ECONOMIC ASSESSMENT
OF REDUCING FLUORIDE
IN DRINKING WATER
November 1985
Abt Associates Inc., Cambridge, Massachusetts
-------
Abt Associates Inc.
55 Wheeler Street, Cambridge, Massachusetts 02138
Telephone • Area 617-492-7100
TWX: 710-3201382
ECONOMIC ASSESSMENT
OF REDUCING FLUORIDE
IN DRINKING WATER
November 1985
Submitted to:
Mr. George Denning
Office of Drinking Water
Environmental Protection Agency
401 M Street SW (WH-550)
Washington, D.C. 20460
An Equal Opportunity Employer
-------
REPORT DOCUMENTATION PAGE
Form . pproved
UM8 Jo Vo3 .old8
e ort — :.‘ en c . — — — - — . ‘ :‘ - — —e • .. —; — s—; i .a cL;
‘e i ’ _—: ——-‘: ::: c’ ——:.:— —sr I r.s Lt C e l—, e : e ‘ r
c :iIe ion y , r’r cn r c rJ e ‘r — - ‘_r:. ’ i’ — .jre s .2r es re. ri:e c st cn Ccer crs I.? , ‘ sQn
2:4 r. on..—22fl’.C -. . _- -. - ‘ cer..cr. eř .ricn .‘iec( iC4 ) 3S) :: srnçt n JC O1
I AGENCY USE ONLY (Ledve Iank) 2 REPORT O TE 3 REPORT TYPE ANO DATES COVERED
Nov. 1985
4. TITLE AND SUBTITLE
Economic Assessment of Peducing Fluoride in
Drinking Water
5 FUNDING NUMBERS
6 AUTHOR(S)
7 PERFORMING ORGANIZATION NAME(S) AND ADORESS(E5)
Abt Associates Lic.
55 Wheeler St.
Cambridge, Massachusetts
8 PERFORMING ORGANIZATION
REPORT NUMBER
9. SPONSORINGIMONITORING AGENCY NAME(S) , ND ADDRESS(ES)
.
U.S. E,nvironmental Protection Agency
Office of Water
401 M St., SW
Washington, DC 20460
10. SPONSORING’ MONITORING
AGENCY REPORT NUMBER
EPA o/Q—86-oO1
-“ ‘
I
II. SUPPLEMENTARY NOTES
1 2a DISTRIBUTION / AVAILABILITY STATEMENT
I 2b DISTRIBUTION CODE
13. ABSTRACT (Ma .esmum200 words)
This report assesses the economic impacts of reducing fluoride in
drinking water. The principal elements of the report are:
definition of the fluoride contamination problem, review of
regulatory alternatives, assessment of the benefits of fluoride
removal, assessment of the costs of removing fluoride, analysis
of regulatory flexibility and paperwork requirements, and exploratio
of uncertainty in the estimates of costs and benefits.
14 SUBJECT TERMS
fluoride, fluoridation, drinking water, cost
15 NUMBER OF PAGES
84
16 PRICE CODE
17. SECURITY CLASSIFICATION 18 SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION
OF REPORT OF THIS PAGE OF ABSTRACT .
unclassified unclassified unclassified
20. LIMITATION OF ABSTRACT
.
unlimited
NSN 7540-01-280-5500
Standard Form 298 Rpv 2 99)
-------
TABLE OF CONTENTS
Chapter 1 EXECUTIVE SUMMARY 1
1.1 Summary of Findings 1
1.2 Problem Definition 2
1.3 Regulatory Alternatives 3
1.4 Assessment of Benefits 4
1.5 Assessment of Costs 5
1.6 Regulatory Flexibility and Paperwork Analyses .7
1.7 Uncertainty of Benefit and Cost Estimates 8
1.8 Summary 9
Chapter 2 PROBLEM DEFINITION 12
2.1 Summary of Health Benefits of Fluoride Removal 12
2.2 Occurrence of Fluoride in Drinking Water 13
2.3 Control Technologies for Fluoride Removal 15
2.4 Management Strategies for Fluoride 15
2.4.1 Market Mechanisms 15
2.4.2 Safe Drinking Water Act 19
2.4.3 State Actions to Control Fluoride 19
2.5 Summary of Issues 19
Chapter 3 REGULATORY ALTERNATIVES 22
3.1 Option 1 22
3.2 Option 2 23
3.3 Option 3 23
-------
Chapter 4 ASSESSMENT OF BENEFITS 25
4.1 Bone Effects 25
4.1.1 Osteosclerosis 25
4.1.2 Osteoporosis 29
4.2 Dental Effects 29
4.2.1 Adverse Dental Effects 29
4.2.2 Fluoride and Caries 30
Chapter 5 ASSESSMENT OF COSTS 34
5.1 Technologies for Reducing Fluoride 34
5.2 Probabilities of Selecting Removal Measures 35
5.3 Computation of Costs 36
5.3.1 Assumptions 36
5.3.2 Cost of Alternatives 38
5.3.3 Conclusions 45
5.4 Monitoring Requirements 47
Chapter 6 REGULATORY FLEXIBILITY ANALYSIS AND PAPERWORK ANALYSIS .49
6.1 Regulatory Flexibility Analysis 49
6.2 Paperwork Analysis 52
APPENDIX A Computation of Number of Cases of Severe and Moderate Dental
Fluorosis Avoided 54
APPENDIX B Costs of Repair of Dental Fluorosis 63
APPENDIX C Uncertainty of Costs and Benefits 67
APPENDIX D Additional Costs for Caries Treatment If a Fluoride MCL is Set 71
-------
LIST OF TABLES
TABLE 1—1 Summary of Issues and Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
TABLE 1—1 (Cont.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
TABLE 2-1 Number of Water Systems Exceeding Fluoride Concentrations 14
TP BLE 4—1 Dean’s Fluorosis Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
TABLE 4-2 Relationship Between Community Fluoride Concentrations and Dental
Caries——Children7to l2Yearsof Age 33
TABLE 5-1 Social Costs to the Nation of Fluoride Removal ...... . .. .. . . ... . . . . .,. .40
TABLE 5-2 Social Costs Per System For Fluoride Removal ...... . . .. .. . .. .41
TABLE 5—3 Utility Cost Impacts——Median Water Rate . 43
TABLE 5-4 Increase in Annual Household Water Bill Attributable to Fluoride
Removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
TABLE 5—5 Utility Cost Impacts 46
TABLE 6-1 Number of Water Systems by Population Served .. . . .. . .51
TABLE A-I Percentage Distribution of Mottled Enamel Scores Among Texas
Children . . 56
TABLE A-2 Percentage Distribution of Mottled Enamel Scores Among Illinois
School Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
TABLE A-3 Dose-Response Relationships For Fluoride in Drinking Water and
Dental Fluorosis . . . . . . . . . . . . . . . . . . . 59
TABLE A-4 Estimated Population Exposed to Fluorides in Drinking Water and
Dental Fluorosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60
TABLE A-5 Number of Cases of Severe and Moderate Dental Fluorosis Avoided
Annually at Alternate MCLS for Fluoride in Drinking Water 62
TABLE B-I Average Costs for Repairing Cosmetic and Functional Damage to
Teeth Due to Fluorosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
TABLE B-2 Avoided Costs for Remedial Repair of Moderate and Severe Dental
Fluorosis if an MCL is Promulgated.... ..... . ..... . .... . .... ....... . .66
-------
Chapter 1
EXECUTIVE SUMMARY
This report assesses the economic impacts of reducing fluoride in drinking water.
The principal elements of the report are: definition of the fluoride contamination
problem, review of regulatory alternatives, assessment of the benefits of fluoride
removal, assessment of the costs of removing fluoride, analysis of regulatory flexibility
and paperwork requirements, and exploration of uncertainty in the estimates of costs and
benefits. This Executive Summary highlights the issues and findings for each of these
elements.
1.1 Summary of Findings
• Fluoride contamination of drinking water (above I mg/I) affects
about 5000 water systems serving 30 million people; however,
only 1300 water systems serving 900,000 people have
concentrations above 2 mg/I, and 300 water systems serving
188,000 people concentrations above 4 mg/I.
• Fluoride contamination can lead to dental fluorosis in children
and, where concentration levels exceed about 4 mg/I, to
skeletal I luorosis, which, under some conditions, may be
crippling.
• EPA can regulate fluoride to protect against dental fluorosis
with a recommended maximum contaminant level of I or 2
mg/i, or to protect against skeletal I luorosis with a
recommended MCL of 4 mg /I, or drop the primary drinking
water regulation for fluoride altogether.
• The benefits of removing fluoride are reduced skeletal fluorosis
(not quantifiable given the existing data) and reduced moderate
and severe dental fluorosis -- 17,100 to 31,400 cases per year if
an MCL of 1 mg/I is adopted, 1700 to 2700 cases per year if an
MCL of 2 mg/I is adopted, and 300 to 500 cases per year if an
MCL of 4 mg/I is adopted.
• The costs of fluoride removal, in present value terms, are $4.95
billion for an MCL of I mg/i, $206 million for an MCL of 2
mg/I, and $43 million for an MCL of 4 mg/I.
• There is a large range of uncertainty around the estimates of
benefits; the uncertainty of costs has not yet been determined.
1
-------
1.2 Problem Definition
A perspective on the fluoride problem can be gained by considering the kinds of
effects on public health and welfare caused by fluoride contamination of drinking water,
the pattern of occurrence of fluoride in drinking water, the technological means of
reducing fluoride, and the institutional means available for managing fluoride
contamination.
Fluoride contamination of drinking water can lead to objectionable dental effects
and adverse skeletal effects. The dental effects of fluoride have been studied for at
least fifty years and the general conclusions are widely agreed upon. Fluoride is
beneficial in reducing caries in children but at concentrations above the dental optimum
(about 1 mg/I or one part per million) moderate or severe dental fluorosis can occur in a
fraction of the population of children. Moderate and severe forms of dental fluorosis are
characterized by brown stains on many teeth and possibly pitting of the teeth.
Osteosclerosis can occur in people exposed to at least 4 mg/I to 8 mg/I of fluoride in
drinking water. This condition is characterized by thickening of the bones and in
extreme situations, by crippling effects. Further detail on dental and skeletal effects is
provided in section 1.3.
Fluoride occurs at concentrations of I mg/I or more in about 5000 public water
systems. Most of these systems are small groundwater systems. At higher
concentrations, the occurrence of fluoride drops off dramatically. For example, there
are about 1300 systems with concentrations above 2 mg/I and about 300 systems with
concentrations above 4 mg/I.
There exist several control technologies for removing fluoride. Those which are
most likely to be selected by utilities wishing to reduce levels of fluoride are centralized
activated alumina, centralized reverse osmosis, and point of use treatments for
households having small children. Other possible measures for removing fluoride are:
optimization of an existing lime softening plant, regionalization, tapping an alternate
source, point of use treatments for all households, and possibly substitution of bottled
water for drinking and cooking purposes.
There are several institutional means which might be used to register the demand
for fluoride removal and increase the supply of fluoride-reduced water. One is the
marketplace. However, market transactions are not likely to lead to an efficient
allocation of resources in this case: people are often uninformed about the presence of
high concentrations of fluoride and about the health and related risks associated with
fluoride; people often cannot put dollar values on health risks avoided; and water is
2
-------
supplied by regulated local monopolies, not by competing suppliers in a marketplace.
One institutional alternative to the marketplace is regulation under the Safe
Drinking Water Act. Under this act, EPA must develop regulations for contaminants
which may have any adverse effect on human health. These regulations may be in the
form of maximum contaminant levels (MCLs) or required treatment techniques. Under
the National Interim Primary Drinking Water Regulations established under the Safe
Drinking Water Act, MCLs were set for fluoride as a function of annual average
maximum daily air temperature. Revised regulations are to be developed by first setting
a recommended MCL. Recommended maximum contaminant levels are to be set at a
level at which no known or anticipated adverse human health effects occur, allowing for
an adequate margin of safety. Then MCLs are to be set as close to the recommended
MCLs as is feasible taking into account the best technology, treatment techniques, and
other means, considering costs. Primary drinking water regulations pertain to
contaminants which may have an adverse effect on human health. Secondary reguations
deal with odor, appearance, or other characteristics of water which adversely affect the
public welfare.
A third institutional alternative is state management of fluoride. States have, in
general, not acted to remove fluoride from drinking water. There has been little
enforcement of the existing standard under the National Interim Primary Drinking Water
Standards. Moreover, several states have petitioned EPA to delete fluoride from the
Primary Drinking Water Regulations because of the high cost of fluoride removal and
because the states believe that fluoride does not present a health risk.
Thus the fluoride “problem” encompasses objectionable dental fluorosis and skeletal
fluorosis which may, under some circumstances, be crippling. Fluoride occurs in the
water supplies of about 5000 water systems at concentrations at or above 1 mg/I but in
much fewer systems at higher concentrations. There exist several technological means
to removing fluoride from drinking water. Market institutions and state management
approaches are not likely to lead to reduced fluoride exposure. Under the Safe Drinking
Water Act, EPA can issue regulations to reduce fluoride exposure and hence reduce the
incidence of dental fluorosis and skeletal fluorosis.
1.3 Regulatory Alternatives
Under the Safe Drinking Water Act, EPA published proposed regulatory alternatives
in the Federal Register of May 14, 1985. Three options were presented.
Option I proposes a primary drinking water regulation based upon protection from
3
-------
moderate and severe dental fluorosis. This option regards dental fluorosis as a health
effect, not simply a cosmetic effect. Psychological impairment due to the disfigurement
resulting from dental fluorosis may also be regarded as a health effect. Under Option I,
EPA could issue a recommended MCL of either I or 2 mg/I, selecting different balances
between caries prevention and fluorosis prevention.
Option 2 proposes a primary drinking water regulation based upon protection from
crippling skeletal fluorosis. A primary recommended MCL of 4 mg/I would be
established to protect against osteoscierosis. In addition, a secondary MCL of 2 mg/I
would be established to provide guidance to the public, states, utilities, and physicians
and dentists about the cosmetic effects of dental fluorosis. Dental fluorosis is regarded
as an objectionable aesthetic effect, but not a health effect under this option.
Notification of the public would be required if fluoride concentrations exceeded 2 mg/I.
