United States       Science Advisory      EPA-SAB-IHEC-ADV-99-
        Environmental       Board             April 1999
        Protection Agency	Washington, DC	www.epa.gov/sab
&EPA
     AN SAB ADVISORY: Building
     Assessment and Survey Evaluation
     (BASE) Study Proposed Data
     Analyses
     PREPARED BY THE INTEGRATED
     HUMAN EXPOSURE COMMITTEE (IHEC)
     OF THE SCIENCE ADVISORY BOARD (SAB)

-------
             UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                                 WASHINGTON B.C. 20460
March 31, 1999
Note to the Reader:
                                                                         OFFICE OF
                                                                   THE ADMINISTRATOR
                                                                   SCIENCE ADVISORY BOARD
       The attached draft report is a draft report of the Science Advisory Board (SAB). The draft
is still undergoing final internal SAB review, however, in its present form, it represents the
consensus position of the panel involved in the review. Once approved as final, the report will be
transmitted to the EPA Administrator and will become available to the interested public as a final
report.

       This draft has been released for general information to members of the interested public
and to EPA staff. This is consistent with the SAB policy of releasing draft materials only when the
Committee involved is comfortable that the document is sufficiently complete to provide useful
information to the reader. The reader should remember that this is an unapproved working draft
and that the document should not be used to represent official EPA or SAB views or advice.
Draft documents at this stage of the process often undergo significant revisions before the final
version is approved and published.

       The SAB is not soliciting comments on the advice contained herein. However, as a
courtesy to the EPA Program Office which is the subject of the SAB review, we have asked them
to respond to the issues listed below. Consistent with SAB policy on this matter, the SAB is not
obligated to address any responses which it receives.

1. Has the Committee adequately responded to the questions posed in the Charge?
2. Are any statements or responses made in the draft unclear?
3. Are there any technical errors?

For further information or to respond to the questions above, please contact:

       Roslyn Edson, Designated Federal Officer
       Committee Operations Staff
       Science Advisory Board (1400)
       US Environmental Protection Agency
       Washington, DC 20460
       (202) 260-3823 Fax: (202) 260-7118
       Email: EDSON.ROSLYN@EPA.GOV

-------
                   -  DRAFT  4, MARCH 31,  1999, DO NOT  QUOTE OR CITE-


                                            MarchS, 1999

        EPA-SAB-IHEC-LTR-99-

        Honorable Carol M. Browner
        Administrator
        U.S. Environmental Protection Agency
        401 M Street, SW
        Washington, DC 20460

 1            Subject:      Advisory on the Building Assessment Survey Evaluation (BASE) Study
 2                          Proposed Data Analyses
 3
 4      Dear Ms. Browner:
 5
 6            The Integrated Human Exposure Committee (IHEC) of the Science Advisory Board met
 7      on March 9, 1999 in Washington, DC to conduct an advisory on the proposed data analyses for
 8      the Building Assessment Survey Evaluation (BASE) study.  The BASE study was designed by the
 9      EPA Office of Radiation and Indoor Air in response to the Agency's responsibility to gather
10      information and disseminate guidance regarding indoor air quality (IAQ) health risks under Title
11      IV of the Superfund Amendment Reauthorization Act and in response to the Government
12      Performance and Results Act (GPRA) Goal 4, Objective 4 which states that "By 2005, 15 million
13      more Americans will live or work in homes, schools, or office buildings with healthier indoor air
14      than in 1994."  In this advisory, the Committee is providing advice on the analysis of the data
15      which has been recently  collected.
16
17            BASE is a cross-sectional multi-year study designed to define key characteristics of IAQ
18      in  100 public and  commercial buildings. The BASE project has four goals:
19
20            a)     to collect baseline data characterizing public and commercial office buildings,
21
22            b)     to establish information on important indoor air parameters for policy decisions
2 3                   and guidance development,
24
25            c)     to examine  the relationships among parameters and between parameters and
2 6                   occupants' perceptions and symptoms, and

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT QUOTE OR CITE-

 1             d)    to serve as a basis for hypothesis development.
 2
 3             The Committee was charged to respond to the following Charge questions:
 4
 5             a)    Are the proposed data analyses the most relevant?;
 6
 7             b)    Does the Committee have advice on additional analyses that should be
 8                   considered?;
 9
10             c)    How should the analyses be prioritized considering the need to address relevant
11                   scientific issues and the most important programmatic goals identified by the
12                   Agency? In prioritizing the analyses, which analyses are essential given the
13                   Agency's need to address relevant scientific issues and the most important
14                   programmatic goals identified by the Agency?; and
15
16             d)    Are there similar analyses (that have been conducted on other data sets) that EPA
17                   should use as guidance in its data analysis efforts?
18
19             The ultimate goal of the BASE study is to improve public health through improvements in
2 0      indoor air quality. To reach this goal, it  is necessary to establish baseline information about the
21      characteristics of indoor air in different buildings, in different locations, and under different
22      conditions.  The BASE data will provide the EPA with building profiles, including distributions
23      for the concentration of various toxicants indoors, building operational characteristics  and
24      frequencies of various  symptom complaints.  Therefore, the IHEC found the BASE  survey to be
25      extremely important and commends the Agency personnel who have steered this complex and
2 6      carefully executed data collection effort  to completion.
27
2 8             Overall, the Committee found the proposed analyses to be quite relevant and extremely
29      useful in providing significant data on the contributions of indoor environments to human
30      exposure and adverse health. The Committee found the overall proposed analyses to be useful in
31      helping the Agency to  meet GPRA Goal 4, Objective 4. The analyses of the study parameters can
32      also be useful in determining good IAQ practices  and, subsequently, in helping the EPA to
33      achieve its GPRA goal of having 5% of the office buildings managed with good IAQ practices by
34      2005. The IHEC highly encourages the  Agency integrate the BASE project into the Agency's

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT  QUOTE  OR CITE-

 1      efforts to analyze cumulative exposure (SAB, 1996) to maximize the impact of BASE on the
 2      overall protection of public health.
 3
 4             The IHEC strongly recommends that the Agency focus on conducting Quality
 5      Assurance/Quality Control on the data followed by evaluating the descriptive statistics, in-depth,
 6      in order to provide critically needed baseline information  on the various parameters that have been
 7      monitored in the 100 commercial and public buildings that were included in the study.  The
 8      Committee urges the Agency to release the information to the public as  soon as the QA/QC and
 9      descriptive statistics analysis is completed. The Committee recommends that the Agency consider
10      conducting more complex analyses such as testing for associations after the descriptive statistical
11      data is released. The Committee makes several recommendations for the subsequent analysis of
12      the data.  The IHEC emphasized the importance of the Agency first assigning the level of
13      acceptable power relative to the declaration  of a significant association  before the Agency tests
14      for associations.
15
16             The IHEC recommends that the Agency  incorporate guidelines regarding the scientific
17      limitations in using the data. Such guidelines would reduce the likelihood that the data are
18      misinterpreted and that invalid associations are inferred and would minimize the likelihood of data
19      dredging, especially given the large number  of variables in the study. The Committee cites a few
2 0      data sets with analyses that EPA may be able to use as guidance in its data analysis efforts and
21      emphasizes the importance of analyzing both the BASE data and  the data from the Office of
22      Research and Development longitudinal study, the Temporal Indoor Monitoring and Evaluation
23      (TIME) Study. Conducted by the EPA's Office of Research and Development, the same core
24      parameters from BASE were collected in a smaller number of buildings  in the TIME study.
25      However, unlike the BASE study, samples in the longitudinal study were taken over different
2 6      seasons (Fortmann, R., 1994)(EPA, 1999).  TIME has the potential, when coupled with the
27      BASE study, to provide valuable information on the relationships between cross-sectional and
2 8      longitudinal studies. Therefore, the Committee strongly encourages the Agency to review and
2 9      compare results from both cross-sectional and longitudinal studies simultaneously to make sure
30      that necessary and comparable analyses are  carried out on the data of both studies.   The
31      Committee also encourages the Agency to establish collaborative relationships with other
32      researchers when developing the strategy to conduct the BASE analyses and when conducting the
33      BASE analyses.
34

-------
                   - DRAFT 4,  MARCH 31,  1999,  DO NOT QUOTE  OR CITE-

 1            The Committee appreciates the opportunity to provide advice to the Agency on the BASE
 2      data analyses and looks forward to receiving a written response from the Assistant Administrator
 3      for Air and Radiation (OAR).
 4
 5                                       Sincerely,
 6
 7
 8
 9                                       Dr. Joan M. Daisey, Chair
10                                       Science Advisory Board
11
12
13
14                                       Dr. Henry A. Anderson, Chair
15                                       Integrated Human Exposure Committee
16                                       Science Advisory Board

