh, C. H., Piersma, A. H.,
. . . Vermeulen, R. (2016). Prenatal Phthalate, Perfluoroalkyl Acid, and Organochlorine
Exposures and Term Birth Weight in Three Birth Cohorts: Multi-Pollutant Models Based
on Elastic Net Regression. Environmental health perspectives, 124, 365-372.
doi: 10.1289/ehp. 1408933
Li, M., Zeng, Z. W., Qian, Z. M., Vaughn, M. G., Suauve, S., Paul, G., . . . Dong, G. H. (2017).
Isomers of perfluorooctanesulfonate (PFOS) in cord serum and birth outcomes in China:
Guangzhou Birth Cohort Study. Environment international, 102, 1-8.
doi:10.1016/j.envint.2017.03.006
Li, S., Peng, Y., Wang, X., Qian, Y., Xiang, P., Wade, S. W., . . . Handelsman, Y. (2019).
Cardiovascular events and death after myocardial infarction or ischemic stroke in an
older Medicare population. Clinical cardiology, 42(3), 391-399.
Li, Y., Barregard, L., Xu, Y., Scott, K., Pineda, D., Lindh, C. H., . . . Fletcher, T. (2020).
Associations between perfluoroalkyl substances and serum lipids in a Swedish adult
population with contaminated drinking water. Environmental Health, 79(33).
Li, Y., Fletcher, T., Mucs, D., Scott, K., Lindh, C. H., Tallving, P., & Jakobsson, K. (2018).
Half-lives of PFOS, PFHxS and PFOA after end of exposure to contaminated drinking
water. Occupational and environmental medicine, 75(1), 46-51.
Liao, S., Yao, W., Cheang, I., Tang, X., Yin, T., Lu, X., . . . Li, X. (2020). Association between
perfluoroalkyl acids and the prevalence of hypertension among US adults. Ecotoxicology
and environmental safety, 196, 110589.
Lin, C. Y., Lee, H. L., Hwang, Y. T., & Su, T. C. (2020). The association between total serum
isomers of per-and polyfluoroalkyl substances, lipid profiles, and the DNA
oxidative/nitrative stress biomarkers in middle-aged Taiwanese adults. Environmental
research, 182, 109064.
Lin, L., & Chu, H. (2018). Quantifying publication bias in meta-analysis. Biometrics, 74(3), 785-
794.
Lin, P.-I. D., Cardenas, A., Hauser, R., Gold, D. R., Kleinman, K. P., Hivert, M.-F., . . . Horton,
E. S. (2019). Per-and polyfluoroalkyl substances and blood lipid levels in pre-diabetic
adults—longitudinal analysis of the diabetes prevention program outcomes study.
Environment international, 129, 343-353.
Final PFAS Rule Economic Analysis
Q-ll
April 2024
-------
FINAL RULE
APRIL 2024
Lind, D. V., Priskorn, L., Lassen, T. H., Nielsen, F., Kyhl, H. B., Kristensen, D. M., . . . Jensen,
T. K. (2017). Prenatal exposure to perfluoroalkyl substances and anogenital distance at 3
months of age in a Danish mother-child cohort. Reproductive Toxicology, 68, 200-206.
Liu, H.-S., Wen, L.-L., Chu, P.-L., & Lin, C.-Y. (2018). Association among total serum isomers
of perfluorinated chemicals, glucose homeostasis, lipid profiles, serum protein and
metabolic syndrome in adults: NHANES, 2013-2014. Environmental pollution, 232, 73-
79.
Lockwood, J. R., Schervish, M. J., Gurian, P., & Small, M. J. (2001). Characterization of Arsenic
Occurrence in Source Waters of U.S. Community Water Systems. Journal of the
American Statistical Association, 96(456), 1184-1193.
doi:10.1198/016214501753381832
Louis, G. M. B., Zhai, S., Smarr, M. M., Grewal, J., Zhang, C., Grantz, K. L., . . . Honda, M.
(2018). Endocrine disruptors and neonatal anthropometry, NICHD fetal growth studies-
singletons. Environment international, 119, 515-526.
Lu, S., & Bartell, S. M. (2020). Serum PFAS Calculator for Adults. Retrieved from
www.ics.uci.edu/~sbartell/pfascalc.html
Ma, S., & Finch, B. K. (2010). Birth Outcome Measures and Infant Mortality. Population
Research and Policy Review, 29(6), 865-891.
Mahmood, S. S., Levy, D., Vasan, R. S., & Wang, T. J. (2014). The Framingham Heart Study
and the epidemiology of cardiovascular disease: a historical perspective. The Lancet,
555(9921), 999-1008.