Option 3 would delete fluoride from the primary drinking water regulations based
upon a finding that levels of fluoride in U.S. drinking water are not associated with any
adverse health effects. A secondary MCL of 2 mg/I would be established, however. This
option presumes that there is an insignificant risk of crippling skeletal fluorosis given the
concentrations of fluoride in drinking water supplies and that dental effects are cosmetic
in nature.
1.4 Assessment of Benefits
Fluoride removal can affect the condition of both bones and teeth. Skeletal effects
are discussed first.
Fluoride induces new bone formation in which calcium is replaced by fluoride. The
result is densification and enlargement of bone--osteosclerosis. Early stages of
osteoscierosis are characterized by slight enlargement of trabeculae in the lumbar
spine. More advanced stages result in densification and enlargement of the pelvis and
vertebral column. Densification then spreads to the ribs and extremities. Several
studies have been conducted in the United States and elsewhere on the relationship
between fluoride and ostesclerosis. There are insufficient data to develop a dose -
response function, so the number of cases of osteoscierosis avoided at different MCLs
cannot be calculated. The literature suggests that mild osteosclerosis is unlikely to
occur at fluoride concentrations in drinking water below about 4 mg/I. Concentrations of
fluoride would probably have to be higher to induce crippling fluorosis. Only a portion of
the population exposed to fluoride would be likely to develop osteosclerosis. The Surgeon
General has indicated that changes in bone density se are not adverse health effects
4
-------
but that crippling fluorosis and arthralgia are adverse health effects.
Fluoride may be beneficial in offsetting osteoporosis. Fluoride has been used as a
treatment for osteoporosis and, in areas with high fluoride concentrations, the
prevalence of osteoporosis may be less pronounced than elsewhere.
The use of fluoride to prevent dental caries is well established. However,
concentrations of fluoride in drinking water above the dental optimum can lead to
moderate or severe fluorosis of the teeth. Moderate fluorosis is characterized by brown
stains on many teeth and severe fluorosis is characterized by pitting of the teeth and
heavy brown stains on many teeth. Dose-response curves can be roughly estimated based
on existing data. These indicate that, as the concentration of fluoride in drinking water
increases, the proportion of children with moderate or severe dental fluorosis increases.
At concentrations above about 2.5 mg/I, perhaps as many as ten to fifteen percent of the
exposed children will develop severe or moderate dental fluorosis. At concentrations
above 4 mg/I, the proportion of exposed children developing severe or moderate fluorosis
may exceed 30% to 40%.
The number of cases of moderate and severe dental fluorosis avoided per year at
various MCLs is estimated at:’
17,100 to 31,400 at an MCL of 1 mg/I
1,700 to 2,700 at an MCL of 2 mg/I
700 to 1,100 at an MCL of 3 mg/I
300 to 500 at an MCL of 4 mg/I
1.5 Assessment of Costs
In order to assess the costs of removing fluoride from drinking water it is necessary
to consider the technologies available for removing fluoride, the probabilities that
utilities would select each of these measures to reduce fluoride, and the capital and
operating costs of fluoride reduction measures. The technologies available for reducing
fluoride are indicated in section 1.2., centralized activated alumina, centralized reverse
osmosis, and point-of-use treatments being most likely to be selected. The probability
‘The ranges reflect two dose-response functions based upon two different studies,
one in Texas and one in Illinois.
5
-------
that a water system exceeding a fluoride standard would select a removal measure
depends on system size, costs of the technology, effectiveness of the technology in
removing fluoride, the population at risk (all households or households with children), and
the likelihood of other inorganic contaminants being present along with fluoride.
Systems serving fewer than 1000 people are most likely to select point of use treatments
employing activated alumina. Only households with small children would receive these
devices. Systems serving over 1000 people would be most likely to select central
activated alumina.
In developing cost figures it was assumed that utilities facing a primary MCL would
act to reduce the concentration of fluoride to the standard but that utilities would not
invest in removal measures if they faced a secondary MCL. This is consistent with past
utility behavior and enforcement of primary MCLs. To provide a range of cost estimates
primary MCLs of I mg/i, 2 mg/I, 3mg/i, and 4 mg/I were analyzed. The cost data
available for this analysis assume that the influent concentration of fluoride is 3.2 mg/I
and that the standard is 2mg/I. Thus, for those removal measures whose costs are
dependent on influent and effluent concentrations, the cost analysis must be regarded as
preliminary.
Applying these assumptions to EPA’s cost and technologies data, the capital costs
(in 1982 dollars) for attaining various MCLs is:
$ 968 million for an MCL = I mg/I
$54.6 million for an MCL 2 mg/I
$11.9 million for an MCL = 4 mg/I
The annual operating costs would be:
$ 268 million for an MCL = I mg/i
$10.2 million for an MCL = 2 mg/I
$ 2.1 million for an MCL 4 mg/I
In terms of present value (calculated for a 20 year time horizon using a real interest rate
of 3% as a social discount rate), the fluoride removal costs are:
$4.95 billion for an MCL = I mg/I
$ 206 million for an MCL = 2 mg/I
$ 43 million for an MCL = 4 mg/I
6
-------
The drop in costs from an MCL of I mg/i to an MCL of 2 mg/I is due to the large number
of systems with fluoride concentrations between 1 mg/I and 2 mg/I.
The increase in the household water bill, as estimated by EPA’s Financing Needs
Model, is likely to be around $40 to $70 per year for households in small communities
which adopt point of use treatment employing activated alumina. This assumes that
costs are spread over all customers of the water system. Other technologies used by
small systems would raise the household water bill more. For larger systems, most would
adopt centralized activated alumina. Household water bills would increase by $18 to $72
per year. Other treatments would cost more. Because of economies of scale, water bill
increases would be lower in larger communities.
Under the National Interim Primary Drinking Water Regulations, water systems
must now monitor for fluoride. Little additional cost for monitoring is anticipated under
the regulatory options considered in this report. Nationwide monitoring costs are about
$170,000 per year.
1.6 Regulatory Flexibility and Paperwork Analyses
The Regulatory Flexibility analysis is concerned with the significance of the effect
of proposed regulations on a substantial number of small businesses. Most water systems
qualify as a small business under Small Business Administration rules. The upper cut-off
for a small business is a water system serving about 50,000 people and nearly all water
systems fall into this category. As a guideline, EPA uses a figure of 20% of affected
entities as being a substantial number. Fewer than 20% of water systems would be
affected by any fluoride MCL and if the MCL were set at 2 mg/I or greater, fewer than
2.5% of water systems would be affected. The degree of impact on affected water
systems can be ascertained by using EPA’s Financing Needs Model. This model indicates
that, for systems serving 500 or fewer people, 7% to 13% may suffer financial difficulty 1
attributable to a fluoride regulation (assuming these systems choose point of use
treatment with activated alumina). For larger systems, 5% to 14% may suffer financial
‘Difficulty is defined as having to increase rates by at least $1.00 per thousand
gallons, or having new capital costs exceed the value of existing assets, or having a water
rate in excess of $3.00 per thousand gallons.
7
-------
hardship attributable to a fluoride regulation (assuming these systems select centralized
activated alumina).
The Paperwork Reduction Act is concerned with minimizing the Federal paperwork
burden for small businesses and state and local governments. The principal paperwork
burden resulting from a fluoride regulation is likely to be monitoring requirements.
Additional monitoring requirements due to changes in existing fluoride regulations are
likely to be very small. Approximately 30,000 samples of water are required per year.
The specifications of this analysis are described in section 5.4 of this report.
Groundwater systems must monitor for fluoride every three years and water systems
using any surface water must monitor for fluoride every year. If the MCL is exceeded
three additional analyses must be completed within one month. If the average
concentration from these four tests exceeds the MCL, a state-designated monitoring
frequency is to be followed until the concentration is less than the MCL on ‘two
successive samples. Typical laboratory costs per test are less than $10 per sample.
1.7 Uncertainty of Benefit and Cost Estimates
There exists uncertainty in the estimates of benefits and costs of fluoride
removal. This uncertainty stems from unexplained variation and error in dose-response
functions, error in the estimation of fluoride occurrence, error in the estimation of the
population served by water systems having fluoride, error in treatment cost estimates,
and errors in estimating the proportion of utilities which would select a given fluoride
removal measure. Some of these uncertainties can be expressed in terms of statistical
confidence intervals. The range of cases of dental fluorosis avoided per year based on
the 95% confidence interval of the slope of the dose - response curve is as follows:
8
-------
Range of Cases of Dental Flurosis Avoided per Year
(95% Confidence Interval)
Moderate Fluorosis
University of Texas NIDR Data Severe Fluorosis
Data ( Illinois) ( NIDR data)
MCL
I mg/I 730 to 28884 0 to 19079 0 to 41085
2 mg/I 605 to 2004 0 to 1322 0 to 2839
3 mg/I 257 to 849 0 to 560 0 to 1196
4 mg/I 107 to 353 0 to 232 0 to 491
Because of the small number of communities involved and the omission of explanatory
variables which would improve goodness of fit, the 95% confidence intervals for the
dose-response functions are quite large. For other factors, especially factors affecting
the cost estimates, sensitivity analyses are appropriate to determine the range of
uncertainty. As of the date of this report, these sensitivity analyses had not been
conducted.
1.8 Summary
Table 1-1 summarizes the main issues and findings of the economic assessment.
9
-------
Table 1-1
SUMMARY OF ISSUES AND FINDINGS
Problem
• Fluoride contamination of drinking water leads to dental fluorosis and skeletal
fluorosis
• About 5000 water systems have fluoride at concentrations greater than 1 mg/I,
but only 1300 have fluoride at concentrations above 2 mg/I
• Control technologies exist to remove fluoride from drinking water, including
activated alumina, reverse osmosis, and point of use treatments
• The marketplace and state management approaches are not likely to be as
effective as EPA regulations in reducing fluoride exposure
Regulatory Alternatives
• Option I: a primary recommended MCL of I mg/i or 2 mg/I to protect against
dental fluorosis
• Option 2: a primary recommended MCL of 4 mg/i to protect against crippling
fluorosis and a secondary MCL of 2 mg/I to protect against aesthetically
objectionable dental fluorosis
• Option 3: no primary drinking water regulation for fluoride but a secondary
MCL of 2 mg/i
10
-------
Table I-I (continued)
Benefits of Fluoride Removal
• Reduced skeletal fluorosis (not quantifiable)
• Reduced moderate and severe dental fluorosis--roughly quantifiable: * 17,100 to
31,400 cases of moderate and severe dental fluorosis
avoided per year if MCL = I mg/I
* 1700 to 2700 cases of moderate and severe dental fluorosis avoided
per year if MCL = 2 mg/I
*300 to 500 cases of moderate and severe dental fluorosis avoided per
year if MCL = 4 mg/I
Costs of Fluoride Removal
• Present value of social costs to the nation
* $4.95 billion if MCL = 1 mg/I
* $206 million if MCL 2 mg/I
* $43 million if MCL 4 mg/I
• Additional monitoring costs negligible
Effects on Small Utilities
• Fewer than 20% affected if MCL: I mg/i, fewer than 2.5% affected if MCL = 2
mg/i or greater.
• Up to about 14% may suffer financial difficulty in complying with the
regulation.
• Little additional paperwork required for monitoring.
Uncertainty
• Large range of uncertainty of benefits; uncertainty of costs not yet determined.
11
-------
Chapter 2
PROBLEM DEFINITION
This chapter reviews the major issues concerning fluoride contamination of drinking
water. The health effects of concentrations of fluoride in excess of the dental optimum
are briefly discussed based upon the more detailed assessment of health benefits of
fluoride removal found in Chapter 4 and Appendix A. The occurrence of fluoride in
drinking water is then summarized. Following this, the technologies available for
removing fluoride from drinking water are reviewed. Then, three mechanisms for
managing fluoride contamination of drinking water are discussed: 1) the market 2) the
Safe Drinking Water Act, and 3) state actions to control fluoride. The issues are
summarized at the conclusion of the chapter.
2.1 Summary of Health Benefits of Fluoride Removal
Much of the literature on fluoride in drinking water is concerned with dental
effects. Both the advantages and disadvantages of fluoride at various concentrations
have been analyzed. The value of fluoride in preventing caries is well established and a
dental optimum concentration of fluoride in drinking water has been set on the basis of
this. At higher concentrations, however, dental fluorosis may occur. The percentage of
children exhibiting moderate or severe cases of dental fluorosis generally increases as
the concentration of fluoride increases. Moderate fluorosis is characterized by brown
stain on the teeth and severe fluorosis is characterized by heavy brown stains and pitting
of the enamel surfaces of many teeth.
The Surgeon General does not consider dental fluorosis to be an adverse health
effect, however.’ The ad hoc committee headed by the Chief Dental Officer of the U.S.
Public Health Service stated that “No sound evidence exists which shows that drinking
water with the various concentrations of fluoride found naturally in public water supplies
EPA, 50 Federal Register 20166; May 14, 1985
12
-------
in the U.S. has any adverse effect on dental health as measured by loss of function and
tooth mortality.” 1
Chapter 4 and Appendix A contain more detail on the benefits of reducing fluoride.
Fluoride in drinking water may also lead to certain skeletal effects. In particular,
bone density may increase with prolonged exposure to fluoride. In its most severe form,
skeletal fluorosis is characterized by the deposition of irregular bone deposits which can
result in crippling in a fraction of the population. Changes in bone can be found in
individuals exposed to more than S mg/i of fluoride in drinking water and possibly in
individuals exposed to 4 to 8 mg/I. Unfortunately, there are insufficient data on skeletal
effects of fluoride to construct a dose-response function and therefore to estimate the
cases of skeletal fluorosis avoided by setting alternative drinking water standards.
Further discussions of the health effects of fluoride in drinking water are found in
Chapter 4.
2.2 Occurrence of Fluoride in Drinking Water
The occurrence of fluoride in drinking water is summarized in Table 2-1. Data are
from the Community Water Supply Surveys and the Rural Water Survey for systems
having low concentrations of fluoride and from the Federal Reporting Data System
(FRDS) for systems having high concentrations of fluorides. 2 It is evident that very few
systems have fluoride concentrations above 2 mg/I and even fewer have concentrations
above 4 mg/I. There are, however, a large number of systems with concentrations of
fluoride between I and 2 mg/I. Systems exhibiting high concentrations of fluoride in
their drinking water tend to be small and obtain their water from undergroundsources.