-------
                   -  DRAFT  4,  MARCH  31,  1999,  DO NOT QUOTE OR CITE-

 1                                            NOTICE
 2
 3
 4             This report has been written as a part of the activities of the Science Advisory Board, a
 5      public advisory group providing extramural scientific information and advice to the Administrator
 6      and other officials of the Environmental Protection Agency. The Board is structured to provide
 7      balanced expert assessment of scientific matters related to problems faced by the Agency. This
 8      report has not been reviewed for approval by the Agency; and hence, the contents of this report
 9      do not necessarily represent the views and policies of the Environmental Protection Agency or
10      other agencies in the Federal government. Mention of trade names or commercial products does
11      not constitute a recommendation for use.
12
13
14
15

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT  QUOTE  OR CITE-

 1                                            ABSTRACT
 2
 3             Overall, the Committee found the proposed analyses to be quite relevant and extremely
 4      useful in providing significant data on the contributions of indoor environments to human
 5      exposure and reported symptoms.  The frequency distributions of the normative data are the
 6      hallmark of this project and should be extremely useful in supplying relevant and useful yardsticks
 7      to practitioners studying indoor air. The Committee found the overall proposed analyses to be
 8      useful in helping the Agency to meet GPRA Goal 4, Objective 4, which states that "By 2005, 15
 9      million more Americans will live or work in homes, schools, or office buildings with healthier
10      indoor air than in 1994."  The analyses of the study parameters can also be useful in determining
11      good IAQ practices and, subsequently, in helping the EPA to achieve its GPRA goal of having
12      5% of the office buildings managed with good IAQ practices by 2005. The IHEC highly
13      encouraged the Agency integrate the BASE project into the Agency's efforts to analyze
14      cumulative exposure in order to maximize the impact of BASE on the overall protection  of public
15      health (SAB, 1996).
16
17             The IHEC strongly recommended that the Agency focus on conducting Quality
18      Assurance/Quality Control on the data and then conduct an in-depth evaluation of the descriptive
19      statistics in order to provide  critically needed baseline information on the various parameters that
2 0      have been monitored in the 100 commercial and public buildings that were included in the study.
21      The Committee urged the Agency to release the information to the public as soon as the QA/QC
22      and descriptive statistics analyses  are completed. It was recommended that more complex
23      analyses, such as testing for  associations, be considered after the baseline data is released. The
24      IHEC provides several recommendations for the subsequent data analyses.  The IHEC
2 5      emphasized the importance of the Agency first assigning the level of acceptable power relative to
2 6      the declaration of a significant association before the Agency evaluates the existing power relative
27      to these criteria.
28
2 9             The IHEC recommended that the Agency incorporate guidelines regarding the scientific
3 0      limitations in using the data. Such guidelines would reduce the likelihood that the data are
31      misinterpreted or that invalid associations are inferred and would minimize the likelihood of data
32      dredging, especially given the large number of variables in the study. The Committee cited a few
33      data sets with analyses that EPA may be able to use as guidance in its data analysis efforts and
34      emphasized the importance of analyzing both the BASE data and the data from the Office of
35      Research and Development longitudinal study, the Temporal Indoor Monitoring and Evaluation

-------
                  - DRAFT 4,  MARCH  31,  1999,  DO NOT  QUOTE OR CITE-

1      Study (TIME). The Committee also encouraged the Agency to establish collaborative
2      relationships with other researchers when developing the strategy to conduct the BASE analyses
3      and while conducting the BASE analyses.
4
5      Keywords:    Building Assessment Survey and Evaluation Study (BASE); indoor air; indoor air
6                  quality (IAQ); indoor environments; human exposure; Government Performance
7                  and Results Act (GPRA); and cumulative exposure; Temporal Indoor Monitoring
8                  and Evaluation Study (TIME).
                                                in

-------
                   - DRAFT 4,  MARCH  31,  1999, DO NOT QUOTE OR CITE-

 1                              U.S. Environmental Protection Agency
 2                                     Science Advisory Board
 3                             Integrated Human Exposure Committee
 4             IHEC Building Assessment Survey and Evaluation (BASE) Advisory Panel
 5
 6      Chair
 7      Dr. Henry A. Anderson, Chief Medical Officer, Wisconsin Bureau of Public Health,
 8            Madison, WI 53703
 9
10      Members
11      Dr. Annette Guiseppi-Elie, Senior Consultant, Dupont Engineering, Corporate Remediation,
12            Wilmington, DE 19880-0027
13
14      Dr. Michael Jayjock, Senior Research Fellow, Rohm and Haas  Co., Research Laboratories,
15            Spring House, PA 19477
16
17      Dr. Lovell Jones, Director, Experimental Gynecology-Endocrinology, Department of
18            Gynecologic Oncology; and Professor of Gynecologic Oncology; MD Anderson Cancer
19            Center, Houston, TX 77030 (Dr. Jones did not participate in this advisory)
20
21      Dr. Michael D. Lebowitz, Professor of Medicine and Epidemiology, Arizona Prevention
22            Center, University of Arizona College of Medicine, Tucson, AZ 85724-5163
23
24      Dr. Kai-Shen Liu, Epidemiologist, California Department of Health Services,
25            Berkeley, CA 94704-9989
26
27      Dr. Thomas McKone, Staff Scientist and Professor, School of Public Health,
2 8            University of California, Berkeley, CA 94720-7360
2 9            (Dr. McKone did not participate in this advisory.)
30
31      Dr. Jerome O. Nriagu, Professor, University of Michigan, School of Public Health,
32            Department of Environmental  and Industrial Health, Ann Arbor, MI 48109-2029
33
34      Dr. Barbara Petersen, President, Novigen  Sciences, Inc., Washington, DC 20036
35
36      Dr. David Wallinga, Senior Scientist, Natural Resources Defense  Council,
37            Washington, DC 20005
38
39      Dr. Charles Weschler, Senior Scientist, Bell Communication Research,
40            Red Bank, NJ 07701
41
                                                 IV

-------
                   - DRAFT 4,  MARCH  31,  1999,  DO NOT  QUOTE OR CITE-

 1      Radiation Advisory Committee Liaison
 2      Dr. Jill Lipoti, Assistant Director for Radiation Protection Programs, Division of Environmental
 3            Safety, Health and Analytical Programs, New Jersey Department of Environmental
 4            Protection, Trenton, NJ 08625-0415
 5
 6      Science Advisory Board Staff
 7      Ms. Roslyn Edson, Designated Federal Official, U.S. Environmental Protection Agency,
 8      Science Advisory Board (1400), 401 M Street, SW, Washington, DC  20460
 9            (202) 260-3823; FAX: (202) 260-7118; edson.roslyn@epa.gov
10
11      Ms. Wanda R. Fields, Management Assistant, U.S. Environmental Protection Agency,
12            Science Advisory Board (1400), 401 M Street, S.W., Washington, DC 20460
13            (202)260-5510 FAX: (202)260-7118;  fields.wanda@epa.gov

-------
                 - DRAFT 4,  MARCH  31, 1999, DO  NOT QUOTE OR CITE-

 l                               TABLE OF CONTENTS
 2
 3
 4     1. EXECUTIVE SUMMARY  	1
 5
 6     2. INTRODUCTION	4
 7           2.1    Background	4
 8           2.2    The Review and Charge	6
 9
10     3. RESPONSE TO CHARGE QUESTIONS  	7
11           3.1    General Findings	7
12           3.2    Charge Question 1: Relevance of the Proposed Data Analyses 	8
13           3.3    Charge Question 2: Additional Analyses That Should Be Considered	10
14           3.4    Charge Question 3:  Prioritizing the Data Analyses	15
15           3.5    Charge Question 4: Similar Analyses for Guidance in BASE Data Analyses ...18
16
17     4. SUMMARY OF RECOMMENDATIONS AND CONCLUSIONS  	22
18
19     REFERENCES CITED	R-l
20
21     APPENDIX A - ACRONYMS AND ABBREVIATIONS	A-l
22
23     APPENDIX B - BASE Core Parameters	B-l
24
25     APPENDIX C - Proposed BASE Analyses 	C-l
26
27     APPENDIX D - GPRA Strategic Goal 4,  Objective 4  	D-l
28
                                            VI