Maine Department of Environmental Protection. (2020). MAINE PFAS DATA (2007 - 2020).
Retrieved from https://wwwl.maine.gov/dep/spills/topics/pfas/PFAS-current-results-
10262020.pdf
Maine Department of Health and Human Services. (2023). Maine PFAS Data (2007 - 2020).
Retrieved from https://www.maine.gov/dhhs/mecdc/environmental-
health/dwp/cet/documents/PFASallResults.pdf
Maisonet, M., Terrell, M. L., McGeehin, M. A., Christensen, K. Y., Holmes, A., Calafat, A. M.,
& Marcus, M. (2012). Maternal concentrations of polyfluoroalkyl compounds during
pregnancy and fetal and postnatal growth in British girls. Environmental health
perspectives, 720(10), 1432-1437.
Manzano-Salgado, C. B., Casas, M., Lopez-Espinosa, M.-J., Ballester, F., Iniguez, C., Martinez,
D., . . . Schettgen, T. (2017). Prenatal exposure to perfluoroalkyl substances and birth
outcomes in a Spanish birth cohort. Environment international, 108, 278-284.
Marks, K. J., Cutler, A. J., Jeddy, Z., Northstone, K., Kato, K., & Hartman, T. J. (2019).
Maternal serum concentrations of perfluoroalkyl substances and birth size in British boys.
International journal of hygiene and environmental health, 222(5), 889-895.
Mathews, T. (2013). Infant mortality statistics from the 2010 period linked birth/infant death data
set. National vital statistics reports : from the Centers for Disease Control and
Prevention, National Center for Health Statistics, National Vital Statistics System, 62(8),
1-26.
McCormick, M. C. (1985). The contribution of low birth weight to infant mortality and
childhood morbidity. New England Journal of Medicine, 312(2), 82-90.
Mclntire, D. D., Bloom, S. L., Casey, B. M., & Leveno, K. J. (1999). Birth weight in relation to
morbidity and mortality among newborn infants. New England Journal of Medicine,
340( 16), 1234-1238.
Final PFAS Rule Economic Analysis
Q-12
April 2024
-------
FINAL RULE
APRIL 2024
Meng, Q., Inoue, K., Ritz, B., Olsen, J., & Liew, Z. (2018). Prenatal exposure to perfluoroalkyl
substances and birth outcomes; an updated analysis from the Danish National Birth
Cohort. International journal of environmental research and public health, 15(9), 1832.
Minatoya, M., Itoh, S., Miyashita, C., Araki, A., Sasaki, S., Miura, R., . . . Kishi, R. (2017).
Association of prenatal exposure to perfluoroalkyl substances with cordblood adipokines
and birth size: The Hokkaido Study on environment and children's health. Environmental
research, 156, 175-182. doi:10.1016/j.envres.2017.03.033
Mishra, S., Ram, B., Singh, A., & Yadav, A. (2018). Birth order, stage of infancy and infant
mortality in India. Journal of biosocial science, 50(5), 604-625.
Missouri Department of Natural Resources. (2023). Missouri PFAS Viewer Tool. Retrieved from
https://modnr.maps. arcgis.com/apps/webappviewer/index.html?id=386c71927569476ebd
2d0e6910424dl7
Mitchell, A. P., Dey, P., Ohn, J. A., Tabatabai, S. M., Curry, M. A., & Bach, P. B. (2020). The
Accuracy and Usefulness of the National Comprehensive Cancer Network Evidence
Blocks Affordability Rating. Pharmacoeconomics, 38(7), 737-745. doi:10.1007/s40273-
020-00901-x
Monroy, R., Morrison, K., Teo, K., Atkinson, S., Kubwabo, C., Stewart, B., & Foster, W. G.
(2008). Serum levels of perfluoroalkyl compounds in human maternal and umbilicalcord
blood samples. Environmental research, 108, 56-62. doi:10.1016/j.envres.2008.06.001
Mora, S. (2016). Nonfasting for Routine Lipid Testing: From Evidence to Action. JAMA Intern
Med, 176, 1005-1006.
National Center for Health Statistics, H., MD. (2019). Health, United States, 2018. Retrieved
from https://www.cdc.gov/nchs/data/hus/husl8.pdf
Negri, E., Metruccio, F., Guercio, V., Tosti, L., Benfenati, E., Bonzi, R., . . . Moretto, A. (2017).