The number of systems exceeding concentrations of 2 mg/I is about 1350. There
are about 500 systems exceeding 3 mg/I, and about 300 systems exceeding 4 mg/I.
‘ibid
2 Report prepared under contract to EPA by JRB Associates “Occurrence of
Fluoride in Drinking Water, Air and Food,” February 9, 1984.
13
-------
Table 2—i
Number of Water Systems Exceeding Fluoride Concentrations
System Size Category (population served )
Concentration 3301—
Exceeding 25—500 501—3300 50,000 50,000+ TOTAL
1 mg/i 3460 935 470 75 4940
2 mg/i 106i 235 49 1 1346
3 mg/i 366 108 20 0 494
4 mg/i 223 47 i2 0 282
Source: EPA, “Occurrence of Fluoride in Drinking Water, Air and Food,” Report
by JRB Associates, February 9, 1984
14
-------
2.3 Control Technologies for Fluoride Removal
Several technologies could be used to remove excess fluoride from drinking
water.’ These include: centralized treatment with activated alumina, centralized
treatment with reverse osmosis, optimization of lime softening at an existing softening
plant, regionalization, tapping an alternate source, substitution of bottled water for tap
water, and point of use treatments. Point of use treatments are assumed to be
purchased, installed, and maintained by the water utility or community. Technologies
available for point of use treatments are activated alumina and reverse osmosis.
The treatment technologies most likely to be selected are centralized activated
alumina, centralized reverse osmosis, and point of use treatments for households having
small children. These technologies and the likelihood of their being selected to remove
fluoride are discussed in Chapter 5.
2.4 Management Strategies for Fluoride
This section reviews the basic management alternatives for removing fluoride.
They are: using the market to register the demand for and supply of fluoride-reduced
water, using the Safe Drinking Water Act, and using state regulatory mechanisms to
control fluoride.
2.4.1 Market Mechanisms
The marketplace is one possible alternative to fluoride regulation. If there were no
contaminant standards for fluoride, could the marketplace efficiently allocate
resources? There are several reasons to think it cannot: a) people are often uninformed
about the presence of high concentrations of fluoride and the health risks associated with
high concentrations of fluoride, b) people typically cannot put a dollar value on health
risks avoided, and c) water is supplied by local monopolies, not by competing utilities.
Thus, there are difficulties in registering the demand for water quality, as a function of
dollars, and in representing competitive costs of supply in the marketplace. The
‘EPA, Office of Drinking Water, “Technologies and Costs for the Removal of
Fluoride from Potable Water Supplies”, Report by V.]. Ciccone and Associates, May 25,
1984. The report was updated on September 10, 1985, but the costs did not change.
15
-------
following discussion addresses these points in more detail.
Some economists argue that the common good is represented by economic
efficiency, a state in which society has maximized its net benefits. Under the conditions
of perfect competition with no externalities 1 , efficiency can be reached through the
process of individual actions in the marketplace. Thus, the characteristics of a perfectly
operating market are used as a standard against which collective or governmental actions
may be judged, and a justification for such collective actions would be “imperfections” in
the market. This discussion examines imperfections in the demand and supply sides of
the market for drinking water quality, and the need for government intervention in that
market.
On the demand side, several strong conditions must be met before a perfectly
operating market can exist. These conditions deal with the preferences of consumers of
water (everyone served by public water systems).
The first condition for a preference function for water quality which will lead to a
demand schedule is that money and water quality are desirable in and of themselves.
This is an easily accepted condition.
The second condition is that money and water quality must be comparable, that is,
each individual must be able to state preferences among any (reasonable) pair of bundles 2
of money and water quality characteristics. A major problem is the absence of market
information. Since people may not have a clear understanding of the relationships
between high concentrations of fluoride in their drinking water and their health, it is
difficult to choose between money and water quality. There are two sources of lack of
information. One is lack of knowledge of the occurrence of fluoride--many people may
be unaware of the presence of fluoride in their drinking water. Second, most people are
not well informed about health risks of high concentrations of fluoride and probably
cannot express trade-of fs between bundles of money and water quality. A study of eight
California communities with drinking water contaminants, including fluoride, indicated
t Externalities are social costs and benefits that are external to the
firm’s or individual’s calculations of private costs and benefits. Pollution is
a common example. A firm would probably not account for the costs to
society of the water pollution it generates when deciding how to maximize
profits. Economists argue that failure of the market to account for these
social costs and benefits leads to inefficiency.
2 lhat is, a market basket of goods and services. For simplicity we talk in terms of
just two goods here -- water quality and all other goods represented as money.
16
-------
that only a small percentage of the population clearly understood the problem, even after
receiving notification of contamination.’ That is, most respondents were not aware of
the contaminant, health problems caused by the contaminant, and personal actions
required to avoid the health problems. In addition, because there are generally no
routine market transactions involving money and water quality, feedbacks about
purchasing too much or too little water quality cannot be relied upon to clarify
preferences among bundles of water quality and money and to allow for corrective
actions as would be the case for goods and services purchased regularly. In short, the
comparability condition is not met and so the demand for fluoride reduction is not well
defined.
Several other conditions must also be met in order to make preference functions
and demand schedules well behaved. Individuals must be able to: (I) state indifference
among several bundles of money and water quality; (2) state preferences in a transitive
manner so that if bundle A is preferred to bundle B and bundle B is preferred to bundle C,
bundle A is preferred to bundle C; and (3) form a continuous line of equally preferred
bundles between bundles which are preferred equally. Because most people’s preference
functions are unlikely to be so precisely defined, these conditions are highly unlikely to
be met and any preference functions inferred from surveys or observed behavior could
very well violate one or more of these conditions. For instance, people may not have
transitive preference functions or may not be able to identify bundles of equal
preference. Failure to meet these conditions will lead to demand schedules from which
conclusions about economic efficiency may be erroneous or impossible to draw. For
example, demand curves may be discontinuous or not exist logically and it may not be
possible to identify a social optimum of net benefits.
On the supply side of the market for drinking water quality, it has been evident for
decades that utilities do not meet the test for a competitive market. Each community,
trailer park, city surburban tract, or other customer base is served by only one water
supplier. Monopolistic conditions exist. Water utilities are natural monopolies 2 in that
they exhibit large economies of scale and competitive services would not be cost
‘William Bruvold and 3ohn Gaston, “Public Notification: Pain or Panacea,” Journal
of the American Water Works Association , vol. 72 (March 1980), pp. 124-127.
“natural monopoly” exists when one firm can supply the market (in this case a
water supply service area) at less cost than two or more firms because of economies of
scale. It is cheaper to build one water supply system per community than two or three.
17
-------
effective. (Moreover, some privately owned water companies are themselves owned by
larger water supply companies). It is generally argued that monopolies may not be
responsive to the demand for product quality and quantity because the pressure of
competing suppliers is absent.
The existence of natural monopolies in water supply has three important
implications for judging the efficiency of water markets. First, water utilities are
regulated by the States with regard to how much they can charge for water. Thus the
market price for water is not a free-market price and efficient allocation of water may
not exist.
Second, because of the natural monopoly-economies of scale characteristics of the
water industry it is not economical to offer several grades of drinking water quality in
one service area. That is, suppliers of drinking water typically cannot treat the water of
only the people who desire low concentrations of fluorides (unless home treatment units
are used but this is often far more expensive than central treatment). Consequently a
perfectly competiti. ”. market within a utility service area which provides grades of
water quality is difficult to maintain. In many cases the water of all customers must be
of the same quality.
Third because of the existence of only one water utility per community, water
suppliers may not correctly perceive the willingness of their customers to pay for water
quality improvements. Consumers cannot switch suppliers. Thus, feedbacks on price as a
function of quality cannot be readily obtained. As a result, many utility managers simply
assume that their customers would not be willing to pay for quality improvements’, even
though this might not be true.
The existence of other regulations and programs pertaining to water supply may
also be regarded as market “imperfections”. These include rate regulations, state and
federal water quality regulations, the availability of grants for water systems, and the
like.
In conclusion, neither the demand nor supply side of the market for drinking water
meets the conditions for a perfectly competitive market. Therefore, economic
efficiency cannot necessarily be achieved by EPA’s refraining from issuing regulations on
fluoride. In fact, market imperfections are so strong that government intervention is
1 Charles Stegman and Georgia Schneider, “The Cost and Effectiveness of Public
Notification of MCL Violations,” Journal of the American Water Works Association , vol.
74 (February 1982), pp. 59-6).
18
-------
required to achieve efficiency. The economic demand for fluoride removal is not
quantifiable, the public is not knowledgeable about fluoride health risks, fluoride
occurrence is poorly understood by the public, suppliers of drinking water are natural
monopolies, and drinking water cannot often be economically provided at different
quality levels in the same service area. Thus the market does not provide a mechanism
for registering the costs and benefits of fluoride reduction and for maximizing the net
benefits of fluoride removal.
2.4.2 Safe Drinking Water Act
As an alternative to the marketplace, Congress passed the Safe Drinking Water
Act. This Act requires the Environmental Protection Agency to develop regulations on
drinking water quality for public water systems. In particular, EPA must develop
• regulations for contaminants which may have any adverse effect on human health.
Regulations may be in the form of maximum contaminant levels (MCLs) or required
treatment techniques.
EPA established National Interim Primary Drinking Water Standards as required by
the Safe Drinking Water Act. Among the contaminants regulated by these interim
standards is fluoride. The regulation specifies the following maximum contaminant
levels for fluoride:
Annual Average Maximum Daily Air
Temperature for the Location in
Which the Community Water System Maximum Contaminant Level
is Situated (degrees Celsius) ( milligrams per liter )
12.0 and below 2.4
12.1 to 14.6 2.2
14.7 to 17.6 2.0
17.7 to 21.4 1.8
21.5 to 26.2 1.6
26.3 to 32.5 1.4
These maximum contaminant levels were intended to reflect the greater volume of water
consumed by peopLe living in warmer areas.
The selection of these maximum contaminant levels for fluoride was based on the
following considerations:
19
-------
• The beneficial effect of fluoride in preventing dental
caries; and
• The adverse health effects of high concentrations of
fluoride, including fluorosis of the teeth.
• Generally available technology for fluoride removal.
The maximum contaminant levels for fluoride were set a twice the level at which
maximum dental benefits are realized.
Revised regulations are to be developed using a two step procedure. First
recommended MCLs are to be set at a level at which no known or anticipated adverse
human health effects occur, allowing for an adequate margin of safety.’ These
recommended MCLs are not enforceable standards, however. The second step requires
EPA to develop an MCL as an enforceable standard. MCLs must be set as close to
recommended MCLs as is feasible taking into account the best technology, treatment
techniques, and other means while considering costs.
The Safe Drinking Water Act distinguishes between primary and secondary
standards. Primary regulations pertain to contaminants which may have an adverse
effect on human health. Secondary regulations deal with odor, appearance, or other
characteristics of water which may cause a substantial number of persons served by the
water system to discontinue use or which may otherwise adversely affect the public
welfare.
2.4.3 State Actions to Control Fluoride
State actions to control fluoride have been of two types: 1) general disinterest in
reducing fluoride, with some exceptions, and 2) opposition to the current fluoride
standard. Telephone conversations with EPA Regional Offices in Regions IV (Atlanta)
and VI (Dallas) indicated that only a few systems in Region IV have acted to remove
fluoride in response to the existing MCL and that no systems in Region VI have acted to
remove fluoride in response to the MCL. Opposition to the current fluoride standard is
reflected in the petition by South Carolina and other States to delete fluoride from the
‘See EPA, )0 Federal Register 20164ff (May 14, 1985) for the proposed
recommended MCL.
20
-------
Primary Drinking Water Regulations. 1 The petition contends that the cost of reducing
fluoride is prohibitive and that fluoride does not present a health hazard but causes only
an aesthetic effect on teeth. Thus state actions have generally not led to reduction of
excess levels of fluoride in drinking water.
2.5 Summary of Issues
Contamination of drinking water by fluoride results in increased incidence of
mottling and pitting of teeth (dental fluorosis) and osteosclerosis. EPA considers the
dental effects not to be health effects. Crippling skeletal fluorosis is considered to be an
adverse health effect, however. Unfortunately, there are insufficient data on skeletal
effects of fluoride to estimate the number of cases of osteosclerosis avoided under
alternative fluoride management strategies.
Fluoride occurs primarily in smaller water systems. Nearly 5000 public water
systems have concentrations of fluoride exceeding I mg/i (ppm), 1300 systems exceed 2
mg/I, and about 300 systems exceed 4 mg/I.
There are several technologies available to control fluoride. The measures most
likely to be adopted are centralized activated alumina, centralized reverse osmosis, and
point of use treatments for households having small children.
The basic means to controlling fluoride are the marketplace, the Safe Drinking
Water Act, and State programs. The marketplace is unable to deal with the issues
primarily because of lack of public information about health risks, because of the
difficulties people have putting a dollar value on avoiding risks from fluoride, and
because of the absence of a competitive market in water since water utilities operate as
regulated natural monopolies. The Safe Drinking water Act provides a means to
controlling fluoride contamination through the existing National Interim Primary
Drinking Water Standards and the establishment by EPA of recommended maximum
contaminant levels and maximum contaminant levels. State actions have not led to
significant reduction of excess fluoride in drinking water and indeed several states have
petitioned EPA to delete fluoride from the primary drinking water standards.
146 Federal Register 58345; December 1, 1981.
21
-------
Chapter 3
REGULATORY ALTERNATIVES
Market mechanisms were shown in Chapter 2 to be ineffective for dealing with
excess fluoride in drinking water. Thus a regulatory approach may be appropriate. This
chapter reviews the regulatory options available to EPA. Three options are discussed
based upon the presentation in the Federal Register of May 14, 1985.’
3.1 Option I
Under this option EPA would promulgate a Primary Drinking Water Regulation to
protect from moderate and severe dental fluorosis. The premises supporting this option
are a) that moderate and severe dental fluorosis are adverse health effects themselves or
b) that the psychological impairment resulting from the cosmetic effects of moderate
and severe dental fluorosis constitute an adverse health effect. The Review Panel on the
Psychological/Behavioral Effects of Dental Fluorosis concluded that “individuals who
have suffered impaired dental appearance as the result of moderate to severe fluorosis
are probably at increased risk for psychological and behavioral problems or
difficulties”. 2 Two suboptions can be considered.