-------
                   -  DRAFT  4, MARCH 31,  1999, DO NOT  QUOTE  OR  CITE-

 l                                 1.  EXECUTIVE SUMMARY
 2
 3            The Integrated Human Exposure Committee (IHEC) of EPA's Science Advisory Board,
 4      supplemented by a radon expert (a liaison from the SAB Radiation Advisory Committee), met on
 5      March 9, 1999 to review the proposed data analyses for the Building Assessment Survey and
 6      Assessment (BASE) study.  The purpose of BASE is to help fill the significant data gap regarding
 7      baseline indoor air quality (IAQ) in public and commercial office buildings in the United States.
 8      This study was conducted by the EPA Office of Radiation and Indoor Air.
 9
10            The Committee addressed the following Charge questions:
11
12            a)     Are the proposed data analyses the most relevant!
13                   Overall, the Committee found the proposed analyses to be quite relevant and
14                   extremely useful in providing significant data on the contributions of indoor
15                   environments to human exposure and reported symptoms. The Committee also
16                   found the overall proposed analyses to be useful in helping the Agency to meet
17                   GPRA Goal 4, Objective 4, which states that "By 2005, 15 million more
18                   Americans will live or work in homes, schools, or office buildings with healthier
19                   indoor air than in 1994."  The analyses of the study parameters can also be useful
20                   in determining good IAQ practices and, subsequently, in helping the EPA to
21                   achieve its GPRA goal of having 5% of the office buildings managed with good
22                   IAQ practices by 2005.  The IHEC highly encouraged the Agency integrate the
23                   BASE project into the Agency's efforts to analyze cumulative exposure (SAB,
24                   1996) in order to maximize the impact  of BASE on the overall protection of public
25                   health.
26
27
28            b)     Does the Committee have advice on additional analyses that should be
2 9                   considered?
30                          The Committee recognized the  significant effort that the Agency has
31                   undertaken in performing this study. As mentioned in the previous section,
32                   overall, IHEC found that the data analysis being proposed is adequate and
33                   comprehensive. In an effort to facilitate getting the normative information in the
34                   published literature, the Committee hesitated to recommend additional analyses

-------
                    - DRAFT 4, MARCH  31,  1999,  DO NOT QUOTE OR CITE-

 1                    that may be useful to perform since those analyses are not as critical in the
 2                    near-term.  Thus, in performing the current analyses, the Committee recommended
 3                    a focus on the QA/QC of the available data that will facilitate publishing quality
 4                    data that may be used by the Agency and others in future evaluations. The IHEC
 5                    provided advice on the subsequent data analysis and commented on: the
 6                    aggregation of the data; uncertainty analysis; the classification of chemicals; testing
 7                    for associations; confounding factors; psychosocial stress; the building symptoms
 8                    index; and the sampling protocol.  The Committee encourages the Agency to work
 9                    with other organizations to define and perform these additional analyses.
10
11             c)      How should the analyses be prioritized considering the need to address relevant
12                    scientific issues and the most important programmatic goals identified by the
13                    Agency? In prioritizing the analyses, which analyses are essential given the
14                    Agency's need to address relevant scientific issues and the most important
15                    programmatic goals identified by the Agency?
16                           The IHEC strongly recommended that the Agency  focus on conducting
17                    Quality Assurance/Quality Control on the data followed by evaluating the
18                    descriptive statistics, in-depth, in order to provide critically needed baseline
19                    information on the various parameters that have been monitored in the 100
2 0                    commercial and public buildings that were included  in the study. The Committee
21                    urged the Agency to release the information to the public as soon as the QA/QC
22                    and descriptive statistics analyses are completed.  The Committee recommended
23                    that the Agency consider conducting more complex analyses such  as testing for
24                    associations after releasing the data on the descriptive analysis.  The IHEC
2 5                    emphasized the importance of the Agency first assigning the level of acceptable
2 6                    power relative to the declaration of a significant association before the Agency
27                    tests for associations.
28
29             d)      Are there similar analyses (that have been conducted on other data sets) that EPA
3 0                    should use as guidance in its data analysis efforts?
31                    There are many similar analyses that have been conducted on other data sets,
32                    essentially too numerous to list. There are multiple  studies in the literature
33                    (including proceedings from Indoor Air & Healthy Buildings) of which the EPA
34                    staff in the Office of Air and Radiation, Office of Radiation and Indoor Air, Indoor
35                    Environments Division are aware. The IHEC cited  four data sets with analyses

-------
                  - DRAFT 4,  MARCH  31,  1999,  DO NOT QUOTE  OR CITE-

1                   that may be useful as guidance in EPA's data analysis efforts.  Those data sets are
2                   from the following studies are: the California Healthy Building Study, the
3                   European Audit Project to Optimize Indoor Air Quality and Energy Consumption
4                   in Office Buildings, the Japanese Office Building Survey and the TEAM study.
5                   The Committee also emphasized the importance of reviewing and comparing
6                   results from both BASE and the Temporal Indoor Monitoring and Evaluation
7                   Study (TIME).

-------
                   -  DRAFT  4,  MARCH  31,  1999,  DO NOT QUOTE OR CITE-

 1                                     2.  INTRODUCTION
 2
 3      2.1     Background
 4
 5             BASE is a cross-sectional, multi-year study designed to define key characteristics of IAQ
 6      in 100 public and commercial buildings.  The BASE project has four goals:
 7
 8             a)     to collect baseline data characterizing public and commercial office buildings,
 9             b)     to establish information on important indoor air parameters for policy decisions
10                   and guidance development (a list of the parameters is included in Appendix B),
11             c)     to examine the relationships among parameters and between parameters and
12                   occupants' perceptions and symptoms, and
13             d)     to serve as a basis for hypothesis development.
14
15             Buildings were randomly selected from cities with a population of at least 100,000 in 10
16      climatic regions.  Businesses were also selected randomly using business listings obtained for a
17      given city. In order for a business to be included in the BASE study, the building
18      owner/management had to be willing to participate, the building could not be highly publicized  as
19      a "sick" or "problem" building, and the building had to meet the study area criteria.  In the BASE
2 0      study, the study area criteria was defined as a building which is served by no more than two air
21      handlers, houses a minimum of 50 employees, and has a maximum of 3 floors.
22
23             In the 100 buildings selected for the BASE study, specific environmental measures were
24      taken, building and heating, ventilation, and air-conditioning (HVAC) characteristics were
25      defined, and occupant questionnaires were completed. A list  of the specific core parameters and
2 6      measurements taken are provided in Attachment B. Examples of environmental measures
27      included temperature, relative humidity and carbon dioxide measurements.  Some of the building
2 8      characteristics that were recorded included building use, occupancy and smoking policy. The
29      Indoor Environmental Quality Questionnaire included questions on job characteristics, health and
30      well-being of the occupant, and work place environmental conditions such as the cleanliness of
31      the workspace and the lighting conditions (EPA,  1994). The sampling sites and the schedule of
32      measurements and equipment were also described in the standardized BASE protocol (EPA,
33      1994).  All measurements were taken over the course of a week, from Monday to Friday.
34      Specifically, some environmental measures  such as relative humidity, carbon dioxide and carbon

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT QUOTE OR CITE-

 1      monoxide were sampled continuously from Tuesday to Thursday. For those samples, the Agency
 2      recorded 5 minute averages. Integrated sampling was used to measure the concentration of
 3      VOCs, particles, and formaldehyde, resulting in 8-9 hour exposure levels. Also, the bioaerosols
 4      were only sampled for 2 minutes and 5  minutes twice on Wednesday.
 5
 6             The EPA has completed field measurements on 100 public and commercial buildings.
 7      Some summary statistics have been completed for VOCs (volatile organic compounds), fungi, and
 8      paniculate matter (PM). Some of the data analysis that is underway includes: QA/QC for the data
 9      for the last 14 buildings included in the study; the evaluation of the representativeness of buildings
10      (compared with the Department of Energy (DOE) survey)(DOE, 1995); the evaluation of the
11      precision and accuracy of the data; and the calculation of ventilation rates of the buildings that
12      were included in the survey.
13
14             The Agency developed a proposed plan for the data analysis that was designed to best
15      address relevant scientific issues and the most important programmatic goals identified by the
16      Agency.  Specifically,  the data analysis plan was developed based on the Government
17      Performance and Results Act (GPRA) goals, especially GPRA Goal  4, Objective 4 which is
18      included in Appendix D. In the proposed BASE analysis plan, EPA has identified and prioritized
19      six types of analyses.  Those analyses include: a) quality assurance/quality control; b)
20      representativeness of the building samples and weighting determinations; c) ventilation rate
21      calculations - % outdoor air, air exchange rate and the amount of cubic feet per minute of air per
22      occupant; d) frequency distribution (for normative data); e) associations; and f)  indices and
23      measures. A description of those analyses is provided below.
24
25             a)     Quality Assurance/Quality Control
2 6                    EPA plans to conduct a  quality assurance/quality control analysis first to review
27                    the data for errors, needed changes, or missing data.  In this QA/QC procedure,
28                    the EPA also plans to determine the accuracy and precision of the data.
29
30             b)     Representativeness of Building Samples and Weighting Determinations
31                    The representativeness of the buildings samples was selected as the second
32                    analysis.  The Agency plans to compare regional frequency distributions of
33                    building characteristics such as occupancy, building age, gross floor area, and the
34                    number of floor to those found in the DOE study; to examine any potential biases