Exposure to PFOA and PFOS and fetal growth: a critical merging of toxicological and
epidemiological data. Critical reviews in toxicology, 10.1080 10408444.2016.1271972,
489-515. doi: 10.1080/10408444.2016.1271972
Nelson, J. W., Hatch, E. E., & Webster, T. F. (2010). Exposure to polyfluoroalkyl chemicals and
cholesterol, body weight, and insulin resistance in the general US population.
Environmental health perspectives, 118(2), 197-202.
New Hampshire Department of Environmental Services. (2021). One Stop Data Portal. In.
New York Department of Health. (2022). PFAS Sampling Results for Public Water Systems in
New York. Available in public docket at: https://www.regulations.gov/docket/EPA-HQ-
QW-2022-0114.
Ngui, E. M., Cortright, A. L., & Michalski, K. (2015). Relationship of paternity status, welfare
reform period, and racial/ethnic disparities in infant mortality. American Journal of
Men's Health, 9(5), 350-359. doi:10.1177/1557988314543906
North Carolina Department of Environmental Quality. (2023). 2022 DEO Sampling of 50 Water
Systems. Retrieved from https://www.deq.nc.gov/kev-issues/2022-pfas-data-pfoa-
pfospdf/download?attachment
North Dakota Department of Environmental Quality. (2020). North Dakota Statewide 2020 Per-
andPolyflaoroalkyl Substances (PFAS) Presence Absence Survey. Retrieved from
https://deq.nd.gov/Publications/MF/PFAS Report 2020.pdf?v=l
North Dakota Department of Environmental Quality. (2021). North Dakota Statewide 2021 Per-
andPolyflaoroalkyl Substances (PFAS) Presence Absence Survey. Retrieved from
https://dea.nd.gov/Publications/MF/PFAS Report 2021.pdf?v=l
Final PFAS Rule Economic Analysis
Q-13
April 2024
-------
FINAL RULE
APRIL 2024
O'Sullivan, A. K., Rubin, J., Nyambose, J., Kuznik, A., Cohen, D. J., & Thompson, D. (2011).
Cost estimation of cardiovascular disease events in the US. Pharmacoeconomics, 29(8),
693-704.
Occupational Safety and Health Administration. (2010). Occupational Exposure to Respirable
Crystalline Silica — Review of Health Effects Literature and Preliminary Quantitative
Risk Assessment.
Oregon Health Authority. (2022). PFAS Screening and Assessment Project Results. Retrieved
from https://vourwater.oregon.gov/pfascountv.php.
Osterman, M. J., Kochanek, K. D., MacDorman, M. F., Strobino, D. M., & Guyer, B. (2015).
Annual summary of vital statistics: 2012-2013. Pediatrics, 135(6), 1115-1125.
Ott, W. R. (1995). Environmental Statistics and Data Analysis (1st ed.). New York: CRC Press,
Lewis Publishers.
Palmer, T. M., & Sterne, J. A. C. (Eds.). (2016). Meta-analysis in Statu: An updated collection
fi'om the Statu Journal (2nd ed.). College Station, TX: Stata Press.
PEACE Trial Investigators. (2004). Angiotensin-converting-enzyme inhibition in stable
coronary artery disease. New England Journal of Medicine, 557(20), 2058-2068.
Pencina, M. J., DAgostino, R. B., Sr., Larson, M. G., Massaro, J. M., & Vasan, R. S. (2009).
Predicting the 30-year risk of cardiovascular disease: the framingham heart study.
Circulation, 119(24), 3078-3084. doi:10.1161/CIRClJLATIONAHA.108.816694
Pencina, M. J., DAgostino, R. B., Sr., & Steyerberg, E. W. (2011). Extensions of net
reclassification improvement calculations to measure usefulness of new biomarkers. Stat.
Med., 30(1), 11-21. doi:10.1002/sim.4085
Perrin, A., Sherman, S., Pal, S., Chua, A., Gorritz, M., Liu, Z., . . . Garrison, L. P. (2015).
Lifetime cost of everolimus vs axitinib in patients with advanced renal cell carcinoma
who failed prior sunitinib therapy in the US. Journal of Medical Economics, 75(3), 200-
209.
Powers, D. A., Frisbie, W. P., Hummer, R. A., Pullum, S. G., & Solis, P. (2006). Race/ethnic
differences and age-variation in the effects of birth outcomes on infant mortality in the
US. Demographic Research, 14, 179-216.
Quah, S. R. (2016). International Encyclopedia of Public Health. Waltham, MA: Academic
Press.
R Core Team. (2021). R: A language and environment for statistical computing. Vienna, Austria:
R Foundation for Statistical Computing.