Suboption A would set a recommended MCL at I mg/i. This balances the
beneficial effects of fluoride in the prevention of tooth decay with the adverse effects of
fluoride on teeth. Up to about 4% of the population in communities now exceeding I
mg/I could be expected to develop moderate or severe dental fluorosis under this
suboption. About 5000 communities now exceed a concentration of I mg/i.
Federal Register 20171 to 20172; May 14, 1985. Note that EPA is considering
dropping the graduated standard for different climates that now exists as part of the
National Interim Primary Drinking Water Standards. See section 2.4.2.
2 R. Kieck, Chairperson, Review Panel on Psychological/Behavioral
Effects of Dental Fiuorosis, Report to Office of Drinking Water, EPA,
November 17, 1984, page 1.
22
-------
Suboption B would set a recommended MCL at 2 mg/I. This option would allow for
greater benefits of fluoride in protecting against caries. Up to about 14% of the
population in communities now exceeding 2 mg/I could be expected to develop moderate
or severe dental fluorosis under this suboption. About 1300 communities would be
affected by this recommended MCL.
3.2 Option 2
Option 2 proposes a primary drinking water standard considering protection from
crippling skeletal effects of fluoride and a secondary drinking water standard based on
cosmetic effects of dental fluorosis. In particular a recommended MCL of 4 mg/I would
be set to protect against crippling skeletal fluorosis with an adequate margin for safety.
Chapter 4 discusses skeletal effects in more detail. Although no dose response curve for
skeletal fluorosis can be developed the evidence indicates that at concentrations below 8
mg/i there is little likelihood of crippling fluorosis occurring. About 300 public water
systems currently exceed 4 mg/I of fluoride in their drinking water.’
As part of Option 2, a secondary MCL of 2 mg/I would be established. Under this
option objectionable dental fluorosis is considered to affect the public welfare but not
present a health hazard. The secondary MCL would provide guidance to states and
utilities and would reflect a balance between the cosmetic effects of fluorosis and the
beneficial effects of fluoride in reducing the incidence of caries. About 1300 public
water systems exceed 2 mg/I of fluoride. Monitoring would be required. In addition,
residents of communities having over 2 mg/I of fluoride in their drinking water would be
notified of the effects of dental fluorosis and the alternatives for prevention.
Physicians, dentists, and public health officials would also be notified when the secondary
MCL is exceeded.
3.3 Option 3
This option would delete fluoride from the Primary Drinking Water Standards on
the premise that dental effects are only cosmetic and crippling bone effects are of
‘About 30% to 40% of the exposed population in these communities would,
however, experience moderate or severe dental fluorosis if the standard was 4 mg/I.
23
-------
minimal risk given concentrations of fluoride found in drinking water. Only a secondary
MCL of 2 mg/I would be published. Monitoring and public notification requirements
would still be put in place under this option to advise the public about concentrations of
fluoride above 2 mg/I. This reflects the Surgeon General’s view’ that naturally
occurring fluoride concentrations in public water systems in the U.S. do not have adverse
effects on general health. It also reflects the Surgeon General’s concern over aesthetic
effects of dental fluorosis: “I encourage communities having water supplies with fluoride
concentrations of over two times optimum to provide children up to age nine with water
of optimum fluoride concentration to minimize the risk of their developing aesthetically
objectionable dental fluorosis.” 2
‘EPA, 50 Federal Register, 20166; May 11 , 1985.
2 Quoted in: 50 Federal Register, 20166; May 14, 1985.
24
-------
Chapter 4
ASSESSMENT OF BENEFITS
Fluoride in drinking water can induce two types of effects as reported in the health
literature. One is changes in bones, including osteoclerosis and reduction in the
likelihood of osteoporosis. The other is changes in teeth, including mottling and pitting
of teeth and, as is well known, prevention of caries. The benefits of reducing fluoride in
drinking water then include 1) reduced incidence of osteosclerosis and 2) reduced
incidence of dental fluorosis including psychological effects resulting from this
disfiguration. This chapter reviews the evidence on the effects of fluoride on bones and
teeth.
4.1 Bone Effects
In bone, calcium ions are stored as the salt, hydroxyapatite. Normally the calcium
ion concentration in the plasma is in equilibrium with the calcium in bone. In order to
maintain this equilibrium calcium ions must be able to travel from plasma to bone and
vice versa in response to hormonal signals. Osteosclerosis is a condition where the
calcium is trapped in bones of unusually high density. Osteoporosis, on the other hand, is
a disease state where calcium is mobilized from the bone. The resulting
demineralization results in a loss of integrity. Osteofluorosis, an accumulation of
fluoride in the bone, can result in osteosclerosis in normal individuals and can be curative
in individuals suffering from osteoporosis.
4.1.1 Osteosclerosis
This section reviews selected literature on the relationship between osteosclerosis
and fluoride in drinking water. Fluoride induces new bone formation which results in
densification and replacement of healthy. This fluoride induced bone is denser than the
bone it replaces due to the presence of fluoroapatite. The Surgeon General has indicated
that changes in bone density per se are not adverse health effects but that crippling
25
-------
fluorosis is an adverse health effect. 1
Skeletal fluorosis may progress in several stages. 2
• Slight radiological changes such as enlargement of trabeculae in
the lumbar spine; this can be further subdivided into two stages.
• Osteosclerosis in the pelvis and vertebral column.
• Increased density and blurring of contours of the pelvis, vertebral
column, ribs, and extremities.
• Greatly increased density of bone with irregular and blurred
contours, all bones being affected. Extremities are thickened
and considerable calcification of ligaments of the neck and
vertebral column occur.
Greater concentration of fluoride in the bone is apparently associated with more severe
stages of osteosclerosis.
Several studies provide evidence relating fluoride in drinking water and
osteoscierosis. Leone et al . (j955)3 compared the effects of fluoride exposure from
drinking water in 116 long-term residents of a high fluoride area (Bartlett, Texas; 8
mg/i) to 121 residents of a low fluoride area (Cameron, Texas; 0.4 mg/I). In the study,
the authors evaluated the effects of fluoride on bone changes as characterized by X-rays
taken in 1943 and 1953.
‘Cited in EPA, 50 Federal Register, 20166; May 14, 1985.
2 Described in EPA, “Final Draft for the Drinking Water Criteria Document on
Fluoride,” TR-540-61F, April, 1985.
3 Leone, N.C. Stevenson, T. Hilbish, M. Sosman, 1955. “A Roentgenologic Study of
a Human Population Exposed to High Fluoride Domestic Water.” Am. 3. Roentg., Radium
Therapy and Nuclear Medicine . 74:874-85.
26
-------
The authors concluded that fluoride-induced bone changes:
• occur in approximately 10 to 15 percent of those ex-
posed to 8 ppm of fluoride in drinking water;
• are not associated with physical findings other than
dental fluorosis;
• cannot definitely be ascribed to fluoride alone;
• may occasionally have a beneficial effect in adult
bone, as in counteracting the normal osteoporotic
changes in the aged;
• are not deleterious if the concentration in water is
less than 8 ppm.
Thus, this study suggests that significant osteoscierosis effects may occur in some
individuals exposed to concentrations of fluoride in drinking water of 8 mg/I or more.
In another study, Stevenson and Watson (1957)1 further evaluated the systemic
effects of excessive fluoride in drinking water. In this study, the authors reviewed
medical records over the period 1943 to 1953. A roentgenologic (x-ray) diagnosis of
fluoride osteosclerosis was recorded on 23 patients’ records out of a total of 170,000 x-
ray examinations of the spine and pelvis of patients living primarily in Texas and
Oklahoma.
The earliest bone changes were noted in the pelvis and vertebral column and
consisted of a slightly increased bone density and a “ground glass appearance.” Slight
roughening of the bones of the forearms or legs was also detected. Of special interest
was the finding of calcification of the sacrospinous and sacrotuberous ligaments in 15 of
the 23 patients. All of the reported bone changes occurred only among patients who
lived in areas with fluoride concentrations from 4 to 8 mg/I in drinking water. Each of
‘Stevenson C., A. Watson, 1957. “Fluoride Osteosclerosis” Am. 3. Roent., Radium
Therapy and Nuclear Medicine . 78: 13-18.
27
-------
these patients had lived in the same fluoride-bearing area throughout his life. Thus, the
study indicates a lowest observed effect level for osteosclerosis of approximately 4
mg/I. However, the authors concluded that the cases of fluoride osteoscierosis which
they saw caused no harmful effects.
Alhava et al . (1980)1 examined post mortem bone specimens from Kuopio, Finland
(where fluoridation started in 1959) and from other locations. The fluoride concentration
in Kuopio was approximately 0.97 ppm while the water samples from outside the Kuopio
area averaged less than 0.1 ppm. The length of exposure to fluoridation prior to death
was between 14-18 years. The findings of this study indicate that: a) fluoride
concentration in bones increases with the age of the person, b) fluoride concentrations in
bone are greater for the Kuopio population which is exposed to higher concentrations of
fluoride in their drinking water, and c) fluoride concentrations in the bones of the Kuopio
population are slightly higher for women than for men.
Studies in India have recorded some of the most severe forms of skeletal fluorosis.
An individual exposed to 9.5 mg/I of fluoride in his drinking water was described as
having irregular outgrowths on bony contours, irregular bone in the joint capsules and
interosseous membranes, heavier bones than normal, enlarged vertebrae and fusing of
some vertebrae, and other health effects in his bones. 2
Charen et al . (1979) examined a dose response curve showing fluoride in bone as a
function of fluoride in drinking water. A linear relationship was found suggesting that 1
ppm of fluoride in drinking water is associated with a fluoride concentration in bones of
about 2000 ppm and that 4 ppm of fluoride in drinking water is associated with a fluoride
concentration in bones of about 7000 ppm. This should be considered the result of
lifetime exposure to fluorides.
On the basis of the studies reviewed it is not possible to develop a dose-response
curve for the crippling effects of skeletal fluorosis. For the purposes of this analysis a
qualitative relationship is all that can be supported: concentrations of fluoride in
drinking water above 8 mg/I are associated with osteosclerosis and possibly with crippling
skeletal fluorosis.
‘Aihava, E.M., Olkkonen, i-I., Kauranen, P. and Kari, Tarja. (1980). “The Effect of
Drinking Water Fluoridation on the Fluoride Content, Strength and Mineral Density of
Human Bone” Acta Orthopedics Scandinavia 51: 4 13-420.
2 Cited in EPA, “Final Draft for the Drinking Water Criteria Document on
Fluoride,” TR-540-6 IF, April, 1985.
3 Charen, 3., Taves, D.R., Stamm, 1W. and Parkins, F.M., 1979. “Bone Fluoride
Concentrations Associated with Fluoridated Drinking Water: Calcified Tissue
International 27: 95—99
28
-------
4.1.2 Osteoporosis
Osteoporosis is an absolute reduction in bone mass resulting from demineralization
of bone. While the cause of osteoporosis is unknown, the increase in bone density
following administration of fluoride has been used therapeutically, but not without
controversy. This controversy concerns whether bone containing fluoroapatite is
physiologically equivalent to normal bone. Bernstein et al. ’ (1966), investigated the
relationship between drinking water fluoride content and the prevalence of osteoporosis
in North Dakota. Fluoride levels were between 4 and 5.8 ppm in the high concentration
area and between 0.15 and 0.3 ppm in the low concentration area. Individuals living in
the high fluoride areas had less incidence of osteoporosis and collapsed vertebrae than
those living in low fluoride areas. The difference in bone density was observed in
females over 55 years of age and in males 55 to 64. A lower incidence of collapsed
vertebrae was observed only in females.
4.2 Dental Effects
Fluoride has both positive and negative effects on teeth. This section briefly
describes these effects. More detail on adverse dental effects may be found in Appendix
A.
4.2.1 Adverse Dental Effects
Dental fluorosis has been classified into degrees of severity by Dean. 2 The Dean
Index, shown in Table 4-I, describes characteristics of teeth affected by fluorosis.
Moderate and severe forms of dental fluorosis result in brown stains affecting all enamel
surfaces and pitting in some teeth.
Several studies have indicated that higher concentrations of fluoride in drinking
‘Bernstein, D. et al. 1966, “Prevalence of Osteoporosis in High - and Low-Fluoride
Areas in North Dakota,” Journal of the American Medical Association , vol, 198, October
31, pp. 85-90.
2 Dean, H.T., 1942. “The Investigation of Physiological Effects by the
Epidemiological Method,” in F.R. Moulton, ed., Fluorine and Dental Health , American
Association for the Advancement of Science, Publication No. 19.
29
-------
water are associated with a greater proportion of children having moderate or severe
dental fluorosis.’ Very roughly, at least 10 to 20 percent of the children in a community
can be expected to suffer from moderate or severe dental fluorosis if the concentration
of fluoride in drinking water exceeds 2.5 to 3 mg/I. Appendix A provides a computation
of the number of children exposed to various fluoride concentrations who are expected to
suffer from moderate or severe fluorosis. Confidence intervals for these calculations are
provided in Appendix C.
An alternative to removing fluoride from drinking water is repair of teeth
exhibiting moderate or severe fluorosis. A drinking water MCL obviates the need to
undertake such repairs. The repair costs avoided by issuing an MCL are estimated in
Appendix B.
Dental fluorosis may also result in health impacts of a psychological nature. 2 The
disfiguration caused by more serious forms of fluorosis is likely to be disadvantageous to
children and adults, affecting self-confidence, social behavior, other people’s perceptions
of them, their influence on others, and other psychological characteristics. The exact
nature of these difficulties has not been studied, however.
4.2.2 Fluoride and Caries
One of the most important public health findings of the twentieth century has been
the effect of fluoride on preventing dental caries. Table 4-2 shows how the average
‘National Institute of Dental Research, “Prevalence of Dental Caries and Dental
Fluorosis in Areas with Optimal and Above-Optimal Concentrations of Fluoride in Their
Community Water Supplies”; and University of Texas, Health Science Center at San
Antoino Dental School, “A Clinical Study of Dental Effects in a Population Exposed to
Water Fluoride Levels in the ‘Critical Zone.”