-------
                   -  DRAFT  4,  MARCH  31,  1999,  DO  NOT  QUOTE  OR  CITE-

 1                   resulting from the building selection process; and to develop nationally
 2                   representative distributions (DOE, 1995).
 3
 4             c)     Ventilation Rate Calculations - % outdoor air, air exchange rate and
 5                   CFM/occupant
 6                   After conducting analyses on the representativeness of the buildings, the Agency
 7                   plans, as the third analysis, to calculate the ventilation rates to  determine the
 8                   percent of outdoor air, air exchange rate and the cubic feet per minute (CFM) of
 9                   air per occupant.
10
11             d)     Frequency Distribution (normative data)
12                   The fourth analysis, the frequency distributions (on normative data) would be
13                   determined for several of the core parameters, including concentrations of
14                   environmental measurements, symptoms reported on the occupant questionnaire,
15                   building maintenance practices, occupant demographics,  comfort parameters (such
16                   as continuous temperature, relative humidity, sound and light), and sources (such
17                   as furnishings, computers, cleaning materials, and cigarette smoke).
18
19             e)     Associations
20                   In its fifth analysis, the Agency proposes to test for associations between the core
21                   parameters such as the relationships between symptoms, demographics,
22                   environmental parameters, and building and HVAC characteristics.
23
24             f)     Developing Indices and Measures
25                   Finally, in its  sixth analysis, the EPA proposes to develop indices and measures
2 6                   including indices for building symptoms, indoor pollutants, and building system
27                   quality.
28
29      2.2     The Review and Charge
30
31             On March 9, 1999, the Integrated Human Exposure Committee met in Washington, DC to
32      conduct an advisory on the proposed data analyses for the Building Assessment Survey and
33      Evaluation (BASE) project.  The IHEC was charged to respond to four  questions. These
34      questions and the responses by IHEC are presented in the next section.
35                        3.  RESPONSE TO CHARGE QUESTIONS

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT QUOTE OR CITE-

 1
 2      3.1    General Findings
 3
 4             Overall, the Committee found the proposed analyses to be quite relevant and extremely
 5      useful in providing significant data on the contributions of indoor environments to human
 6      exposure and reported symptoms.  The frequency distributions of the normative data are the
 7      hallmark of this project and should be extremely useful in supplying relevant and useful yardsticks
 8      to practitioners studying indoor air. The Committee also found the overall proposed analyses to
 9      be useful in helping the Agency to meet GPRA Goal 4, Objective 4.  The analyses of the study
10      parameters can be useful in determining good IAQ practices which can ultimately help the EPA to
11      achieve their GPRA goal of having 5% of the office buildings managed with good IAQ practices
12      by 2005. The IHEC recommends that the Agency integrate the BASE project into the Agency's
13      efforts to analyze cumulative exposure.
14
15             The IHEC strongly recommends that the Agency first focus on conducting Quality
16      Assurance/Quality Control on the data and then focus on evaluating the descriptive statistics, in-
17      depth, in order to provide critically needed baseline information on the various parameters that
18      have been monitored in the 100 commercial and public buildings that were included in the study
19      since this is the heart of BASE. The Committee urges the Agency to release the information to
20      the public as soon as the QA/QC and descriptive statistics analysis is completed. The Committee
21      recommends that the Agency considers conducting more complex analyses such as testing for
22      associations only after the data on the descriptive analysis has been released.   The IHEC
23      emphasizes the importance of the Agency first assigning the level of acceptable power relative to
24      the declaration of a significant association before the Agency tests for associations.
25
2 6             The IHEC strongly recommends that the Agency, for additional analyses, incorporates
27      guidelines regarding the scientific limitations in using the data, e.g. to reduce the likelihood that
2 8      the data are misinterpreted or that invalid associations are inferred.  It will also be necessary to
2 9      developed procedures to minimize the likelihood of data dredging, especially given the large
30      number of variables in the study. This may be particularly important in using the data on
31      symptoms in conjunction with the data on the building characteristics.  The Committee also cites a
32      few data sets with analyses that EPA may be able to use as guidance in its data analysis efforts
33      and emphasizes the importance of analyzing both the BASE data and the data from the Office of
34      Research and Development longitudinal study,  the Temporal Indoor Monitoring and Evaluation
35      (TIME) study. The Committee also encouraged the Agency to establish collaborative

                                                   7

-------
                    -  DRAFT 4,  MARCH 31,  1999, DO NOT  QUOTE  OR  CITE-

 1      relationships with other researchers when developing the strategy to conduct the BASE analyses
 2      and while conducting the BASE analyses.
 3
 4      3.2    Charge Question 1: Relevance of the Proposed Data Analyses
 5
 6      Are the proposed analyses the most relevant?
 7             The Committee found the proposed analyses to be quite relevant and  extremely useful
 8      to those who are  concerned with the contribution of indoor environments to human
 9      exposures. IHEC commends the Agency personnel who have steered this complex and
10      carefully  executed effort to completion. The Committee also found the overall proposed
11      analyses to be useful in helping the Agency to meet GPRA Goal 4, Objective 4.  For example,
12      the associations between symptoms, environmental parameters and building and HVAC
13      characteristics could be useful in determining good IAQ practices. These practices could then
14      be publicized to help building owners to improve their IAQ practices and to help EPA to
15      achieve their GPRA goal of having 5% of the office buildings managed with good IAQ
16      practices  by 2005.  The IHEC strongly recommends that the Agency integrates the BASE
17      project into its cumulative exposure efforts such as the NHEXAS project and the Cumulative
18      Exposure Project (SAB, 1996).  As an evaluation tool to assist the Agency further in
19      responding to this question, the IHEC recommends that the EPA consider use a matrix,
2 0      assigning point values  according to the contribution of each analysis to each GPRA or
21      program goal.  This type of system could make it easier for the Agency to identify high
22      priority analyses.
23
24
25             The Committee recommends that the Agency modifies the prioritization of the data
2 6      analyses to begin with  Quality Assurance/Quality Control analyses, followed by an analysis of
27      frequency distributions, and then an analysis of ventilation rate calculations. The frequency
28      distribution analysis should also include an examination of the shape of the distribution in
29      addition to an examination of medians and interquartile ranges. Before proceeding with the
30      frequency distribution analysis, the Agency should  reevaluate non-detect values using more
31      recent approaches (including simulation/Monte Carlo analyses).  The IHEC recommends that
32      the analyses of the associations and indices be placed at the bottom of the priority list of the
33      data analyses. More specific advice on the prioritization of the data analyses is provided in the
34      Committee's response to Charge question 3 in Section 3.4 of this report.
35

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT  QUOTE  OR CITE-

 1             a)     Associations
 2
 3             There was some concern that there may be a temptation to select the types of
 4             associations based on curiosity as opposed to significance, especially given the large
 5             number of associations that the Agency could analyze.
 6
 7             b)     Averaging the data
 8             There was also a concern that important information on outliers and sensitive
 9             populations would be lost by averaging the data. Specific recommendations regarding
10             these concerns are provided in the Committee's response to Charge question 2 in
11             Section 3.3 in the discussion on the aggregation of the data.
12
13             c)     Reliance on subjective reports of symptoms
14             There was a concern that the current study  depends solely on more subjective reports
15             of symptoms, rather than actual signs of health effects.  The occupant questionnaire
16             included several subjective questions about employee health and well-being. Inclusion
17             of analyses on both signs and symptoms would have provided a more complete
18             picture of indoor air quality. Therefore, the Committee recommends that in future
19             studies on indoor air quality the Agency consider including more objective health data
2 0             such as physical exams and biomarkers. For example, a short physical exam could
21             include observations for dermatitis or wheezing.  Also, several biomarkers exists for
22             determining the presence of various chemicals or their metabolites  in the body,
23             including several VOCs and pesticides.
24
25
26             d)     Potential sources of indoor air quality lacking adequate data in the BASE
27                   study
2 8             The Committee recommends that, in future studies, the Agency include more
2 9             information on pesticide exposure, allergens, and cleaning agents in order to provide a
30             more complete picture of their significance as sources for IAQ complaints and IAQ-
31             related health effects. For example, the EPA Indoor Environmental Quality Survey,
32             does not mention pesticides or pesticide exposures. The monitoring checklist on page
33             D-4 of the supplement does include pesticides.  However, this informal monitoring,
34             apparently performed only twice during one day of the study, depends on actual
35             observations of pesticide application which would be unlikely during the hours when