Racsa, P. N., Whisman, T. R., & Worley, K. (2015). Comparing two tyrosine kinase inhibitors
for treatment of advanced renal cell carcinoma in Medicare and commercially insured
patients. Current Medical Research and Opinion, 57(10), 1933-1940.
Rappazzo, K. M., Coffman, E., & Hines, E. P. (2017). Exposure to perfluorinated alkyl
substances and health outcomes in children: a systematic review of the epidemiologic
literature. International journal of environmental research and public health, 14(1), 691.
Ratnasiri, A. W., Parry, S. S., Arief, V. N., DeLacy, I. H., Halliday, L. A., DiLibero, R. J., &
Basford, K. E. (2018). Recent trends, risk factors, and disparities in low birth weight in
California, 2005-2014: A retrospective study. Maternal health, neonatology and
perinatology, 4(1), 1-13.
Raudenbush, S. W. (2009). Analyzing effect sizes: Random-effects models. The handbook of
research synthesis and meta-analysis, 2, 295-316.
Final PFAS Rule Economic Analysis
Q-14
April 2024
-------
FINAL RULE
APRIL 2024
Ray, J. G., Park, A. L., & Fell, D. B. (2017). Mortality in infants affected by preterm birth and
severe small-for-gestational-age birth weight. Pediatrics, 140(6).
Regli, S., Chen, J., Messner, M., Elovitz, M. S., Letkiewicz, F. J., Pegram, R. A., . . . Wright, J.
M. (2015). Estimating potential increased bladder cancer risk due to increased bromide
concentrations in sources of disinfected drinking waters. Environmental Science &
Technology, 49(22), 13094-13102.
Reinhorn, D., Sarfaty, M., Leshno, M., Moore, A., Neiman, V., Rosenbaum, E., & Goldstein, D.
A. (2019). A cost-effectiveness analysis of nivolumab and ipilimumab versus sunitinib in
first-line intermediate-to poor-risk advanced renal cell carcinoma. The oncologist, 24(3),
366-371.
Rice, W. S., Goldfarb, S. S., Brisendine, A. E., Burrows, S., & Wingate, M. S. (2017).
Disparities in infant mortality by race among Hispanic and non-Hispanic infants.
Maternal and child health journal, 21(1), 1581-1588.
Robledo, C. A., Yeung, E., Mendola, P., Sundaram, R., Maisog, J., Sweeney, A. M., . . . Louis,
G. M. B. (2015). Preconception maternal and paternal exposure to persistent organic
pollutants and birth size: the LIFE study. Environmental health perspectives, 123(1), 88-
94.
Rodriguez-Barranco, M., Tobias, A., Redondo, D., Molina-Portillo, E., & Sanchez, M. J. (2017).
Standardizing effect size from linear regression models with log-transformed variables
for meta-analysis. BMC medical research methodology, 17(1), 1-9.
Roger, V. L., Go, A. S., Lloyd-Jones, D. M., Benjamin, E. J., Berry, J. D., Borden, W. B., . . .
Ford, E. S. (2012). Heart disease and stroke statistics—2012 update: a report from the
American Heart Association. Gradation, 125(1), e2.
Rowley, D. L., & Hogan, V. (2012). Disparities in infant mortality and effective, equitable care:
Are infants suffering from benign neglect? Annual Review of Public Health, 33, 75-87.
RStudio Team. (2020). RStudio: Integrated Development Environment for R. Retrieved from
https://www.rstudio.com/
Riicker, G., Schwarzer, G., Carpenter, J., & Olkin, I. (2009). Why add anything to nothing? The
arcsine difference as a measure of treatment effect in meta-analysis with zero cells.
Statistics in medicine, 28(5), 721-738.
Russell, R. B., Petrini, J. R., Damus, K., Mattison, D. R., & Schwarz, R. H. (2003). The changing
epidemiology of multiple births in the United States. Obstetrics & Gynecology, 101(1),
129-135.
Sagiv, S. K., Rifas-Shiman, S. L., Fleisch, A. F., Webster, T. F., Calafat, A. M., Ye, X., . . .
Oken, E. (2018). Early-pregnancy plasma concentrations of perfluoroalkyl substances
and birth outcomes in Project Viva: confounded by pregnancy hemodynamics? American
journal of epidemiology, 757(4), 793-802.
Savitz, D. A., Stein, C. R., Elston, B., Wellenius, G. A., Bartell, S. M., Shin, H.-M., . . . Fletcher,
T. (2012). Relationship of perfluorooctanoic acid exposure to pregnancy outcome based
on birth records in the mid-Ohio Valley. Environ. Health Perspect., 720(8), 1201-1207.
doi: 10.1289/ehp. 1104752
Seo, S.-H., Son, M.-H., Choi, S.-D., Lee, D.-H., & Chang, Y.-S. (2018). Influence of exposure to
perfluoroalkyl substances (PFASs) on the Korean general population: 10-year trend and
health effects. Environment international, 113, 149-161.