2 R. Kleck, Chairperson, Review Panel on Psychological/Behavioral Effects of
Dental Fluorosis, Report to Office of Drinking Water, EPA, November 17, 1984.
30
-------
score for decayed, missing, and filled surfaces per tooth decreases as fluoride
concentration in drinking water increases, according to a recent study. ” 2 Analyses
conducted over the past fifty years have identified an optimum concentration of fluoride
in drinking water to promote dental health. This optimum depends on climate and water
intake but is about one part per million. (The reduction of cases of dental fluorosis due
to removal of fluoride from drinking water may be offset by increased dental caries as
fluoride concentrations drop. Appendix D provides an estimate of the costs of treating
additional caries at various MCLs.)
1 E. Collins, et al. “Analysis of Costs for the Treatment of Dental Fluorosis,”
Health Effects Research Laboratory, EPA, Cincinnati, Table 8 (p. 28).
2 Higher levels of fluoride are associated with lower scores for decayed,
missing, and filled surfaces (DMFS) per tooth. However, at concentrations of
fluoride above about 4 mg/I, higher scores for decayed, missing, and filled
surfaces and higher caries frequency have been observed. (W. Butler, V. Segretto,
and E. Collins, 1985, “Dental Mottling and DMFS Among School-Aged Children in
Texas;” and B. Forsman, 1974, “Dental Fluorosis and Caries In High Fluoride
Districts in Sweden,” Community Dent. Oral Epidemiol . 2: 132-148).
As the fluoride concentration in drinking water increases, the proportion of the
population with moderate or severe fluorosis increases. Moreover, the DMFS
score is sometimes higher for children with moderate or severe fluorosis than f or
children with lesser degrees of fluorosis. (This relationship between DMFS and
severity of fluorosis has not been found consistently at all age groups of children,
however.) It is thought, but has not been conclusively demonstrated, that the
pitting effects of moderate and severe dental fluorosis lead to the higher DMFS
scores observed at higher concentrations of fluoride in drinking water. Hence, at
higher fluoride levels, the incidence of caries may increase.
31
-------
Table 4-I
DEAN’S FLUOROSIS INDEX
Category Weight Description
Normal Enamel 0.0 Normal.
Questionable 0.5 Normal translucency varied by a few
white flecks or white spots.
Very Mild 1.0 Small, opaque, paper-white areas
scattered over the teeth, involving less than 25
percent of the surface. Tips of cusps of
bicuspid and second molars are commonly
affected.
Mild 2.0 White opaque areas are more extensive but do
not involve more than 50 percent of the
surface.
Moderate 3.0 All enamel surfaces are involved and
brown stain is a frequent disfiguring feature.
Severe 4.0 All enamel surfaces are involved with pitting
in many teeth. Brown stains are widespread
and teeth often present a corroded - like
appearance.
Source: H.T. Dean, “The Investigation of Physiological Effects by The Epidemiological
Method.” In P.R. Moulton, ed., Fluorine and Dental Health , American Association for the
Advancement of Science, Publication No. 19
32
-------
Table 4-2
Relationship Between Community Fluoride Concentrations
and Dental Caries -- Children 7 to 12 Years of Age
Average Score for
Community Fluoride Decayed, Missing and
Concentration Filled Surfaces per Tooth
0-0.9 mg/I .38
1.0-1.9mg/i .24
2.0-2.9 mg/i .16
3.0-3.9 mg/I .18
4.0-4.9 mg/I .13
Source: E. Collins, et al., “Analysis of Costs For the Treatment of Dental
Fluorosis,” Health Effects Research Laboratory, U.S. Environmental Protection
Agency, Cincinnati, Table 8 (p. 28).
33
-------
Chapter 5
ASSESSMENT OF COSTS
This chapter analyzes the costs of removing fluoride from drinking water. The
technologies for removing fluoride are discussed in Section 5.1. Section 5.2 addresses the
probability of utilities’ selecting these various removal measures. Computations of costs
to the nation and costs facing individual affected utilities and their customers are
presented in Section 5.3, followii g a discussion of the assumptions used to estimate
these costs. Section 5.4 presents an analysis of the costs of monitoring requirements for
fluoride.
5.1 Technologies for Reducing Fluoride
EPA has identified seven technologies for reducing fluoride, some of which have
several variations.’ These are:
• Activated alumina in a centralized treatment plant
• Reverse osmosis in a centralized treatment plant
• Optimized lime softening at an existing softening plant
• Bottled Water
- Used by all households in the community
- Used by households with small children (30% of households)
• Point of use treatment
‘EPA, “Technologies and Costs for the Removal of Fluoride from Potable Water
Supplies,” prepared by V.3. Ciccone and Associates, May 25, 1985.
34
-------
- With activated alumina
- With reverse osmosis
• Regionalization (i.e. interconnection with a nearby water
system which does not exceed the fluoride MCL).
• Alternate Source (i.e. tapping a new source of water which
does not exceed the fluoride MCL).
The point of use, regionalization, and bottled water alternatives apply only to
systems serving fewer than 10,000 persons. Centralized treatment is treatment which
occurs at the water system treatment plant rather than at each participating dwelling
unit.
5.2 Probabilities of Selecting Removal Measures 1
A water utility finding fluoride in its water supply in excess of a standard could
select any of several methods of treatment. EPA considered the following factors in
estimating the probabilities of a utility selecting a treatment:
• system size;
• capital and operating and maintenance costs;
• effectiveness in removing fluorides;
• the population at risk (all households or those with small
children); and
• likelihood of other inorganic contaminants being present along
with fluorides.
On the basis of these considerations, probabilities for selecting treatments were
estimated (assuming the MCL were exceeded) as follows:
‘The basic methodology used in this section was first developed in EPA, “Water
Utility Financing Study,” prepared for the Office of Drinking Water by Temple, Barker,
and Sloane (n.cf.).
35
-------
• Systems serving fewer than 1,000 people
-- 85% select point-of-use (activated alumina) for 30%
of the households (those households with small children)
-- 5% select point-of-use (reverse osmosis) for 30% of
the households (those households with small children)
-- 10% obtain a variance
• Systems serving 1,001 to 10,000 people
-- 90% select central activated alumina
-- 5% select central reverse osmosis
-- 5% obtain a variance
• Systems serving more than 10,000 people
-- 90% select central activated alumina
-- 10% select central reverse osmosis
The decision to use reverse osmosis as opposed to activated alumina is likely to
depend on the presence of other contaminants in the water supply. Systems with several
inorganics, including fluoride, would probably select reverse osmosis whereas systems
whose only inorganic contaminant is fluoride would probably select activated alumina.
5.3 Computation of Costs
5.3.1 Assumptions
Several sets of assumptions were adopted to compute the costs of removing
fluoride from drinking water. Costs are computed for the nation as a whole and for
individual affected utilities and their customers.
The first set of assumptions defines the standards to which water is to be treated .
For this calculation, MCLs of 1 mg/I, 2 mg/I, 3 mg/I, and 4 mg/I are assumed. This will
provide a range of costs on which judgements about standards can be made.
The second set of assumptions pertains to the presumed behavior of utilities under
different regulatory and rionregulatory options. In particular:
36
-------
• For promulgating an MCL as a primary standard, it is assumed
that the regulation would be enforced and that all systems
exceeding the MCL would treat their water to meet the
MCL. However, some systems can be expected to be granted a
variance.
• For a secondary standard of 2 mg/I or for a notification
requirement, it is assumed that no additional systems would
adopt fluoride removal measures. Systems failing to comply
with the current primary standard would probably not
undertake the cost of removing fluorides if it were no longer
believed that fluorides present a threat to health as a
secondary standard would imply.
• For a monitoring and notification requirement, it is unlikely
that any additional systems would act to remove fluorides
from drinking water. This requirement is already in place.
(4OCFR 141.32).
To confirm that these assumptions are reasonable, telephone contacts were made with
EPA Regional Offices in Regions IV (Atlanta) and VI (Dallas). Apparently no systems
have acted to remove fluorides in response to the existing MCL in Region VI but a few
have done so in Region IV. A monitoring and notification requirement was thought by the
Regional Offices to be unlikely to induce systems to remove fluorides. A secondary
standard of 2 mg/I would probably have little effect although both Regions indicated
that the states can enforce these standards. Publicity on fluoride contamination in
connection with a secondary standard may induce a few systems to remove fluorides.
However, because this is likely to be a small number of systems, the costs are not
analyzed here and it is assumed that a secondary standard would not lead to fluoride
removal actions.
The third set of assumptions is concerned with the characteristics of the treatment
technologies adopted by utilities to remove fluorides. EPA’s cost and technologies
studies to date’ have been prepared only for reducing an influent concentration of
fluoride of 3.2 mg/I to a standard of 2 mg/I. The costs based on these assumptions are
applied to all influent and effluent combinations in the calculations presented below,
recognizing that these costs may have to be revised as EPA prepares additional cost
1 EPA, “Technologies and Costs for the Removal of Fluoride from Potable Drinking
Water Supplies,” prepared by V.3. Ciccone and Associates, May 25, 1984. This report was
updated on September 10, 1985, but the costs did not change.
37
-------
information.
The specific features of the treatment technologies expected to be adopted are as
follows:
• For centralized activated alumina, 60% of the water is treated
and blended with 40% of the raw water to reach the MCL.
• For centralized reverse osmosis, 50% of the water is treated
and blended with the remaining raw water.
• For point of use treatment all households will receive fluoride
removal equipment.
Recall from Section 5.2, that the other possible treatments are not likely to be selected.
5.3.2 Cost of Alternatives
Two types of cost are presented. One is the cost to society of investing in fluoride
removal measures. This is the opportunity cost to the nation of purchasing and operating
fluoride treatments. Constant 1982 dollars are used. A real social discount rate of 3%
over a time horizon of 20 years is used to compute the present value of costs. 1 The real
social discount rate reflects the opportunity cost of a risk-free investment in constant
dollars.
The other perspective on costs is that of individual utilities. They must pay a
market rate of interest and so prevailing nominal discount rates appropriate to various
system size categories are used. EPA’s Financing Needs Model estimates water rates and
the proportion of systems satisfying certain financial criteria for well managed systems
adopting various water treatments.
All systems must currently monitor for fluorides and no new monitoring costs would
be incurred as a result of the regulatory alternatives unless the monitoring schedules
were changed (except that dropping all standards would obviate the need for even
monitoring for fluorides). Because the monitoring requirement is now in effect and no
new schedule for monitoring is proposed, no costs for monitoring are included in the
costing analysis.
‘This discount rate represents a long run historical average, not the higher real
rates experienced in the 1980s. Real discount rates exclude the effects of inflation and
are consistent with the use of constant dollars.
38
-------
Table 5-I shows the capital and operating and maintenance costs, the present value
of fluoride removal costs, and the annualized value of fluoride removal costs to society
for all systems which would act to remove fluoride assuming an MCL were set.’ The
results are shown by system size category. Table 5-2 shows these costs to society per
system by system size category. Only systems exceeding the MCL are included in the
calculations leading to Table 5-2.
The cost to society of fluoride removal measures drops off quickly as the MCL is
relaxed. This is because the number of systems exceeding the MCL declines rapidly as
the MCL increased. Per system costs are about the same at each MCL. 2
The cost estimates shown in Tables 5-I and 5-2 are based on the assumption that
systems serving fewer than 1000 people would install point of use equipment in 30% of
the households. This assumption is consistent with the goal of protecting children from
dental fluorosis. If instead the goal of fluoride regulation is to protect all individuals
from skeletal osteoscierosis, the analysis must reflect 100 percent protection. In
particular, systems serving fewer than 1000 persons would install centralized treatment.
Under this assumption, costs for the two smallest system size categories at the
MCL level of four mg/I have been estimated as follows. For the category of systems
serving populations between 25 and 500, national capital costs would be $3.8 million and
O & M costs $0.5 million, for a total annual cost of $0.7 million and a present value of
$10.5 million. Social costs per system for the 25-500 size catory are $18,000 capital cost
and $2,500 annual 0 & M cost, giving a present value of $51,700.
For the category of systems serving populations between 501 and 3300, national
capital costs would be $4.7 million and 0 & M costs $0.6 million, for a total annual cost
of $0.9 million and a present value of $14 million. Social costs per system for the 501-
3300 category are $92,200 capital cost and $11,800 annual 0 & M cost, giving a present
value of $51,700.
These estimates imply changes in the total social costs to the nation as follow:
‘These costs consider only the costs of the technologies to remove fluoride.
Reduction of fluoride in some drinking water systems will also cause the incidence of
caries to increase. Appendix D provides an estimate of the costs of repairing these
additional caries.
2 Differences are largely due to rounding error and the inclusion of very large
systems in the size category 50,000 for an MCL of I but not for an MCL of 2
since there are no very large systems with fluoride concentrations above 2 mg/I.
39
-------
Table 5—1
Social Costs to the Nation of Fluoride Removal
(millions of 1982 dollars)
MCL System Size Category (population served )
330 1—
25—500 501—3300 50,000 50,000+ TOTAL
MCL = 1 mg/i
• capital costs 19.7 66.8 285.9 595.3 967.6
• annual 0&M Costs 5.4 10.9 50.6 201.2 268.0
• present value* 100.0 229.0 1038.7 3588.6 4954.8
• annualized cost* 6.7 15.4 69.8 241.2 333.0
MCL = 2 mg/i
• capital costs 5.9 16.7 29.9 2.0 54.6
• annual 0&M costs 1.7 2.9 5.0 0.5 10.2
• present value* 31.2 59.8 104.3 9.4 206.4
• annualized cost* 2.1 4.0 7.0 0.6 13.9
MCL = 3 mg/i
• capital costs 2.2 8.6 12.5 0.0 23.4
• annual O&M costs 0.6 1.4 2.0 0.0 4.2
• present value* 11.1 29.4 42.3 0.0 85.9
• annualized cost* 0.7 2.0 2.8 0.0 5.8
MCL = 4 mg/i
• capital costs 1.3 3.8 6.8 0.0 11.9
• annual 0&M costs 0.4 0.6 1.1 0.0 2.1
• present value* 7.3 12.7 23.2 0.0 43.1
• annualized cost* 0.5 0.9 1.6 0.0 2.9
*Computed with a 3% real discount rate over 20 years.