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT  QUOTE  OR CITE-

 1             the study was done. The IHEC was also concerned that the BASE protocol does not
 2             include specific information about cleaning agents or specific allergens. Some cleaning
 3             agents are potential sources of indoor exposure to hazardous chemicals via
 4             aersolization and dermal exposure. The building maintenance workers, who typically
 5             clean at night, would, in general, have more detailed knowledge of these sources than
 6             building occupants.
 7
 8      3.3    Charge Question 2: Additional Analyses That Should Be Considered
 9
10             Does the Committee have advice on additional analyses that should be considered?
11             The Committee recognizes the significant effort that the Agency has undertaken in
12      performing this study.  As mentioned in the previous section, overall, the data analysis being
13      proposed is adequate and comprehensive.  In an effort to facilitate getting the normative
14      information in the published literature, the Committee hesitates to recommend additional
15      analyses that may be useful to perform since those analyses are not critical in the near-term.
16      Thus, in performing the current analyses, the Committee recommends a focus on the QA/QC
17      of the available data which will facilitate publishing quality data that may be used by the
18      Agency and others in future evaluations.  The Committee encourages the Agency to work
19      with other organizations to define and perform these additional analyses.
20
21             The Committee provides several recommendations on some of these subsequent
22      analyses. However, the contextual framework and the issues to be explored using the data
23      will have a strong influence on whether additional analysis will be required. Since the
24      analytical models are still being developed, the Committee recommends that the following
25      factors be considered in the data  analysis.
26
27
                                                 10

-------
                    - DRAFT  4, MARCH  31,  1999,  DO NOT  QUOTE  OR CITE-

 1             a)     Aggregation of the data
 2                    There are concerns that in aggregating data, valuable information may be lost.
 3                    The need to provide the normative frequency distributions is viewed as more
 4                    important than developing nationally representative distributions, particularly
 5                    for example, for any parameters that are best described on a regional basis
 6                    (e.g., construction type).
 7
 8                    Another concern is that by averaging the data into frequency distributions, one
 9                    could lose the power of looking at the outlier data.  Thus, in developing
10                    distributions, there should be a discussion of the statistical rigor, or at a
11                    minimum the level of confidence/uncertainty, with which any distribution is
12                    developed. The shape of the distributions should be examined, not just the
13                    medians and interquartiles. Also, the relevance of outliers should be addressed.
14
15                    However, it may be appropriate to group some types of parameters, including
16                    chemicals, to reduce the total number of analyses. Any effort to do so should
17                    be carefully examined to ensure that the grouping is appropriate as described
18                    below.
19
20             b)     Classes of chemicals and. individual chemicals
21                    An effort should be made to classify the chemicals into a small number of
22                    categories in order to facilitate the use of the data in risk analysis.  There are
23                    several classification schemes that relate the chemical properties of compounds
24                    to the toxicity potential.  However, the IHEC also recommends that the
25                    Agency analyzes individual chemicals in addition to classes of chemicals since
2 6                    both may be found have a significant effect on indoor air quality.  The IHEC
2 7                    recommends that the Agency start with structure-activity relationships when
2 8                    conducting analyses on chemical classes.
29
30                    Although the use of "toxicity equivalence units" has inherent flaws, the fact
31                    that people are exposed to a mixture of indoor air contaminants at any given
32                    time should be realized.  In September 1998, the IHEC reviewed the
33                    disproportionate impact methodologies that the Agency was proposing to use
34                    to help it respond to complaints filed under Title VI of the Civil Rights Act of
35                    1964 that allege discriminatory effects from the issuance of pollution control

                                                  11

-------
                    - DRAFT  4,  MARCH  31,  1999,  DO  NOT QUOTE OR CITE-

 1                    permits by states or other governmental bodies that receive financial assistance
 2                    from EPA.  Those methodologies included the use of "toxicity equivalence
 3                    units."(SAB,  1999).
 4
 5             c)     Uncertainty analyses
 6                    The uncertainty in data for any collected parameter should be addressed, as far
 7                    as possible. The Committee found the proposal for performing the stated
 8                    QA/QC analyses to be appropriate.  However, the IHEC recommends that the
 9                    Agency include the level of confidence in the reported data for factors (e.g.,
10                    smoking policy) that cannot be absolutely defined. It was also recommended
11                    the non-detect values be re-evaluated using more current approaches like
12                    Monte Carlo simulation methods.
13
14             d)   Associations
15                    Although categories of associations are listed, no specific ones are defined.
16                    After providing the normative data, this type of analyses is considered the  most
17                    useful. Methods to perform the analyses include simple pattern analysis using
18                    the raw data, development of building profiles, and multivariate and stratified
19                    analyses. When conducting multivariate and stratified analyses, the Agency
2 0                    should consider using demographic risk factors and building risk factors as
21                    covariates and/or effect modifiers. (Risk indices could be calculated for
22                    individuals and for buildings, as well as the other indices indicated.  These risk
23                    indices could then be used in the analyses). This is especially important in
24                    looking at the simplistic symptoms (and indices) in relation to environmental
25                    parameters (where other "complaints"  are treated  as confounders) and/or
26                    building/HVAC characteristics. Before conducting multivariate analyses, the
27                    Agency should first determine if there is colinearity between any of the core
2 8                    parameters such as between contaminants and between contaminants and
2 9                    HVAC characteristics. However, care should be exercised in performing these
3 0                    analyses because of the danger of defining cause and effect interpretations
31                    where none exist.
32
33             e)     Confounding factors
34                    Confounding  factors must be considered in any effort to establish causal
35                    relationships in the data.  For instance, in analyzing the data, the Agency

                                                  12

-------
                   -  DRAFT  4,  MARCH  31,  1999,  DO NOT QUOTE OR CITE-

 1                   should control for the effects of existing medical conditions when trying to
 2                   establish the relationship between exposure and reported symptoms since the
 3                   symptoms themselves may not be independent variables. The Committee
 4                   expressed some concern about the utility of the occupant questionnaire given
 5                   the possibility of a healthy worker bias since sick workers may not have been
 6                   captured during the questionnaire process.  The study was probably not
 7                   designed to address confounders and cofactors. Therefore, it is particularly
 8                   important that the EPA clearly document the limitations in use of the data in
 9                   the exposure-response assessment. For future analyses of the worker well-
10                   being, the IHEC stresses the importance of obtaining information from
11                   employees on sick leave.
12
13             f)     Psychosocialfactors
14                   The psychosocial  factors (work stressors, at-home stressors,  ergonomic factors
15                   of lighting, office comfort and proximity to windows) should be included in the
16                   assessment of relationships between exposure and health outcomes
17                   (symptoms).  The association of psychosocial  stressors with the reported
18                   symptoms may be particularly important in the population under study.
19
20             g)     Building Symptoms Index (BSI)
21                   The IHEC found that concept of Building Symptoms Index (BSI) needs further
22                   development to increase its power as a risk assessment tool.  In its present
23                   form, the ranking of individual symptoms are summed to obtain the value for
24                   BSI. The assumption that each factor contributes equally to  the "total"
25                   symptom may be unrealistic in many instances. The IHEC recommends that
2 6                   the Agency consider using a weighting which takes into account the
27                   importance of key risk factors (in the indoor environment) and the severity of
2 8                   the symptoms.
29
30                   Since the "symptoms"  are so non-specific, the development of "indices" that
31                   may be subsequently used in cause and effect relationships should be carefully
32                   evaluated. Any subsequent interpretations using these  "indices" should be
33                   based on analyses that relate effects not only to exposure (because of the
34                   presence of the material) but should be compared to "environmental levels"
                                                 13

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO NOT  QUOTE  OR CITE-

 1                   that would be expected to produce "effects" consistent with symptoms or other
 2                   adverse effects.
 3
 4             h)     Sampling Schedule
 5                   Some environmental measures such as relative humidity, carbon dioxide and
 6                   carbon monoxide were sampled continuously from Tuesday to Thursday.  For
 7                   those samples, the Agency has recorded 5 minute averages.  However,
 8                   integrated sampling was used to measure the concentration of VOCs, particles,
 9                   and formaldehyde, resulting in an 8-9 hour exposure level.  Also, the
10                   bioaerosols were only sampled for 2 minutes and 5 minutes twice on
11                   Wednesday.  The Committee was concerned that measurements without
12                   continuous data, such as the VOCs, particles and formaldehyde would be  of
13                   limited utility in testing for associations with acute health effects such as
14                   asthma.
15
16         The Committee offers the following general discussion to assist the EPA in putting
17      context around additional  studies as well as  specific recommendations on additional/alternate
18      analyses, should the Agency have the resources to perform these. The Committee
19      recommends that the Agency considers:
20
21             a)     analyzing in other indoor environments including residences and day care
22                   centers;
23
24             b)     identifying one environmental parameter that is a good indicator of overall
2 5                   building air quality so that all environmental parameters do not have to be
2 6                   measured;
27
28             c)     analyzing indoor air to protect "sensitive" populations as well (the current
2 9                   analyses are designed to protect the "average" person);
30
31             d)     determining the correlation between outdoor air quality and indoor air quality;
32
33
34
                                                 14

