Participant Reactions to Medical Screening: A Survey of Satisfaction With the C8 (PFOA) Health Project

Participant Reactions

to Medical Screening:

A Survey of Satisfaction

With the C8 (PFOA)

Health Project

Clare Malone

*, Gu¨lnaz C¸ı

g

*,

Phil Brown

, and Alan Ducatman

Abstract

We report participant perceptions of the 2005–2006 C8 Health Project, a massive medical monitoring effort in response to perfluorooctanoic acid (C8) in West Virginia and Ohio. The C8 Health Project consisted of a health survey (n¼ 69,030), blood testing for ten per- and polyfluoroalkyl substances, and 50þ laboratory tests (n¼ 66,899). A randomly selected subgroup was surveyed in 2007 on (1) demographics (2) satisfaction with the project, and (3) perceptions of outcomes such as contribution to personal/family, community health, and links to health outcomes. The response rate was 573/1500 (38.2 percent). Most (92.7 percent) characterized their participation experience as “excellent” or “good,” and most (96.2 percent) considered the project very “important,” “important,” or “moderately important.” No demographic variable predicted impor-tant changes in satisfaction or perception of project importance. We conclude that responses to the survey indicate strong positive assessments of project benefits.

Keywords

PFOA, PFAS, DuPont, C8 study, participant satisfaction

1

Northeastern University, Boston, MA, USA

2

Istanbul University - Cerrahpasa, Turkey

3

West Virginia University, Morgantown, WV, USA *Clare Malone and Gu¨lnaz C¸ıg are equal first co-author Corresponding Author:

Phil Brown, Northeastern University, 318INV 360 Huntington Ave, Boston, MA 02115, USA. Email: p.brown@northeastern.edu

NEW SOLUTIONS: A Journal of Environmental and Occupational Health Policy 2019, Vol. 29(2) 186–204 ! The Author(s) 2019 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/1048291119850109 journals.sagepub.com/home/new

Introduction

This paper analyzes participant satisfaction with the C8 Health Project, a large community-based environmental health survey (n¼ 69,030). The C8 Health Project arose from a Wood County West Virginia circuit court-supervised 2004 settlement agreement in principle (Leach, et al. v. E.I. E.I. DuPont de Nemours & Co.), negotiated between attorneys for a plaintiff class and attor-neys representing DuPont, concerning perfluorooctanoic acid (PFOA, C8HF15O2, also called “C8”) contamination of six drinking water districts

along the Ohio River in Ohio and West Virginia. Its deliverables in principle included a cash payment to persons whose drinking water was affected, health education and community projects for the benefit of the plaintiff class, water treatment in six affected water districts (including private wells) in West Virginia and Ohio, worker and community health studies, a science panel to evaluate the results of the studies, and a subsequent medical monitoring fund in the event that the health studies demonstrated a “probable link” between PFOA exposure and any human disease. This broad agreement in principle led to the C8 Health Project, an extensive demographic and health survey, including a large set of clinical laboratory tests, serum concentrations of PFOA and other perfluor-oalkyl acid exposures, a secure, central database for use by the C8 Science Panel, a group of three epidemiologists appointed by the court to analyze the data for health outcomes, and a West Virginia University open access website for rapid, visual communication of summary data. Enrollment in the C8 Health Project began in July 2005 and ended thirteen months later. A description of the settlement and project methods has been published elsewhere.1

This Appalachian experience with PFOA contamination has gained addition-al significance in recent years for two reasons. Water contamination by per- and polyfluoroalkyl substances (PFAS) chemicals has been recognized at many industrial sites, as well as military, airport, and fire training sites that have become known as sources of drinking water contamination due to the use of aqueous fire-fighting foam.2 A map of U.S. contamination sites is viewable online and includes Environmental Protection Agency (EPA) Unregulated Contaminant Monitoring Rule data (https://www.ewg.org/research/update-map ping-expanding-pfas-crisis#.W50OlJNJGMK). And, an increasing number of human health concerns have arisen about PFAS exposure. The C8 Health Project played a role in that increased understanding.