Note: totals may not add due to rounding
40
-------
Table 5—2
Social Costs Per System For Fluoride Removal
• (thousands of 1982 dollars)
MCL System Size Category (population served )
3301—
25—500 501—3300 50,000 50,000+ Average
System
MCL = 1 mg/i
• capital costs 6.1. 78.1 693.9 9929.8 213.1
• annual O&M costs 1.7 12.7 122.8 3357.7 59.0
• present value* 31.4 267.0 2520.9 59883.9 1090.9
MCL = 2 mg/i
• capital costs 6.2 65.2 622.9 3041.4 43.4
• annual 0&M costs 1.8 11.3 104.2 760.3 8.1
• present value* 33.0 233.3 2173.1 14352.7 163.9
MCL = 3 mg/i
• capital costs 6.6 72.9 625.0 NA 49.7
• annual 0&M costs 1.8 11.9 100.0 NA 8.9
• present value* 33.4 249.9 2112.7 NA 182.1
MCL = 4 mg/i
• capital costs 6.4 74.5 581.2 NA 44.8
• annual 0&M costs 2.0 11.8 94.0 NA 7.9
• present value* 36.2 250.1 1979.7 NA 162.3
*Computed with a 3% real discount rate over 20 years.
NA means not applicable
NOTE: Systems using purchased water are excluded to avoid double counting
since their water would be treated by their suppliers.
41
-------
capital cost increases from $11.9 to $15.3 million, annual 0 & M costs rise from $2.1 to
$2.2 million, annual total cost increases from $2.9 to $3.2 million and present value rises
from $43.1 to $47.7 million. This increase is about 10 percent of previous estimates, and
falls within the expected error of the estimates.
Impacts of treatments as perceived by water utilities are shown in Table 5_3•1
These impacts are expressed as water rates that would have to be charged by utilities
under the current status (no fluoride treatment) and assuming they would adopt a fluoride
removal measure. Water rates are computed by the Financing Needs Model assuming
that a predetermined earnings or operating surplus level must be met (based on a debt
service ratio, pretax interest coverage, or return on equity) and that water usage would
not go down if costs and prices went up. The impact of fluoride treatment is the
difference in water rates between the current status rate and the rate with the
treatment in place. For example, in publicly owned systems serving over 50,000 people,
the current status rate is $0.86 per thousand gallons and the rate with central activated
alumina is $1.07 per thousand gallons. Therefore the impact on a system choosing
central activated alumina is $1.07 minus $0.86, or $0.21 per thousand gallons.
The impact of higher water rates resulting from fluoride removal is shown in Table
5-4 in terms of increases in annual household water bills. These increases assume a
typical household purchases 100,000 gallons of water per year. EPA’s Financing Needs
Model provides for no adjustment for decreased use of water as water prices increase. 2
Another way to view utility level impacts is to estimate the percentage of systems
which could pass the following three tests simultaneously.
O a rate increase less than $1.00 per thousand gallons;
• a ratio of new treatment capital costs to assets less than 1.0;
and
‘Recall that EPA has not yet studied the costs under alternative influent and
effluent concentrations. Thus Table 5-3 and others based on the Financing Needs Model
pertain to any influent-effluent combination.
2 The increase in water bills is simply the median water rate with the fluoride
treatment minus the current status rate in Table 5-3; this is then multiplied by 100.
42
-------
Fluoride
Removal Measure ______
Current Status
Central Activated
Alumina
Central Reverse
Osmosis
Point of Use (30%
of Households)
*Populat ion served.
Source: Financing Needs Model
Public Ownership
330 1—
______ _________ 50,000
1.00
1.27
501—3300
1.47
2.19
Table 5—3
Utility Cost Impacts——
Median Water Rate
($ per thousand gallons, 1982 dollars)
System Size* and Ownership Category
_______________________________________ Private_Ownership
3301—
______ ________ ______ _______ ______ ________ 50,000 _______
1 • 66
1.96
25—500
3 • 80
4.75
50,000+
0.86
1 • 07
7.61
2 5—500
7.50
8 • 81
501—3300
3 • 54
4.18
2.42
1.98
4.15
1.98
2.59
50, 000+
1 .35
1.53
2.29
11.17 6.50
o Activated Alumina 4.32
o Reverse Osmosis 5.15
2.99
8.21
9 • 04
3.94
4 • 64
-------
Table 5—4
Increase in Annual Household Water Bill Attributable to Fluoride Removal*
(1982 dollars)
System Size** and Ownership Category
Public Ownership Private Ownership
Fluoride
Removal 3301— 3301—
Measure 25—500 501—3300 50,000 50,000+ 25 5OO 501—3300 50,000 50,000+
Central Activated $ 95 $ 72 $ 27 $ 21 $ 131 $ 64 $ 30 $ 18
Alumina
Central Reverse $381 $328 $142 $112 $ 367 $296 $133 $ 94
Osmosis
Point of Use (30% of
households)
o Activated Alumina 52 51 71 40
o Reverse Osmosis 135 112 154 110
* Assuming 100,000 gallons of water used per household before and after fluoride treatment
** Population served.
-------
• a total cost (rate) less than $3.00 per thousand gallons.
The Financing Needs Model makes such estimates and the results are shown in
Table 5-5. The impact of fluoride treatment is the difference between the percentage of
systems meeting this test under current status (with no new fluoride treatment) and the
percentage of systems meeting the test with a fluoride removal measure in place. For
example, 37.0% of public systems serving 25 to 500 people pass the financial tests before
fluoride treatment (i.e., with the current status) but only 18.5% pass the financial tests
after fluoride treatment with central activated alumina. Thus the impact is that 37.0%
minus 18.5% equals 18.5% of these systems would be significantly affected by a fluoride
regulation.
From the utility perspective, the following conclusions can be drawn:
• For systems serving fewer than 3,300 people, point of use
activated alumina treatment is likely to be least costly
assuming that it could be used under conditions specific to
such systems.
• Central activated alumina is least costly for larger systems
and relatively inexpensive for systems serving fewer than
3,300 people.
• Reverse osmosis is generally more expensive but it may be
selected by systems for which activated alumina is ineffective
or inappropriate.
5.3.3 Conclusions
For purposes of standard-setting, EPA judges the feasibility and cost of treatments
considering the availability of treatments for larger, well-run, metropolitan and regional
systems. In the case of fluoride, large systems can meet any MCL under consideration at
a cost of $18 to $21 per household per year using centralized activated alumina. If
appropriate, optimized lime softening could be used by large systems at a cost of $9 per
household per year. Centralized reverse osmosis could also be used by large systems at a
cost of about $100 per household per year. The costs of centralized activated alumina
and optimized lime softening are clearly affordable. Therefore, feasibility and cost are
not, in general, constraints in setting an MCL. However, few large systems are likely to
exceed a fluoride MCL.
45
-------
Table 5—5
Utility COBt Impacts
Percent of Systems Satisfying Financial Criteria*
System Size** and Ownership Category
Public Ownership Private Ownership
3301— 3301—
25—500 501—3300 50,000 50,000+ 25—500 501—3300 50,000 50,000+
Current Status 37.0 70.3 83.3 83.9 20.8 42.9 78.4 73.2
Central Activated 18.5 59.5 78.2 78.8 83 32.1 64.7 68.3
Alumina
Central Reverse 0.0 0.0 6.4 26.3 0.0 0.0 59 51.2
Osmosis
Point of Use (30%
of Households)
o Activated Alumina 29.6 64.9 8.3 39.3
o Reverse Osmosis 14.8 21.6 0.0 17.9
*Rate increase less than $1.00 per thousand gallons, new capital cost to asset ratio less than 1.0, and total cost
(rate) less than $3.00 per thousand gallons.
**Population served
Source: Financing Needs Model.
-------
Families served by small systems that are likely to violate an MCL face
significantly larger costs. Moreover, some systems may not be able, as a practical
matter, to comply. In the case of fluoride, the engineering and cost analysis shows that
there are three treatment technologies generally available (taking cost into
consideration) even to the smallest systems. These are centralized activited alumina,
targetted point-of-use activated alumina, and targetted point-of-use reverse osmosis,
where the targetting is to households having small children. Annual per household cost
ranges from $52 to $154 for the smallest systems. Assuming that the Administrator
makes a finding that such costs are affordable these would be designated as generally
available technologies for the smallest systems.’
5.4 Monitoring Requirements
Under current regulations (40 CFR 141.23) community groundwater systems must
monitor for fluoride every three years and community water systems using any surface
water must monitor for fluoride annually. If the MCL is exceeded three additional
analyses must be completed within one month. If the average concentration from these
four tests exceeds the MCL, a state-designated monitoring frequency is to be followed
until the concentration is less than the MCL on two successive samples.
There are approximately 48,854 groundwater systems and 10,958 surface water
systems in the U.S. The number of analyses for fluoride each year is therefore about
27,000. This is “baseline” sampling. In addition, those surface water systems exceeding
the MCL would have to take at least three additional samples per year and at least one
third of the groundwater systems exceeding the MCL would have to take at least three
additional samples per year.
Considering this baseline plus the occurrence of fluorides exceeding various MCLs
(Table 2-1), the minimum number of samples per year would be:
• For “baseline” sampling 27,000 plus
• For an MCL of I mg/I 6,400 additional samples, or
• For an MCL of 2 mg/I 1,400 additional samples, or
• For an MCL of 3 mg/I 500 additional samples, or
• For an MCL. of 4 mg/I 300 additional samples.
‘Exemptions would be available where deemed appropriate.
47
-------
The laboratory cost per analysis is about $6 using the electrode method. This
method is the most commonly used one. Thus baseline laboratory costs would be about
$162,000 per year. This would be the cost if no systems violated the MCL. At an MCL
of 2 mg/I the costs for additional sampling by systems now exceeding the MCL would add
at least $8,400 per year to this figure.
Under proposed regulations described in 40 CFR 141.23, systems with fluoride
concentrations less than 2 mg/I would be required to monitor one time in each ten year
period. Ground water systems that are likely to exceed 2 mg/I would be required to
monitor every three years. Surface water systems that are likely to exceed 2 mg/I
would be required to monitor every year (Pages 50 - 57, proposed notice.)
Ground-water Surface-water
Systems Systems
likely to
exceed 2 mg/I 1,324 22
not likely to
exceed 2 mg/I 47,530 10,936
Under these regulations, 6,310 samples per year would be required at a cost of
$37,860 per year.’
It should be noted that these are the current analytical requirements. Under the
alternatives examined here, no additional monitoring costs would be imposed.
‘Systems < 2 mg/I:
Ground-water (.1 x 47,530) 4,753 samples/year
Surface-water(.1 x 10,936) 1,094
Total 5,847
Systems > 2 mg/I:
Ground-water (.33 x 1,324) 441 samples/year
Surface-water (I x 22) 22
.463
48
-------
Chapter 6
REGULATORY FLEXIBILITY ANALYSIS AND PAPERWORK ANALYSIS
6.1 Regulatory Flexibility Analysis
The Regulatory Flexibility Act (RFA) was enacted on September 19, 1980 to
require all agencies to explicitly consider small entities in their regulatory design and
implementation process. More specifically, regulatory agencies are to try to minimize
the disproportionate burden that falls on small entities. The three primary objectives of
the RFA are to:
• increase agency awareness of their regulatory impact on small
entities;
• compel agencies to explicitly analyze, explain and publish their
regulatory impacts on small entities; and
• encourage agencies to provide regulatory relief to small entities,
while accomplishing their statutory mandates.
These objectives are accomplished through the requirements of regulatory
flexibility analyses for all existing and proposed regulations. If a regulation does not
have a “significant” impact on a “substantial” number of small entities, then the
regulatory flexibility analysis will consist of a certification to that effect.
Prior to conducting a regulatory flexibility analysis, the agency must define a small
entity. The RFA defines small entities to include small businesses, organizations, and
governments (PL 96-354, Section 601 [ 6]). Small businesses are defined as any business
which is independently owned and operated and not dominant in its field (15 U.S. Code,
section 632). Small organizations are defined as any non-profit enterprise which is
independently owned and operated and is not dominant in its field. Lastly, small
governments are defined as those city, county, town, township, village, school district or
special district governments serving a population of less than 50,000 persons (Regulatory
Flexibility Act, PL 96-354, Sections 601 [ 4] and 601 [ 5]). Some public water systems are
publicly owned, some are privately owned, and some are ancillary to other enterprises
such as hospitals or mobile home parks. According to EPA’s 1980 Survey of Operating
49
-------
and Financial Characteristics of Community Water Systems’ there are 26,424 publicly
owned water systems of which 98% serve fewer than 50,000 persons. See Table 6-1.
For the case of privately owned water systems, firms primarily engaged in water
supply (SIC 4941) are small businesses under the Small Business Administration criteria if
their annual receipts are less than $3.5 million ( Federal Register , vol. 49, no. 28 (Feb. 9,
1984], p. 5035). Applying the Consumer Price Index for water and sewerage maintenance
(February 1980/ February 1984) to this figure, the cut-off for a small water utility would
be $2.4 million in 1980 dollars. EPA’s 1980 Survey of Operating and Financial
Characteristics of Community Water Systems indicates that a population of 50,000
persons is about the cut-off for revenues of $2.4 million. For privately owned systems
serving 25,000 to 50,000 persons, revenues in 1980 averaged $1,966,900, and for privately
owned systems serving 50,001 to 75,000 persons, revenues in 1980 averaged $3,156,000.
The number of privately owned small water systems is shown in Table 6-1.
There is some question as to whether privately owned water systems serving fewer
than 50,000 persons qualify as small businesses. Some privately owned systems are not
independently owned as there are many holding companies in the water supply business;
in particular American, Consumers, Continental, and General Water Works Companies
own a number of water utilities. In addition, nearly every privately owned water
company is a natural monopoly in its market area, thereby bringing into question its
dominance in its field. However, the Small Business Administration considers only a
national basis for each industry, not a local basis ( Federal Register , Vol. 49, No. 28, Feb.
9, 1984, p. 5039), so most water utilities would not be dominant in the entire U.S.
marketplace.