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT  QUOTE  OR CITE-

 1             e)      conducting longitudinal studies, e.g., evaluating overlapping buildings in the
 2                    ORD study that was carried out over time; and
 3
 4             f)      determining whether the buildings with the highest level of concern for a given
 5                    environmental parameter are also found to be the same buildings with the
 6                    highest levels of concern for another parameter ( e.g., determining if buildings
 7                    with the highest levels of fungi were also the same buildings with the greatest
 8                    amount of water damage and if the buildings with the highest levels of VOCs
 9                    are the same buildings with the most reported symptoms).
10
11      3.4    Charge Question 3:  Prioritizing the Data Analyses
12             How should the analyses be prioritized considering the need to address relevant
13             scientific issues and the most important programmatic goals identified by the
14             Agency? In prioritizing the analyses, which analyses are essential given the Agency's
15             need to address relevant scientific issues and the most important programmatic goals
16             identified by the Agency?
17
18             The Committee recommends that the Agency first discard incomplete and unreliable
19      data before it analyzes the data. The Agency's proposed data analysis is included as Appendix
20      C and is listed in order of priority.  The Committee concurs with the Agency's placement of
21      the Q A/QC at the top of the list. However, the Committee's recommendations for the
22      prioritization of the other data analyses is different from that proposed by the Agency.  The
23      IHEC recommends that the EPA focuses on the outlined frequency distribution first, after
24      conducting the QA/QC. The  IHEC recommends analysis on the frequency distribution for the
25      second analysis since baseline data is the heart of BASE.  This analysis provides baseline
2 6      information on the various parameters that have been monitored in the 100 commercial and
27      public buildings that were included in the study.  After conducing the analysis on the
2 8      frequency distributions, the Agency should then calculate the ventilation rates. The
2 9      Committee places the analyses on the representativeness of the buildings and the analyses on
30      association at the bottom of the list of priorities.  If, at some point, regional weighting factors
31      are developed and employed,  the Committee urges the Agency to be quite explicit when
32      presenting data summaries to indicate whether the data is weighted or unweighted.
33
34
35      Frequency Distributions

                                                  15

-------
                    -  DRAFT  4,  MARCH 31,  1999,  DO  NOT QUOTE OR  CITE-

 1             The frequency distributions of the normative data are the hallmark of this project and
 2      should be extremely useful in supplying relevant and useful yardsticks to practitioners studying
 3      indoor air.  It is the Committee's understanding that the overall shape and central portions
 4      (25th to 75th percentiles) of these distributions are reasonably well defined and will provide
 5      much of value of this study.   The majority of the Committee was of the opinion that the
 6      Agency should direct a lesser, somewhat modest effort and level of attention to the tails of
 7      these distributions. They could be studied for the lessons they may hold. Indeed, these tails
 8      or "outliers" may represent a fundamentally different population and this could be important
 9      information.  A reasonable level of sensitivity analysis could be conducted to provide more
10      information and insight relative to these tails. There was a minority opinion that the Agency
11      could lose significant information on subpopulations if it does not analyze the tails of the
12      distributions.
13
14      Ventilation Rate
15             The Committee recommends that the Agency calculates the ventilation rates rather
16      than have numerous users of the database repeat the exercise and possibly make mistakes.
17      There may be some issues of relevancy with regard to the ventilation rate calculations and
18      their association with occupant symptoms.  Clearly the "core" zones receive less fresh air
19      than "perimeter" zones but the anonymity of the survey data will prevent the placement of
20      these occupants in either zone type.  Also, the critical details of the treatment of infiltration
21      and the issue of whether the representative area was in a "core" or "perimeter" area should be
22      sorted out.
23
24              The Committee anticipates that there will be significant uncertainties in the ventilation
25      rate calculations and recommends that these uncertainties should be clearly stated. The issues
2 6      and problems associated with using CO2 as a surrogate for ventilation rate are well established
2 7      and should be a well-documented caveat in the reports that will describe the BASE data.  In
2 8      those situations where it is possible, comparisons between ventilation rates calculated using
2 9      the CO2 approach versus the temperature approach would be interesting. As an aside for
30      future consideration, one IHEC Member has suggested the possibility of using an incidental
31      outdoor air contaminant as a tracer penetrating within the building to directly measure
32      infiltration. This could be naturally occurring contaminants or SF6 from a distant upwind
33      source.
34
35      Associations

                                                  16

-------
                    - DRAFT  4,  MARCH  31,  1999,  DO NOT  QUOTE  OR CITE-

 1             As mentioned previously, the Committee agreed with the Agency's placement of the
 2      analyses of the associations and indices toward the bottom of the priority listing of analysis.
 3      Clearly, it will be most challenging to focus on the associations that have the greatest impact
 4      for public health; that is, those areas with the greatest practical significance.  They are worthy
 5      and relevant projects but should only be implemented with the best in managerial and
 6      statistical acumen to first assign the level of acceptable power before it tests for associations.
 7
 8             If reasonably well-documented and adequately powered associations are established
 9      between symptoms, environmental parameters and building and HVAC characteristics then
10      such calculations could be useful in determining good IAQ practices. These can then be
11      publicized to help building owners improve their IAQ practices and help EPA achieve their
12      GPRA goal of having 5% of the office buildings managed with good IAQ practices by 2005.
13      Some potential associations to include in the analyses are:
14
15             a)     water damage vs. biological contamination as an indicator of biologicals,
16
17             b)    biologicals vs. asthma incidence  (and biologicals vs. other symptoms),
18
19             c)     type of filtration vs. PM2 5 and type of filtration vs. PM10, and
20
21             d)    indoor volatile organic compounds (VOC) levels vs. ventilation rate.
22
23      As mentioned above, the appropriate and meaningful analyses of the associations will be very
24      challenging given the uncertainties and other limitations of the data.  The Committee
25      commends the EPA for sharing the data set and exploring the possible associations with
2 6      individuals outside the agency.
27
28
                                                  17

-------
                    - DRAFT 4,  MARCH 31,  1999,  DO  NOT  QUOTE  OR CITE-

 1      3.5    Charge Question 4: Similar Analyses for Guidance in BASE Data Analyses
 2
 3             Are there similar analyses (that have been conducted on other data sets) that EPA
 4             should use as guidance in its data analysis efforts?
 5      There are many similar analyses that have been conducted on other data sets, essentially too
 6      numerous to list(including several studies described in the proceedings for Indoor Air &
 7      Healthy Buildings).  The IHEC acknowledges that the EPA personnel associated with the
 8      BASE study are well aware of most of this literature  and cites four studies with analyses that
 9      may be particularly useful as guidance to the EPA for its data analyses.  Those studies are: the
10      California Healthy Building Study, the European Audit Project to Optimize Indoor Air Quality
11      and Energy Consumption in Office Buildings, the Japanese Office Building Survey and the
12      TEAM study.
13
14      The California Healthy Building Study
15             The California  study included investigations on the relationships between the type of
16      ventilation system, VOC  levels, and office worker symptoms in 12 office buildings in the San
17      Francisco Bay Area(Mendell, M.J.,  1996). Using data from this study, J. Ten Brinke, J.
18      Daisey and co-workers from the Lawrence Berkeley  National Laboratory tested seven VOC
19      exposure metrics in terms of their ability to predict complaints among office workers (Ten
20      Brinke, J.T., et.al, 1998). Although some of the metrics were not statistically significant
21      predictors of symptoms, the analysis of the data resulted in at least one statistically significant
22      predictor of symptoms. The BASE study may lend itself to a similar examination.
23
2 4      The European Project
25             The European project compared IAQ parameters across different countries using 56
2 6      office buildings in 9 European countries (Bluyssen, P.M.,  1995). The Agency may find some
27      of the analyses used in the European project useful for analyzing the regional differences of
28      the IAQ core parameters included in the BASE study. However, in it's recommendations on
29      the prioritization  of the data analyses, the IHEC placed analyses on the representativeness of
3 0      the buildings at the bottom of the list of priorities.  Also, the Committee urges that the Agency
31      exercise caution in conducting such an analysis to insure that such regional differences in IAQ
32      measurements and questionnaire data are not "averaged" out. The IHEC also recommends
33      that the Agency indicate whether data is weighted or unweighted if at some point, regional
34      weighted factors  are developed and employed.
35