Shearer, J. J., Callahan, C. L., Calafat, A. M., Huang, W. Y., Jones, R. R., Sabbisetti, V. S., . . .
Hofmann, J. N. (2021). Serum concentrations of per- and polyfluoroalkyl substances and
Final PFAS Rule Economic Analysis
Q-15
April 2024
-------
FINAL RULE
APRIL 2024
risk of renal cell carcinoma. Journal of the National Cancer Institute, 113(5), 580-587.
doi:10.1093/jnci/djaal43
Shi, Y., Yang, L., Li, J., Lai, J., Wang, Y., Zhao, Y., & Wu, Y. (2017). Occurrence of
perfluoroalkyl substances in cord serum and association with growth indicators in
newborns from Beijing. Chemosphere, 169, 396-402.
doi: 10.1016/j. chemosphere .2016.11.050
Sidik, K., & Jonkman, J. N. (2005). A note on variance estimation in random effects meta-
regression. Journal of Biopharmaceatical Statistics, 75(5), 823-838.
Singh, G. K., & Stella, M. Y. (2019). Infant mortality in the United States, 1915-2017: Large
social inequalities have persisted for over a century. International Journal of Maternal
and Child Health and AIDS, 5(1), 19.
Solis, P., Pullum, S. G., & Frisbie, W. P. (2000). Demographic models of birth outcomes and
infant mortality: An alternative measurement approach. Demography, 57(4), 489-498.
South Carolina Department of Health and Environmental Control. (2020). PFAS Sampling
Results: Surface Water-Sourced Community Drinking Water Systems. Retrieved from
https://scdhec.gov/BOW/pfas-sampling-results
South Carolina Department of Health and Environmental Control. (2023). PFAS Sampling
Results: Ground Water-Sourced Community Drinking Water. Retrieved from
https://scdhec.gov/sites/default/files/media/document/Raw%20Data O.pdf
Stan Development Team. (2020). Stan. Retrieved from http://mc-stan.org/
Stan Development Team. (2021). Stan Modeling Language Users Guide and Reference Manual,
Version 2.26. Retrieved from https://mc-stan.org/
Starling, A. P., Adgate, J. L., Hamman, R. F., Kechris, K., Calafat, A. M., Ye, X., & Dabelea, D.
(2017). Perfluoroalkyl substances during pregnancy and offspring weight and adiposity at
birth: examining mediation by maternal fasting glucose in the Healthy Start Study.
Environ Health Perspect, 125, 067016-067011. doi:10.1289/EHP641
StataCorp. (2013a). Statu 13 logit Command Reference Manual. Retrieved from College Station,
TX: https://www.stata.com/manualsl3/rlogit.pdf
StataCorp. (2013b). Statu 13 Margins Command Reference Manual. Retrieved from College
Station, TX: https://www.stata.com/manualsl3/rmargins.pdf
StataCorp. (2019). Statu Statistical Software: Release 16. College Station, TX: StataCorp LP.
Steenland, K., Barry, V., & Savitz, D. (2018). Serum Perfluorooctanoic Acid and Birthweight:
An Updated Meta-analysis With Bias Analysis. Epidemiology, 29(6), 765-776.
doi: 10.1097/EDE.0000000000000903
Steenland, K., Tinker, S., Frisbee, S., Ducatman, A., & Vaccarino, V. (2009). Association of
perfluorooctanoic acid and perfluorooctane sulfonate with serum lipids among adults
living near a chemical plant. American journal of epidemiology, 170(10), 1268-1278.
Steenland, K., & Woskie, S. (2012). Cohort mortality study of workers exposed to
perfluorooctanoic acid. American journal of epidemiology, 176, 909-917.
Surveillance Research Program - National Cancer Institute. (2020a). SEER*Stat software
Incidence-Based Mortality - SEER Research Data, 18 Registries. Retrieved from:
seer, cancer. gov/ seerstat
Surveillance Research Program - National Cancer Institute. (2020b). SEER*Stat software
Incidence - SEER Research Limited-Field Data, 21 Registries. Retrieved from:
seer, cancer. gov/ seerstat
Final PFAS Rule Economic Analysis
Q-16
April 2024
-------
FINAL RULE
APRIL 2024
Tennessee Department of Environment and Conservation. (2023). Metro Nashville PFAS
Sampling Data. Retrieved from https://www.tn.gov/environment/policv/pfas/non-tdec-
sampling.html
Thom, T., Kannel, W., Silbershatz, H., & D'Agostino, R. (2001). Cardiovascular diseases in the
United States and prevention approaches. Hurst 's the heart, 7, 3-17.