All ancillary systems serve fewer than 500 people according to EPA’s 1980 survey.
These might be small entities although the main part of the activity to which water is
supplied could be too large to qualify as a small business or may be owned by a larger
organization. Moreover, some of these ancillary systems are not businesses but rather
hospitals or schools. There are 16,907 ancillary water systems; see Table 6-1. It is not
possible to disaggregate these systems by size of the parent organization, however.
The number of water systems affected by a fluoride MCL is shown in Table 2-1.
EPA’s guidelines on compliance with the Regulatory Flexibility Act (April 12, 1983)
‘EPA, “Survey of Operating and Financial Characteristics of Community Water
Systems,” prepared by Temple, Barker and Sloane, 1982.
50
-------
Table 6-1
Number of Water Systems by Population Served
Population Served
330 1-
25-500 501-3300 50,000 50,000+ TOTAL
Publicly Owned 8,932 11,544 5,455 493 26,424
Privately Owned 12,591 2,239 802 108 15,740
Ancillary 16,907 0 0 0 16,907
Total 38,430 13,783 6,257 601 59,071
Source: EPA, Survey of Operating and Financial Characteristics of Community Water
Systems, 1982.
51
-------
indicate that, in general, a substantial number of small entities is more than 20% of
these entities. Tables 6-1 and 2-1 show that fewer than 20% of the systems would be
affected by an MCL of I mg/I. Even fewer would be affected by an MCL of 2 mg/I or
more. By the 20% rule fluoride regulations would not affect a substantial number of
small water utilities.
Cost impacts on small water systems are summarized in Tables 5-3 and 5-5. Table
5-5 suggests the magnitude of significant cost impacts. The difference in the percentage
of systems passing the financial tests with and without fluoride treatment indicates that
some small systems may face financial difficulties as a result of fluoride regulations.
For example, in publicly owned systems serving fewer than 500 people, nearly all will
select point-of-use activated alumina, and about 7 percent are likely to find this
expensive. For privately owned systems serving fewer than 500 people, nearly all will
select point-of-use activated alumina and about 12.5 percent are likely to find this
expensive.
6.2 Paperwork Analysis
Among the purposes of the Paperwork Reduction Act (Public Law 96-511; 94 STAT
2812) are:
• Minimization of the Federal paperwork burden for individuals,
small businesses, state and local governments, and other persons;
and
• Minimization of the cost to the federal government of collecting,
maintaining, using and disseminating information.
Water utilities and state agencies will be required to maintain records on
monitoring for fluoride and this is likely to be the largest component of paperwork
associated with fluoride regulations. The Paperwork Reduction Act is intended to
minimize the burden imposed on utilities and states while meeting the need to protect
the public health and welfare under the Safe Drinking Water Act.
EPA is required to submit to the Office of Management and Budget proposed
information collection requests. EPA must also submit a copy of a proposed rule
containing an information collection requirement no later than publication of a notice of
proposed rulemaking in the Federal Register . In addition, when a final rule is published
In the Federal Register , EPA must explain how any collection of information
requirements respond to public comments. The Office of Management and Budget
52
-------
determines the necessity and practical utility of the information being requested and if
approval of the request is made, 0MB will issue a control number.
In order to determine whether a specific water system exceeds a maximum
contaminant level for fluoride, each water system must monitor its water. EPA, the
states, water utilities, and the public would use monitoring information to determine
whether fluoride exceeds the MCL and to help determine appropriate courses of action if
it does.
The monitoring and notification requirements of a fluoride regulation are likely to
constitute the paperwork burden of water utilities. As indicated in Section 5.4, the
additional monitoring requirements of a fluoride regulation are negligible since fluoride
is now regulated and monitoring is already required. Data collection on fluoride is
already authorized by 0MB. The number of water samples required per year at an MCL
of 2 mg/I would be 28,400. The greatest respondent burden would fall on surface water
systems exceeding an MCL. They would have to take at least four samples per year. A
certified or state laboratory would conduct the analysis for fluorides.
53
-------
APPENDIX A
COMPUTATION OF NUMBER OF CASES OF SEVERE
AND MODERATE DENTAL FLUOROSIS AVOIDED
This appendix provides numerical estimates of the number of cases of severe and
moderate dental fluorosis avoided if various MCLs are adopted. The MCLs examined are
I mg/I, 2 mg/I, 3 mg/i, and 4 mg/I. These provide a range of standards from which to
judge some of the benefits of reducing fluoride.
An ad hoc committee headed by the Chief Dental Officer of the U.S. Public Health
Service has stated that “No sound evidence exists which shows that drinking water with
the various concentrations of fluoride found naturally in public water supplies in the U.S.
has any adverse effect on dental health as measured by loss of function and tooth
mortality.”’ The Surgeon General also stated that “...I encourage communities having
water supplies with fluoride concentrations of over two times optimum to provide
children up to age nine with water of optimum fluoride concentration to minimize the
risk of their developing aesthetically objectionable dental fluorosis.” 2 The data provided
in this appendix suggest the magnitude of cases of objectionable dental fluorosis avoided
at various MCLs.
It has been conclusively demonstrated that fluoride in drinking water causes dental
fluorosis in some portion of the population. In moderate and severe cases, tooth surfaces
have brown stains or pitting or both.
The Dean Index, developed in the late 1930s and early 1940s, is the standard system
used to classify the degree of fluorosis. 3 The index contains six categories ranging from
“normal” and “questionable” to “severe”. See Table 4-1. This analysis is limited to
‘Quoted in EPA, 50 Federal Register, 20166; May 14, 1985.
3 Dean, H.T., 1942, “The Investigation of Physiological Effects by the
Epidemiological Method,” in F.R. Moulton, ed., Fluorine and Dental Health , American
54
-------
evaluation of the two most serious categories - “moderate” and “severe” - as they
represent actual destruction of the dental enamel.
Two studies provide data with a sufficiently wide range of fluoride concentrations
to permit estimation of dose-response relationships for fluorosis. 1 These relationships
are needed to project the number of cases of fluorosis avoided by reducing fluoride
concentrations. Both studies evaluated fluorosis in children since fluorosis is most likely
to occur before six years of age. In addition, both studies took into account the
concentration of fluoride in the drinking water of the towns in which the children lived.
In particular, the University of Texas study team evaluated the prevalence of fluorosis in
a total of 2,602 children in 16 Texas communities in which fluoride concentrations in
drinking water ranged from 0.3 to 4.3 times the dentally-defined optimal level; the NIDR
study evaluated 807 children in seven Illinois communities with fluoride concentrations
ranging from 1.0 to 4 times optimum. Although the optimum will vary from place to
place due to climatic differences, the optimum is about one part per million or slightly
less. Thus a community with water having a fluoride concentration of twice the dental
optimum would exhibit a fluoride concentration of about two parts per million. Tables
A-I and A-2 present the percentage distributions of children by degree of fluorosis by
level of fluoride in drinking water. Note the difference between the dose-response rates
for moderate fluorosis between the two studies.
The data from these studies can be used to determine the number of cases of
severe or moderate fluorosis avoided in children under six years of age if fluoride
concentrations were reduced. It is assumed that fluorosis occurs due to exposure from
birth to six years of age.
The empirical data are insufficient to model the form of the mathematical
relationship (e.g., linear, log-linear, etc.) between fluoride concentrations in drinking
water (dosage) and the proportion of children with fluorosis (response). This analysis
assumes that the relationship is linear, a reasonable assumption in the low dose range.
Association for the Advancement of Science, Publication No. 19.
‘National Institute of Dental Research, “Prevalence of Dental Caries and Dental
Fluorosis in Areas with Optimal and Above-Optimal Concentrations of Fluoride in Their
Community Water Supplies;” and University of Texas, Health Science Center at San
Antonio Dental School, “A Clinical Study of the Dental Effects in a Population Exposed
to Water Fluoride Levels in the ‘Critical Zone.”
55
-------
Table A—i
PERCENTAGE DISTRIBUTION OF MOTTLED ENAMEL SCORES
AMONG TEXAS CRIL.DREN
Water
Fluoride Z of children with
Study (Factor of Number Moderate Severe
Community Optimal) Examined Fluoroeis* Fluorosis*
New Braunfels 0.3 103 0.0 0.0
San Marcos 0.3 223 0.0 0.0
San Antonio 0.4 126 0.0 0.0
Kingsville 1.0 361 0.3 0.0
Alvin 1.3 211 0.9 0.0
Angleton 1.3 187 1.1 0.0
Kerrville 1.4 128 0.0 0.0
Alpine 2.3 23 13.0 0.0
Littlefield 2.3 109 14.7 0.0
Fort Stockton 2.5 301 3.3 0.0
Uillsboro 2.7 200 4.0 0.0
Monahans 2.7 170 13.5 0.0
Perryton 2.7 90 6.7 0.0
Abernathy 2.9 67 32.8 0.0
Gatesville 3.1 113 4.4 0.0
Taylor 4.3 190 31.1 0.5
*Each subject was classified on the basis of severest form of enamel mottling recorded for two or more teeth.
Classification made according to Dean (1942).
Source: University of Texas, Health Science Center, San Antonio Dental School, “A Clinical Study of the Dental
Effects in a Population Exposed to Water Fluoride Levels in the ‘Critical Zone’,” Table V.C.3.
56
-------
Table A—2
PERCENTAGE DISTRIBUTION OF MOTTLED ENAMEL SCORES AMONG ILLINOIS SCHOOL CHILDREN
Water Fluoride Percentage Distribution of Mottled Enamel Sources*
(Multiple ______________________________________________________________________________
Study of Optimal Number Very
Community Dental Level) Examined Normal Questionable Mild Mild Moderate Severe
(0.0) (0.5) (1.0) (2.0) (3.0) (4.0)
Kewanee 1 336 56.0 29.5 7.4 4.8 1.8 0.6
Monmouth 2 143 18.2 28.7 23.1 16.8 8.4 4 .9
Abingdon 3 192 22.9 26.0 15.1 19.8 7.8 8.3
Elmwood
Ipava 4 136 12.5 15.4 16.9 25.0 7.4 22.8
Bushnell
Table Grove
*Each subject was classified according to Dean’s Index on the basis of the most severe form of enamel mottling
recorded for two or more teeth.
Source: National Institute of Dental Research, “Prevalence of Dental Caries and Dental Fluorosis in Areas
with Optimal and Above Optimal Concentrations of Fluoride in their Community Water Supplies.”
-------
Scattergrams of the dosage and response variables within the range of data observed do
not support a non-linear relationship.
The calculation of the number of cases avoided annually was performed in several
steps. In the first step, the percentage of children having a given degree of fluorosis was
regressed on the concentration of fluoride in drinking water. Concentration of fluoride is
measured as a multiple of the optimal level. Since the University of Texas data
indicated a non-zero percentage of “severe” fluorosis in only one of the 16 groups of
children studied, no equation was derived for “severe” fluorosis using these data. The
regression equations and coefficients of correlation for the two sets of data are
presented in Table A-3.
In the second step, two sources of data were combined to estimate the total
population exposed to various concentrations of fluoride. The first data source was
EPA’s estimate of the number of systems by size category by fluoride concentration.’
To obtain the population served by each system, the average population by system size
category was obtained from EPA’s 1980 survey of community water systems. 2 The
average population was then multiplied by the number of systems in each size category
to obtain the total population exposed to a given fluoride concentration. The estimated
population exposed to fluoride by concentration category is shown in Table A-4.
In the third step, the regression equations in Table A-3 were used to calculate the
percentage of children with severe or moderate fluorosis at various concentrations of
fluoride in drinking water. Concentration levels of 1, 1.5, 2, 2.5, . . ., 8.5 mg/I were
utilized.
Data from the Bureau of Census indicate that 10.1 percent of the population is
under six years of age. 3 Therefore, in the fourth step, the population data presented in
Table A-4 were multiplied by .101 to derive the number of children in the age group birth
to six.
‘EPA report prepared under contract by JRB Associates, “Occurrence of Fluoride
in Drinking Water, Air and Food,” February 9, 1984.
2 EPA, “Survey of Operating and Financial Characteristics of Community Water
Systems,” 1982, Table 11-3.
3 U.S. Bureau of the Census, Current Population Reports. Preliminary Estimates of
the Population of the United States by Age, Sex, and Race: 1970 to 1981 . Series P-25,
No. 917.
58
-------
Table A-3
DOSE-RESPONSE RELATIONSHIPS FOR FLUORID IN DRINKING WATER
AND DENTAL FLUOROSIS’
Study
Degree o
Fluorosis University of Texas NIDR
Moderate y = -5.51+6.79x r=.73 y = 2.30+l.62x r=.68
Severe see note 3 y = -8.35+7.OOx r=.94
percentage of children with fluorosis of the degree of severity
indicated.
x = fluoride concentration in drinking water, expressed as the number
times the dental optimum concentration.
Analysis uses ordinary least squares regressions.
defined by Dean (1942).
3 Only one of the 16 exposure groups showed a non-zero percentage of children with
“severe” fluorosis. Therefore, no regression analysis was performed.
59
-------
Table A—A
ESTIMATED POPULATION EXPOSED TO FLUORIDES IN DRINKING WATER
Number of People Exposed at Concentration Indicated
(milligrams
Len iter)
I I
Population
Served by
Water System
Average
Population
(per system)
25 — 100
57
>1
1> - !
I
73,017
24,168
101 — 500
245
I
273,910
I 66,395
501 — 1000
782
I
210,358
65,688
1001 — 3300
1,819
I
783,989
78,217
3301 — 10,000
5,165
I
1,723,735
1138,360
10,001 — 25,000
16,935
1,066,905
50,805
25,001 — 50,000
37,157
2,192,263
74,314
50,001 — 75,000
62,830
I
753,960
I 62,830
75,001 —100,000
88,035
528,210
I 0
100,001 —500,000
209,950
I
9,867,650
I 0
500,001 —1 Million
106,830
I
4,240,980
I 0
1 Million +
2,342,736
I
7.028.208
I 0
I
>3—4
1
>4 .1
3,591
I
,751 I
19,600
29,890
25,024
18,168 I
52,751
I
41,837 I
34,590
I
57,650 I
16,935
I
33,870 I
37,157
I
0
0
I
Ol
0
01
0
I
01
0
I
01
I
I
I
I
I
TOTAL I 28,743,185 560,777
1189,648
1187,172
I
-------
In the fifth step the net reduction in cases of severe and moderate fluorosis was
calculated, assuming that the systems exceeding a given standard would reduce fluoride
levels to, but not below, the standard. The net reduction in cases is the total number of
cases that would occur in the absence of treatment minus the number of cases that would
occur at the level set by the standard. This calculation provides the net reduction in
cases within the entire six-year age category (ages 0 to 6). To obtain an annual average
rate of fluorosis avoided for this age category, the number of cases avoided for the
entire age category was divided by six. The results of this calculation are presented in
Table A-5.