                                                 18

-------
                    - DRAFT 4,  MARCH 31,  1999, DO NOT  QUOTE  OR  CITE-

 1      The Japanese Study
 2             The Japanese study included 131 office buildings in four major cities.  The proportion
 3      of buildings with indoor environmental measurements exceeding acceptable levels according
 4      to Japanese guidelines (e.g. 1200 ppm for CO2, 10 ppm for CO, temperature within
 5      17°C-28°C, relative humidity within 40-70% etc.) were assessed (Building Management
 6      Education Foundation, 1988).  There are no comparable IAQ guidelines in this country.
 7      However, once the frequency distributions of measured variables have been determined, it
 8      would be easy to assess the proportion of measurements exceeding certain levels (e.g.,
 9      outdoor air standards or indoor levels recommended by other national or international
10      organizations).
11
12      The EPA Team Study
13             Some of the analyses used in the latter stages of the TEAM study should be useful as
14      guidance (and are probably being used for this purpose). The TEAM studies of volatile
15      organic compounds in several U.S. cities, and the Particle (PTEAM) studies in California were
16      important in characterizing the normal ranges of residential indoor and personal air exposures
17      for US populations (EPA, 1996; 1997). In the latter stages of the TEAM study, the VOC
18      data were fitted to log-normal distributions. The fits were reasonably good and the approach
19      provided a convenient way to summarize a large amount of data. A similar approach may be
2 0      useful in summarizing selected environmental measurements in the BASE data set.
21
22      Comparing Data Sets
23             The buildings included in these European and Asian studies are, in many ways, quite
2 4      different from the buildings in the BASE data set. These differences include design,
25      construction materials, furnishings, and the types and manner of operation of the HVAC
2 6      systems.  Nonetheless, besides using the above-mentioned studies for guidance in the
27      developing the analyses plans for BASE, using the European and Asian studies to compare the
2 8      results with those from BASE could also generate some insight into the understanding of
2 9      building problems.  For example, the comparison of questionnaire data between BASE study
30      and the National  Institute for Occupational Safety and Health (NIOSH) investigations of
31      complaint buildings has revealed some factors associated with complaints in office buildings
32      (Brightman, H.S., 1997).  Also, the data from the various studies can be compared to outdoor
33      air standards or recommended indoor levels by various national or international organizations.
34       Any cross-study analysis should include information on indoor sources of exposure.  Such
                                                 19

-------
                    -  DRAFT 4,  MARCH 31,  1999, DO NOT  QUOTE  OR  CITE-

 1      data may provide insight on those parameters that account for differences in indoor air quality
 2      across found across studies.
 3
 4      Complementary Longitudinal Study
 5             Along with the BASE study, there is a complementary longitudinal study referred to as
 6      the Temporal Indoor Monitoring and Evaluation (TIME) study.  (Fortmann, R., 1994)(EPA,
 7      1999)).  Conducted by the EPA's Office of Research and Development, the longitudinal study
 8      collected the same core parameters in a smaller number of buildings. However, unlike the
 9      BASE study, samples in the longitudinal study were taken over different seasons. A portion
10      of the buildings were included in both studies to ensure comparability and provide some
11      information on the relationship between cross-sectional and longitudinal measurements.  The
12      TIME study has the potential, when coupled with the BASE study, to provide valuable
13      information on the relationships between cross-sectional and longitudinal studies. Therefore,
14      the Committee strongly encourages the Agency to review and compare results from both
15      cross-sectional and longitudinal studies simultaneously to make sure that necessary and
16      comparable analyses are carried out on the data of both studies.
17
18      Study on Water Damaged Buildings
19             Swedish researchers Jan Sundell and Carl-Gustof Bornehag have assembled a database
20      that contains studies of water damaged buildings (Sundell, J. and Bornehag, C,1998). The
21      Committee recommends that the EPA investigators contact Sundell and Bornehag for
22      guidance on examining associations between water damage and building complaints for
23      consideration in the analysis of the BASE data set.
24
2 5      Studies from other data sets
2 6             Studies from other data sets, not necessarily building studies, are also
27      relevant in terms of guidance for data analyses. The EPA should consider reevaluating non-
28      detect values using some of the more recently described approaches, including simulations and
2 9      Monte Carlo methods. The Agency should also identify outliers and decide on a consistent
30      procedure for treating them. If there is co-linearity between contaminants or between
31      contaminants and HVAC characteristics, these should be considered before all parameters are
32      included in multivariate analyses.
33
34             In examining potential associations, the EPA should also consider some of the
35      associations that have been reported in other building studies.  However, the EPA should be

                                                 20

-------
                    -  DRAFT  4,  MARCH 31,  1999,  DO NOT  QUOTE  OR CITE-

 1      selective;  it should first consider those potential associations with the largest practical
 2      significance. When multivariate and stratified analyses are performed, demographic and
 3      building risk factors might be included as covariates and/or effect modifiers. (Risk indices
 4      could be calculated for individuals and for buildings, as well as the other indices that have
 5      been mentioned.) This is especially important in looking at the relatively simple symptom
 6      categories in relation to environmental parameters and/or building/HVAC characteristics.
 7
 9
10
11
12
13
14
15
16
17
18
                                                  21

-------
                   -  DRAFT  4,  MARCH  31,  1999,  DO NOT QUOTE OR CITE-

 1          4.  SUMMARY OF RECOMMENDATIONS AND CONCLUSIONS
 2
 3
 4             In this report the IHEC has made a number of recommendations for the BASE data
 5      analyses plan:
 6
 7             a)     The IHEC concurs with the Agency's placement of Quality Assurance/Quality
 8                   Control as a first priority for the data analysis.
 9
10             b)     After Quality Assurance/Quality  Control,  the  Agency  should  analyze  the
11                   descriptive statistics, in-depth, to provide a baseline of information about the
12                   characteristics of indoor air in the 100 commercial and pubic buildings included
13                   in the study.
14
15             c)     The baseline data containing the descriptive  statistics should be released to the
16                   public as soon as the statistical analysis has been completed.
17
18             d)     The BASE study should be integrated into the Agency's other efforts to analyze
19                   cumulative exposure in order to maximize the impact of BASE on the overall
2 0                   protection of pubic health.
21
22             e)     Before testing  for  associations, the Agency must first  assign  the level  of
2 3                   acceptable power relative to the declaration of a significant association.
24
25             f)     The EPA should incorporate guidelines regarding the scientific limitations in using
2 6                   the data to reduce the likelihood that the data are misinterpreted and that invalid
2 7                   associations are inferred and to reduce the likelihood of data dredging, especially
2 8                   given the large number of variables in the study.
29
30             g)     The Agency should review and compare results  from both the cross-sectional
31                   study (BASE) and the longitudinal study (TIME) simultaneously to  make sure
32                   that necessary and comparable analyses are carried out on the data of both studies
33
34
35

                                                22

-------
                   - DRAFT  4,  MARCH 31,  1999,  DO  NOT QUOTE  OR CITE-

 1
 2                                 REFERENCES CITED
 3
 4      *Brightman, H.S., Womble,  S.E., Ronca, E.L., and Girman, J.R. 1996. Baseline
 5            Information on Indoor Air Quality  in Large Buildings (BASE '95). Proceedings of
 6            Indoor Air'96. Vol.  3, pp.1033-1038.
 7
 8      *Brightman, H.S, Womble,  S.E., Girman,  J.R.,  Sieber, W.K., McCarthy, J.F., Buck, R.J.,
 9            and Spengler, J.D. 1997. Preliminary  Comparison of Questionnaire Data From Two IAQ
10            Studies: Occupant and Workspace Characteristics of Randomly Selected Buildings and
11            "Complaint" Buildings. Presented at the Healthy Buildings/IAQ 97 Conference in
12            Washington, DC, September 1997.
13
14      Bluyssen, P.M., Fernandes, EDO, Fanger, P.O., Groes, L.,  Clausen, G., Roulet, C.A.,
15            Bernhard, C.A.,  and  Valbjorn, O. 1995. European Audit Project to Optimize Indoor
16            Air Quality and Energy Consumption in Office Buildings.  TNO Building and
17            Construction Research, March 1995.
18
19      Building Management Education Foundation. 1988.  Report of a Survey of Building
20            Hygienic conditions (Japanese).  Building Management Education Foundation, March
21            1988.
22
23      DOE, 1995. Commercial Buildings Energy  Consumption Survey.  U.S. Department of
24            Energy. Website location: http://www.eia.doe.gov/emeu/cbecs/char95.
25            March 29,1999.
26
27      EPA, 1987, The Total Exposure Assessment Methodology  (TEAM) Study: Summary and
28            Analysis: Volume 1, USEPA Office  of Research and Development, Washington, DC,
29            EPA/600/6-87/002a,  June 1987.
30
31
                                              R-l