Thomford, P. (2002). 104-week dietary chronic toxicity and carcinogenicity study with
perfluorooctane sulfonic acid potassium salt (PFOS; T-6295) in rats.
U.S. Bureau of Economic Analysis. (2021). Disposable personal income: Per capita
[A229RC0A052NBEA]. Retrieved from:
http s: //fred. stl oui sfed. or g/seri es/ A229RC 0 AO 5 2NBE A
U.S. Bureau of Economic Analysis. (2023). Gross Domestic Product: Implicit Price Deflator
[GDPDEF], retrievedfrom FRED, Federal Reserve Bank of St. Louis;
hltps: fred.stIouisfed.org series GDP DEV. October 8, 2023.
U.S. Bureau of Labor Statistics. (2022a). Medical care in U.S. city average, all urban consumers,
not seasonally adjusted, 1980-2022 CUUR0000SAM.
U.S. Bureau of Labor Statistics. (2022b). Table 1. Employment Cost Index for total
compensation, by occupational group and industry. Employment Cost Index Historical
Listing - Volume III. Retrieved from https://www.bls.gov/web/eci/eci-current-nominal-
.dollar.pdf
U.S. Census Bureau. (2017). American Community Survey Single-Year Estimates. American
Community Survey Demographic and Housing Estimates. Retrieved from:
https://data.census.gov/cedsci/
U.S. Census Bureau. (2020). Annual County Resident Population Estimates by Age, Sex, Race,
and Hispanic Origin: April 1, 2010 to July 1, 2019, downloaded at
https: ii )i )i 2.census.isov programs-surveys popest on May 3, 2022.
U.S. Energy Information Administration. (2021). Annual Energy Outlook 2021. Retrieved
August 10, 2021 https://www.eia.gov/outlooks/aeo/
U.S. Environmental Protection Agency. (2000). Arsenic Occurrence in Public Drinking Water
Supplies. EPA-815-D-00-001. Washington, D.C.
U.S. Environmental Protection Agency. (2001). Water Treatment Plant Model Version 2.0.
Retrieved from https://www.epa.gov/sites/default/files/2017-
03/documents/wtp model v. 2.0 manual 508.pdf
U.S. Environmental Protection Agency. (2005a). Guidelines for Carcinogen Risk Assessment.
/•./VI 630 I'-03 00IF. Washington, D.C. Retrieved from
https://www.epa.gov/sites/default/files/2013-09/documents/cancer guidelines final 3-
25-05.pdf
U.S. Environmental Protection Agency. (2005b). Occurrence and Exposure Assessment for the
Final Long Term 2 Enhanced Surface Water Treatment Ride, EPA 815-R-06-002. (EPA
815-R-06-002).
U.S. Environmental Protection Agency. (2008). Final Ozone NAAOSRegulatory Impact
Analysis. (EPA 452/R-08-003). Research Triangle Park, North Carolina
U.S. Environmental Protection Agency. (2010). Guidelines for Preparing Economic Analyses.