As can be seen from Table A-5, the number of moderate cases of dental fluorosis
avoided per year is between 55 and 230 if the MCL is set at 4 mg/I. The number of
moderate cases of dental fluorosis avoided rises to between 4,486 and 18,806 per year if
the MCL is set at I mg/i. The large jump in the number of cases avoided from an MCL
of 2 mg/I to an MCL of I mg/I is due to the occurrence of a large number of systems
having a fluoride concentration between I and 2 mg/I.
Only the NIDR data provide information on severe cases of fluorosis avoided per
year; the University of Texas data indicate few severe cases of fluorosis at the observed
levels of fluoride concentration. Thus, one must be cautious about accepting the NIDR
results as applied to this problem. With this caveat in mind, the number of severe cases
of dental fluorosis avoided per year at a standard of 4 mg/I is 237. This rises to 12,641
severe cases avoided per year at a standard of 1 mg/I. The large jump in the number of
cases avoided from an MCL of 2 mg/I to an MCL of 1 mg/I is due to the occurrence of a
large number of systems having a fluoride concentration between I and 2 mg/I.
Confidence intervals for these estimates are derived in Appendix C. These
confidence intervals consider only errors in the dose-response curve.
61
-------
Table A-5
Number of Cases of Severe and Moderate Dental Fluorosis Avoided Annually
at Alternate MCLS for Fluoride in Drinking Water
MCL
Number of Cases of Fluorosis
Avoided per Year 4 mg/I 3 mg/I 2 mg/I I mg/i
Moderate’
• Systems serving 25- 12-55 25-109 54-228 142- 597
500 people
• Systems serving 501- 17-71 44-185 102-426 315-1316
3300 people
• Systems serving 3301- 25-105 62-261 147-617 948-3972
50,000 people
• Systems serving over 0 0 8- 36 3083-12920
50,00 people
• All systems 2 55-230 132-554 311-1307 4486-18806
Severe 3
• Systems serving 25- 56 112 234 502
500 people
• Systems serving 501- 72 188 437 1065
3300 people
• Systems serving 107 267 635 2860
3301-50,000 people
• Systems serving over 0 0 37 8211
50,000 people
• All systems 2 237 570 1346 12641
‘Low value is from NIDR data, high value is from University of Texas data.
2 Total may not add due to rounding.
3 Values are from NIDR data.
62
-------
APPENDIX B
COSTS OF REPAIR OF DENTAL FLUOROSIS
If fluoride MCLs are not promulgated, then the effects of moderate and severe
dental fluorosis can be remedied by cosmetic and functional repairs to the teeth of
affected children. The children who would have avoided moderate or severe fluorosis if
an MCL were promulgated (Table A-5) would thus require tooth repairs to offset the lack
of an MCL. Table B-I shows the average cost for cosmetic and functional repairs for
various degrees of fluorosis’. These data are from dentists’ evaluations of repair costs
for minimum care (assuming the patient has a limited budget) and optimum care
(assuming no financial restrictions on the patient) 2 . Fifty-five children were used in the
study. These cost data must be used with caution because of the small number of
children in some fluorosis categories and the wide variance in costs estimated by the
dentists.
Minimum repair costs are lower than optimum repair costs for two reasons. First,
minimum repair procedures were performed less frequently per subject than optimum
repair procedures. Second, one of the most widely used optimum repair procedures is
much more expensive than the most commonly used minimum repair procedures. This
expensive optimum procedure is a porcelain gold crown. 3
‘Note that avoided costs are not a measure of the economic value of benefits.
2 Edwin Collins, University of Texas Health Science Center at San Antonio, letter of June
18, 1985, based on data in E. Collins, et al., “Analysis of Costs for the Treatment of
Dental Fluorosis,” Health Effects Research Laboratory, EPA, Cincinnati.
3 me distribution of the most common repair procedures and their costs are as follows:
Average Number of
Procedures per Dentist
Minimum Repair Performed on 55 Subjects Standard Cost
Bleach 60.67 $ 50.00
Amalgam - 2 surface (perm) 27.83 30.00
Amalgam - I surface (perm) 27.17 22.00
Composite resin-I surface 24.50 28.50
Acid etch 23.33 28.50
63
-------
A maximum contaminant level and repairs to teeth can achieve similar dental
results. With an MCL, the costs of cosmetic and functional repairs to remedy moderate
and severe cases of fluorosis would be avoided, however. These avoided costs for the
nation are shown in Table B-2. These Costs were estimated by multiplying the number of
moderate and severe cases of fluorosis avoided per year if an MCL were promulgated
(Table A-5) by the optimum cost for repair of cosmetic and functional damage to teeth
due to fluorosis from Table B-I.
In some cases additional future repair measures may be needed. The costs of these
measures have not yet been estimated but they are thought to be small for optimum
treatment. Only maintenance measures would be needed. For minimal treatment (which
was not used in Table B-2), extensive future repairs will be needed over the lifetime of
the patient. These future costs have not yet been estimated.
Acrylic lamination 20.17 80.00
Mastique 8.33 80.00
Optimum Repair
Bleach 89.33 $ 50.00
Porcelain gold crown 75.67 $412.50
Pit and fissure sealant 58.17 5.00
Composite resin 34.83 28.50
Amalgam - I surface (perm.) 30.17 22.00
Amalgam - 2 surface (perm.) 30.00 30.00
Acid etch 23.33 28.50
Data are from Collins et al., Table 1 (pp. 19-21) and Table 2 (p. 22)
64
-------
TABLE B-i
Average Costs for Repairing Cosmetic and Functional Damage to
Teeth Due to Fluorosis
Severity of Fluorosis Minimum Repair Cost Optimum Repair Cost
Mild $ 88.80 $ 210.65
Moderate 130.17 481.30
Severe 230.78 $1,063.91
SOURCE: Edwin Collins, University of Texas Health Science Center at San Antonio,
letter of 3une 18, 1985, based on E. Collins, et al., “Analysis of Costs for the
Treatment of Dental Fluorosis,” Health Effects Research Laboratory, U.S.
Environmental Protection Agency, Cincinnati, Appendix H, p. 43.
Costs based on national fee scale.
Severity of fluorosis is based on Dean’s Index.
65
-------
TABLE B-2
Avoided Costs for Remedial Repair of Moderate and Severe
Dental Fluorosis if an MCL is Promulgated
Avoided Annual Costs for Remedial Repair
of Dental Fluorosis
MCL (millions of dollars)
I mg/I $15.6 - $22.5
2 mg/I 1.6 - 2.1
3mg/I 0.7-0.9
4 mg/I 0.3 - 0.4
NOTE: Costs refer to optimal repair using a national fee scale.
66
-------
APPENDIX C
UNCERTAINTY OF COSTS AND BENEFITS
The estimates of health benefits such as cases of dental I luorosis avoided and of
fluoride removal costs are subject to error. There are two ways to incorporate the
uncertainty introduced by these errors into the analysis of costs and benefits of fluoride
removal. One is a statistical method involving confidence intervals and the other is
sensitivity analysis. This appendix indicates how these approaches could be used in the
economic assessment of fluoride regulations.
The statistical approach uses confidence intervals based on the sampling
distribution of the relvant variables. For health benefits, the number of cases of
fluorosis is the product of the probability of fluoride occurring in drinking water in a
given concentration interval, the number of water systems in a given size category, the
probability of fluorosis occurring given a concentration in drinking water, and the number
of people served by a water system of a given size. Occurrence, number of systems,
dose-response, and population served are all random variables. In order to determine
confidence intervals for the product of these random variables it is necessary to know
the sampling distributions of each of the random variables. The status of this
information is as follows:
• Occurrence data - - for higher concentrations of fluoride, theoretically all water
systems are included in the estimate of occurrence. There is therefore no
sampling error, although there is measurement error. A sample was used for
lower concentrations and so a sampling error can be estimated for this set of
systems.
• Number of systems - - presumably EPA has identified the universe so there is no
sampling error.
• Dose-response function--confidence intervals can be calculated for the dose-
response function. For instance, the confidence interval of the coefficient of the
dosage variable in a regression of proportion of children with fluorosis (response)
on the concentration of fluoride in drinking water (dose) is estimated below.
There are insufficient date to compute a dose-response function for
67
-------
osteoscierosis, however.
o Number of persons served by a water system -- confidence intervals could be
calculated from EPA’s 1980 survey of community water systems.
A critical component of uncertainty is the dose-response function. The uncertainty
associated with the dose-response function is reflected in the 95% confidence intervals
of the coefficient of x (fluoride concentration) in the regression equations reported in
Table A-3. These are as follows:
Approximate 95%
confidence interval
Coefficient of of coefficient of.
Equation fluoride concentration fluoride concentration
University of Texas
• moderate fluorosis 6.79 3.15 to 10.43
NIDR
• moderate fluorosis 1.62 -3.65 to 6.89
• severe fluorosis 7.00 -0.84 to 14.84
Given the small number of groups of communities in the NIDR data base (4), one would expect a ver
large range of uncertainty in the dose-response functions.
68
-------
The uncertainty of the benefits of reducing fluoride are shown below using the 95%
confidence interval for the slopes of the three dose-response functions.
Range of Cases of Dental Fluorosis Avoided per Year
(95% Confidence Interval)
Severe Fluorosis
______ ( NEDR data )
o to 19079
Oto 1322
Oto 560
Oto 232
0 to 41085
0 to 2839
Oto 1196
Oto 491
For costs of contaminant removal, there are also several components: cost data
for individual systems, probabilities of utilities selecting a given treatment, the number
of water systems, and the proportion of water systems exceeding the standard. The
status of this information is as follows:
• Cost data for individual systems -- these data are a nonrandom sample of
engineering analyses and design rules of thumb which, because they are
nonrandom, cannot be expressed in terms of confidence intervals.
• Treatment selection probabilities - - these are based on expert judgement and
therefore it is not appropriate to derive statistical confidence intervals.
• Number of water systems -- this is presumably a count of the universe so a
confidence interval is irrelevant.
• Proportion of water systems exceeding the standard -- this is based on a
Moderate Fluorosis
University of Texas NIDR Data
Data ( Illinois )
MCL
I mg/I
2 mg/I
3 mg/!
4 mg/I
730 to 28884
605 to 20004
257 to 849
107 to 353
69
-------
combination of the universe of water systems (for high fluoride concentrations)
and a sample of water systems (for lower fluoride concentrations). Thus a
confidence interval depends on fluoride concentration.
When multiplied together these data are not appropriate for use in constructing a
statistical confidence interval.
A sensitivity analysis may be more appropriate than a statistical analysis to
determine the effects of uncertainty in the cost data. Expert judgement could be used to
establish a range of costs of fluoride removal for various system size categories and a
range of probabilities of utilities selecting various treatments. Thus a range of national
costs could be estimated.
70
-------
• APPENDIX D
ADDITIONAL COSTS FOR CARIES TREATMENT IF A FLUORIDE MCL IS SET
Removing fluoride from drinking water may cause an increased incidence of caries
in children. This appendix provides a calculation of the costs of treating the expected
number of additional caries in children resulting from reducing the concentration of
fluoride in drinking water.
The following assumptions were made:
a. The population of interest is children exposed to excess fluoride, ages 7 to 12.’
b. The relevant treatment costs are those given by Collins et al. 2
c. Annual costs to the nation at a given MCL are (C* - C 1 ) P 1 /6, where C 1 is the caries
treatment cost per person at fluoride concentration category i, C is the caries
treatment cost per person at the fluoride level just below the MCL, P 1 is the number
of children ages 7 to 12 exposed to fluoride concentration category i , and the sum
is divided by 6 to annualize the costs over the six years of age included.
For example, the treatment cost for caries if the fluoride concentration is 1.0 to
1.9 times optimal is $101 and the treatment cost for caries if the fluoride concentration
is between 4.0 and 4.9 times optimal is $55. There are 17,607 children aged 7 to 12
‘U.S. Bureau of the Census, 1980 Census of Population
2 E. Collins et al, “Analysis of Costs for the Treatment of Dental Fluorosis; Health
Effects Research Laboratory EPA, Cincinnati Table 8. The costs are $160 per child if
the community fluoride level is 0 - 0.9 mg/I; $101 if fluoride concentrations are 1.0 - 1.9
mg/I, $67 at 2.0-2.9 mg/I, $76 at 3.0-3.9 mg/I and $55 at 4.0 - 4.9 mg/I.
3 Obtained by multiplying the exposed population from Table A-4 by the proportion
of the population in age category 7 - 12.
71
-------
exposed to fluoride concentrations of 4.0 times optimal or more. The additional cost of
treating their increased caries if an MCL of 2 is imposed is ($101-$55) x 17,607
$809,922. Annualized, this is $134,987.
Carrying out these calculations, the annualized additional costs for caries
treatment if a fluoride MCL is established are as follows:
• MCL: 1 mg/I $27,875,000
• MCL = 2 mg/I 507,000
• MCL 3 mg/I 35,000
• MCL = 4 mg/I 62,000
(The anomalous low value for an MCL of 3 mg/I is due to the lower caries treatment cost
reported by Collins et al. for fluoride concentrations of 2.0 to 2.9 times optimal than for
a fluoride concentration of 3.0 to 3.9 times optimal. This calculation assumes no
increase or decrease in caries treatment costs if the current concentration of fluoride
was 3.0 to 3.9 times optimal and the MCL were set at 3 mg/I).
The net result of this exercise is to add only a very small cost to the annualized
national costs of fluoride removal, less than 5% for an MCL of 2, 3, or 4 mg/I, but
slightly more for an MCL of I mg/I.
72
------- |