-------
                   -  DRAFT 4,  MARCH 31,  1999,  DO NOT QUOTE  OR CITE-

 1      EPA, 1994, A Standardized EPA Protocol for Characterizing Indoor Air Quality in Large
 2            Office Buildings, Indoor Air Division, Office of Radiation and Indoor Air, USEPA,
 3            Washington, DC and Atmospheric Research and Exposure Assessment Laboratory,
 4            Office of Modeling, Monitoring Systems, and Quality Assurance, USEPA, Research
 5            Triangle Park, North Carolina. June 1, 1994.
 6
 7      EPA, 1996. The Particle Team, PTEAM Study: Analysis of the Data, Final Report,
 8            Volume 3, USEPA Office of Research and Development, Washington, DC,
 9            EPA/600/R-95/098, August 1996.
10
11      EPA, 1999. TIME Study - Overview and Summary Data for Six Initial Buildings. Website
12            location: http://www.epa.gov/iaq/base/summary.html
13
14      Fortmann, R., Clayton, R., Highsmith, V.R., and Nelson, C.J. 1994. The U.S. EPA/ORD
15            Large Building Study: Results of the Initial Survey of Randomly Selected GSA
16            Buildings. Presented at the 1994 Air and Waste Management Association
17      Symposium. Website location: http://www.epa.gov/iaq/base/awma.html.
18            March 29, 1999.
19
20      *Girman, J.R, Womble, S.E., and Ronca, E.L. 1995. Developing Baseline Information on
21            Buildings and Indoor Air Quality (BASE '94): Part II - Environmental Pollutant
2 2            Measurements and Occupant Perceptions. Presented at Healthy Buildings '95,
23            September 11-14, 1995.
24
25      *Hadwen, G.E., McCarthy, J.F., Womble, S.E., Girman, J.R., and Brightman, H.S. 1997.
2 6            Volatile Organic Compound Concentrations in Office Buildings in  the Continental
2 7            United States. Presented at the Healthy Buildings/IAQ 97 Conference in
28            Washington, DC, September 1997.
29
30      Mendell, M.J., Fisk, W.J., Deddens, J.A., Seavey, W.G., Smith, A.H., Smith, D.F., Daisey,
31            J.M., and Goldman, L.R. 1996. Elevated Symptom Prevalence Associated with
32            Ventilation Type in Office Buildings. Epidemiology, Vol.	pp.  583-589
33
34
                                               R-2

-------
                   - DRAFT  4,  MARCH 31,  1999,  DO NOT QUOTE  OR  CITE-

 1      SAB, 1996. An SAB Report: The Cumulative Exposure Project, A Review of the Office of
 2            Planning, Policy and Evaluation's Cumulative Exposure Project (Phase 1) by the
 3            Integrated Human Exposure Committee, September 1996, EPA-SAB-fflEC-ADV-96-
 4            004.
 5
 6      SAB, 1999.  An SAB Report: Review on Disproportionate Impact Methodologies, A Review
 7            by the Integrated Human Exposure Committee (THEC) of the Science Advisory Board
 8            (SAB). EPA-SAB-fflEC-99-007.
 9
10      Sundell, J. and Bornehag, C. 1998. (Personal communication with C. Weschler in 1998)
11
12      Ten Brinke, J.T., Selvin, S., Hodson, A.T., Fisk, W. J., Mendell, M.J., Koshland, C.P., and
13            Daisey, J.M. 1998. Development of New Volatile Organic Compounds (VOC)
14            Exposure Metrics and Their Relationship to "Sick Building Syndrome" Symptoms.
15            Indoor Air,  Vol. 8, pp. 140-152.
16
17      *Womble, S.E., Axelrad, R., Girman, J.R., Thompson, R., and Highsmith, R. 1993. EPA
18            BASE Program - Collecting Baseline Information on Indoor Air Quality. Proceedings
19            of Indoor Air  '93, Vol. 1, pp. 821-825.
20
21      *Womble, S.E., Girman, J.R., Ronca, E.L., Axelrad, R., Brightman, H.S., and McCarthy, J.F.
22            1995. Developing Baseline Information on Buildings and Indoor Air Quality (BASE
23            '94): Part I- Study Design, Building Selection, and Building Descriptions. Presented at
24            Healthy Buildings '95, September 11-14, 1995.
25
26      *Womble, S.E., Ronca, E.L., Girman, J.R., and Brightman, H.S. 1996. Developing Baseline
27            Information on Buildings and Indoor Air Quality (BASE '95) in IAQ 96,  Paths to
28            Better Building Environments, Ed. Kevin Y. Teichman. American Society of Heating,
2 9            Refrigerating and Air-Conditioning Engineers, Inc.
30
31
32
33
34            *     Material reviewed by the Committee prior to the meeting.
                                              R-3

-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28



BASE
BSI
CFM
CO
CO2
CV
DOE
GPRA
HVAC
IAQ
fflEC
NHANES
NHEXAS
NIOSH
ORD
PM
PTEAM
QA/QC
SAB
SF6
TEAM
TIME
VAV
VOCs

 - DRAFT  4, MARCH  31,  1999, DO NOT QUOTE OR CITE-

APPENDIX A - ACRONYMS AND ABBREVIATIONS
         Building Assessment Survey Evaluation
         Building Symptoms Index
         cubic feet per minute
         carbon monoxide
         carbon dioxide
         constant volume
         Department of Energy
         Government Performance and Results Act
         Heating, Ventilation and Air-Conditioning
         Indoor Air Quality
         Integrated Human Exposure Committee
         National Health and Human Nutrition Examination Survey
         National Human Exposure Assessment Survey
         National Institute for Occupational Safety and Health
         Office of Research and Development
         particulate matter
         Particle Total Exposure Assessment Methodology
         Quality Assurance/Quality Control
         Science Advisory Board

         Total Exposure Assessment Methodology
         Temporal Indoor Monitoring and Evaluation Study
         variable air volume
         Volatile Organic Compounds
                            A-l

-------
- DRAFT 4, MARCH 31,  1999, DO NOT QUOTE OR CITE-




      APPENDIX B - BASE Core Parameters
                    B-l

-------
- DRAFT 4, MARCH 31, 1999, DO NOT QUOTE OR CITE-




     APPENDIX C - Proposed BASE Analyses
                     C-l

-------
                   -  DRAFT  4, MARCH 31,  1999, DO NOT  QUOTE OR CITE-

 1                  APPENDIX D - GPRA Strategic Goal 4, Objective 4
 2
 3
 4      By 2005, 15 million more Americans will live or work in homes, schools, or office buildings
 5      with healthier indoor air than in 1994.
 6
 7            More specifically, to reduce lung cancer, respiratory diseases, and other health
 8      problems, 11.5 million more Americans will be exposed to healthier indoor air in their homes
 9      by the mitigation of 700,000 homes with high radon levels, the construction of one million
10      homes with radon-resistant construction techniques, a reduction in the proportion of
11      households in which children 6 and under are regularly exposed to smoking from 27% in 1994
12      to 15%, and a reduction in the number of children and low-income populations exposed to
13      indoor air pollutants which worsen or trigger asthma episodes.  To reduce health problems in
14      the nearly 10 million children made ill annually from indoor air problems in schools,  15% of
15      the nation's schools will adopt good IAQ practices consistent with EPA's "Tools for Schools"
16      guidance. To reduce lAQ-related illness from contaminated air in the workplace, 5% of office
17      buildings will be managed with good IAQ practices consistent with EPA's "Building Air
18      Quality" guidance.
19
                                               D-l

-------
                   - DRAFT  4, MARCH 31, 1999,  DO  NOT QUOTE  OR CITE-

 l                                 DISTRIBUTION LIST
 2
 3
 4      Administrator
 5      Deputy Administrator
 6      Assistant Administrators
 7      Deputy Assistant Administrator for Science, ORD
 8      Director, Office of Science Policy, ORD
 9      EPA Regional Administrators
10      EPA Laboratory Directors
11      EPA Headquarters Library
12      EPA Regional Libraries
13      EPA Laboratory Libraries
14      Library of Congress
15      National Technical Information Service
16      Congressional Research Service

-------