U.S. Environmental Protection Agency. (2011a). The Benefits and Costs of the Clean Air Act
fi'om 1990 to 2020. Final Report - Rev. A. Retrieved from
https://www.epa.gov/sites/default/files/2015-07/documents/fullreport rev a.pdf
Final PFAS Rule Economic Analysis
Q-17
April 2024
-------
FINAL RULE
APRIL 2024
U.S. Environmental Protection Agency. (201 lb). Exposure Factors Handbook, 2011 Edition
(Final). (EPA-600-R-09-025F). U.S. Environmental Protection Agency, Washington, DC
U.S. Environmental Protection Agency. (2012). Revisions to the Unregulated Contaminant
Monitoring Regulation (UCMR 3) for Public Water Systems; Final Ride. Federal Register
Retrieved from https://www.federalregister.gov/documents/2012/05/02/2012-
9978/revisions-to-the-unregulated-contaminant-monitoring-regulation-ucmr-3-for-public-
water-systems
U.S. Environmental Protection Agency. (2015). Benefit and Cost Analysis for the Effluent
Limitations Guidelines and Standards for the Steam Electric Power Generating Point
Source Category. (EPA-821-R-15-005). US EPA, Office of Water, Washington DC
20460
U.S. Environmental Protection Agency. (2016a). Health effects support document for
Perfluorooctane Sulfonate (PFOS). (EPA 822-R-l6-002). Retrieved from
https://www.epa.gov/sites/production/files/2016-05/documents/pfos hesd final 508.pdf
U.S. Environmental Protection Agency. (2016b). Health effects support document for
Perfluorooctanoic Acid (PFOA). (EPA 822-R-16-003). Retrieved from
https://www.epa.gov/sites/production/files/2016-05/documents/pfoa hesd final-plain.pdf
U.S. Environmental Protection Agency. (2017). Monitoring Unregulated Drinking Water
Contaminants: UCMR 3 (2013-2015) Occurrence Data. Retrieved from
https://www.epa.gov/dwucmr/occurrence-data-unregulated-contaminant-monitoring-
rule#3
U.S. Environmental Protection Agency. (2019). Benefit and Cost Analysis for the Proposed
Revisions to the Effluent Limitations Guidelines and Standards for the Steam Electric
Power Generating Point Source Category [DCN SE07401], (EPA-821-R-19-011). US
EPA, Office of Water, Washington, DC 20460
U.S. Environmental Protection Agency. (2021a). Analysis of Cardiovascular Disease Risk
Reduction as a Result of Reduced PFOA and PFOS Exposure in Drinking Water. (EPA
815-D-21-001). Retrieved from
https://sab.epa.gov/ords/sab/f?p=100:19:9687145615924:::RP.19:P19 ID:963
U.S. Environmental Protection Agency. (2021b). Safe Drinking Water Information System 2021
Quarter 4 Database. Retrieved from https://www.epa.gov/ground-water-and-drinking-
water/safe-drinking-water-information-svstem-sdwis-federal-reporting
U.S. Environmental Protection Agency. (2022). Review of EPA's Analyses to Support EPA's
National Primary Drinking Water Ridemaking for PFAS. EPA-SAB-22-008. Washington,
D.C. Retrieved from
https://sab.epa.gov/ords/sab/f?p=l00:18:16490947993:::RP.18:P18 ID:2601#report
U.S. Environmental Protection Agency. (2023a). Economic Analysis of the Proposed Regulation
of Methylene Chloride Under TSCA Section 6(a). (RIN2070-AK70). Washington, D.C.
Retrieved from https://www.govinfo.gov/content/pkg/FR-2023-05-03/pdf/2Q23-
09184.pdf
U.S. Environmental Protection Agency. (2023b). Environmental Benefits Mapping and Analysis
Program - Community Edition User's Manual. Washington, D.C. Retrieved from
https://www.epa.gov/sites/default/files/2015-04/documents/benmap-
ce user manual march 2015.pdf
Final PFAS Rule Economic Analysis
Q-18
April 2024
-------
FINAL RULE
APRIL 2024
U.S. Environmental Protection Agency. (2024a). Appendix - Office of Water Final Human
Health Toxicity Assessment for Perflaorooctane Sulfonic Acid (PFOS). EPA-815-R-24-
007. Washington, D.C.
U.S. Environmental Protection Agency. (2024b). Office of Water Final Human Health Toxicity
Assessment for Perflaorooctane Sulfonic Acid (PFOS). EPA-815-R-24-009. Washington,
DC.
U.S. Environmental Protection Agency. (2024c). Office of Water Final Human Health Toxicity
Assessment for Perfiuorooctanoic Acid (PFOA). EPA-815-R-24-008. Washington, D.C.
U.S. Environmental Protection Agency. (2024d). Per- and Polyfluoroalkyl Substances (PFAS)
Occurrence & Contaminant Background Support Document for the Final PFAS National
Primary Drinking Water Regulation. EPA-815-R-24-013.
U.S. Environmental Protection Agency. (2024e). Technical Support Document - Technologies
and Cost for Removing Per- and Polyfluoroalkyl Substances (PFAS) from Drinking
Water. EPA-815-R-24-012.
U.S. Office of Management and Budget. (2003). Circular A-4: Regulatory Analysis.
Washington, DC: OMB
U.S. Office of Management and Budget. (2023). Circular A-4: Regulatory Analysis.
Washington, DC: OMB Retrieved from https://www.whitehouse.gov/wp-
content/uploads/2023/1 lZCircularA-4.pdf
Uno, H., Cai, T., Pencina, M. J., DAgostino, R. B., & Wei, L. J. (2011). On the C-statistics for
evaluating overall adequacy of risk prediction procedures with censored survival data.
Stat. Med., 50(10), 1105-1117. doi:10.1002/sim.4154
Valvi, D., Oulhote, Y., Weihe, P., Dalgard, C., Bjerve, K. S., Steuerwald, U., & Grandjean, P.
(2017). Gestational diabetes and offspring birth size at elevated environmental pollutant
exposures. Environment international, 107, 205-215.
Vermont Department of Environmental Conservation. (2023). State PFAS Monitoring Results
Received. Retrieved from https://dec.vermont.gov/water/drinking-water/water-qualitv-
monitoring/pfas
Verner, M. A., Loccisano, A. E., Morken, N. H., Yoon, M., Wu, H., McDougall, R., . . .
Longnecker, M. P. (2015). Associations of Perfluoroalkyl Substances (PFAS) with Lower
Birth Weight: An Evaluation of Potential Confounding by Glomerular Filtration Rate
Using a Physiologically Based Pharmacokinetic Model (PBPK). Environmental health
perspectives, 725(12), 1317-1324. doi: 10.1289/ehp. 1408837
Vieira, V. M., Hoffman, K., Shin, H., Weinberg, J. M., Webster, T. F., & Fletcher, T. (2013).
Perfiuorooctanoic acid exposure and cancer outcomes in a contaminated community: A
geographic analysis. Environmental health perspectives, 121(3), 318-323.
Virginia Department of Health. (2021). Virginia Per and Polyfluoroalkyl Substances (PFAS) in
Drinking Water Sample Study. Retrieved from
https://www.vdh.virginia.gov/content/uploads/sites/14/2021/Q9/VA-PFAS-Sample-
Studv-Summary.pdf
von Hippel, P. T. (2015). The heterogeneity statistic 12 can be biased in small meta-analyses.
BMC medical research methodology, 75(1), 1-8.
Wan, X., Zhang, Y., Tan, C., Zeng, X., & Peng, L. (2019). First-line nivolumab plus ipilimumab
vs sunitinib for metastatic renal cell carcinoma: a cost-effectiveness analysis. JAA4A
Oncology, 5(4), 491-496.
Final PFAS Rule Economic Analysis
Q-19
April 2024
-------
FINAL RULE
APRIL 2024
Wang, Y., Adgent, M., Su, P.-H., Chen, H.-Y., Chen, P.-C., Hsiung, C. A., & Wang, S.-L.
(2016). Prenatal exposure to perfluorocarboxylic acids (PFCAs) and fetal and postnatal
growth in the Taiwan Maternal and Infant Cohort Study. Environmental Health
Perspectives, 124(11), 1794-1800.
Washino, N., Saijo, Y., Sasaki, S., Kato, S., Ban, S., Konishi, K., . . . Saito, K. (2009).
Correlations between prenatal exposure to perfluorinated chemicals and reduced fetal
growth. Environmental health perspectives, 117(4), 660-667.
West Virginia Department of Health and Human Resources. (2023). Drinking Water Watch
Data. Retrieved from https://dww.wvdhhr.org/DWWpublic/.
Whitworth, K. W., Haug, L. S., Baird, D. D., Becher, G., Hoppin, J. A., Skjaerven, R., . . .
Wilson, R. (2012). Perfluorinated compounds in relation to birth weight in the Norwegian
Mother and Child Cohort Study. American journal of epidemiology, 175(12), 1209-1216.
Whooley, M. A., De Jonge, P., Vittinghoff, E., Otte, C., Moos, R., Carney, R. M., . . . Feldman,
M. D. (2008). Depressive symptoms, health behaviors, and risk of cardiovascular events
in patients with coronary heart disease. Jama, 300(20), 2379-2388.
Wickham, H. (2007). Reshaping Data with the reshape Package. Journal of statistical software,
27(12), 1 -20. doi: 10.18637/jss.v021.il2
Wickham, H. (2016). ggplotl: elegant graphics for data analysis: springer.
Wickham, H., Fran
-------
FINAL RULE
APRIL 2024
maternal determinants and associations with birth outcomes. Journal of hazardous
materials, 380, 120867.
Yang, Q., Guo, X., Sun, P., Chen, Y., Zhang, W., & Gao, A. (2018). Association of serum levels
of perfluoroalkyl substances (PFASs) with the metabolic syndrome (MetS) in Chinese
male adults: A cross-sectional study. Science of the total environment, 621, 1542-1549.
doi: 10.1016/j.scitotenv.2017.10.074
Zhang, Y., Beesoon, S., Zhu, L., & Martin, J. W. (2013). Biomonitoring of perfluoroalkyl acids
in human urine and estimates of biological half-life. Environmental Science &
Technology, ¥7(18), 10619-10627.
Final PFAS Rule Economic Analysis
Q-21
April 2024
------- |