A little less than one year after the September 2004 agreement in principle, and only months after the final implementation details were hammered out, a massive enrollment project began. A little less than two years after the agreement in principle, enrollment was closed. Soon after, summary data concerning findings began to appear on the West Virginia University website; remnants of the exten-sive data portrayed can be reviewed and seen at https://www.hsc.wvu.edu/resoff/ research/c8/. C8 Science Panel studies using the C8 Health Project data began to

appear in 2009. An early list of C8 Science Panel peer publications is available at http://www.c8sciencepanel.org/publications.html. These publications provided a broader range of data (both corroborative and contradictory) compared to pre-viously published industry reports (i.e., Olsen et al.3; Olsen and Zobel4). They also provided scientific insight into the reliability of public positions of industry. (Contemporary industry positions are summarized by Environmental Working Group at https://www.ewg.org/research/credibility-gap-toxic-chemicals-food-packaging-and-duponts-greenwashing/dupont-press-quotes, concerning whether PFOA and other PFAS exposure represent a health risk for humans.)

Under the Community Projects rubric of the settlement, faculty and staff at West Virginia University School of Medicine created a participant satisfaction/ perception survey. It was mailed to a random subset of project participants approximately six months after survey enrollment closed. In the absence of literature concerning the outcomes of settlement agreements in toxic chemical contamination cases, our contribution is to demonstrate (1) the importance of satisfaction analyses for such cases, (2) the potential for high satisfaction in such cases, especially in light of a strong community engagement model, as well as (3) the influence that the data collected may have had on peer thinking about the topics studied.

Background

History of the DuPont Contamination and the C8 Health Project

This well-known contamination incident in the mid-Ohio Valley has been widely documented (especially in Lyons5), and since our purpose here is to analyze par-ticipant satisfaction, we provide only a brief historical description. Beginning in 1951, DuPont used PFOA in the production of TeflonVR

in Parkersburg, WV; Lyons5 reported how DuPont was early on aware of water contamination, animal toxicology, and human health effects which it did not report to affected residents or the EPA. Residents learned about that contamination much later, following the discovery process in a lawsuit based on cattle die-offs on land pur-chased for PFOA disposal. The legal discovery process uncovered evidence of DuPont’s concealment of important data, and much of the relevant data are available in an EPA docket (EPA AR-226). Litigation led by Cincinnati, Ohio attorney Robert Bilott led to the September 2004 settlement agreement between the plaintiff class and DuPont. Review of the actual testimony leading to the agreement in principle reveals that the settlement origins of the C8 Health Project5,6and especially its emphasis upon putting settlement assets into partici-pant enrollment and clinical data collection were heavily influenced by Harry Deitzler, a West Virginia attorney and a past president of the West Virginia State Bar. The testimony transcript is available at https://www.hpcbd.com/ dupont/Transcript-of-the-Final-Hearing.pdf. The transcript also demonstrates

participation of community leaders such as Joseph Kiger who still remain engaged in asking DuPont to address the history of environmental PFOA pollution (see, for example, the “Keep Your Promises” website including its news stories and document archive at https://keepyourpromisesdupont.com/).

When all details were worked out, participants were eligible if they resided, worked, or attended school in six affected water districts in West Virginia and Ohio (see map of the study area, Figure 1) with contamination above fifty parts per trillion. Brookmar Inc. was created limited liability company to design, publicize, and implement the survey project under court supervision. It was led by A. Paul Brooks, a Parkersburg, WV physician and former president of the West Virginia Medical Licensing Board. His partner was Arthur A. Maher, a former local area hospital leader. Dr. Brooks sometimes asked himself (oral communication with Alan Ducatman) if DuPont representatives may have assumed that a local family physician would be unable to create an effective survey, enroll a substantial population, and capture relevant data in a relatable data warehouse for an effective environmental health project. Brookmar’s C8 Health enrolled 69,030 participants, estimated to include greater than eighty percent of adults in affected water districts.1 Participants were tested at four mobile sites. Both the project’s health survey (http://www.hsc.wvu.edu/media/ 4542/c8-health-project-questionnaire-v7-29-05.pdf) and the list of laboratory

Figure 1. Map of C8 project study area. Source: Shin et al.7

test results received by participants (http://www.hsc.wvu.edu/resoff/research/c8/ results/clinical-laboratory-tests/) are available online. Dr. Brooks insisted that each participating individual receive their mailed personal results. He also cre-ated emergency flags for patients whose laboratory data indiccre-ated a need for emergency medical care. During the thirteen-month-long enrollment period, approximately twelve individuals were telephoned and advised to present them-selves to their physician or the emergency department immediately. The precip-itating causes were previously undiagnosed diabetes, blood disorders such as undiagnosed leukemia, and potentially life-threatening electrolyte abnormality (personal communication with Alan Ducatman).

The implementation of the settlement agreement also established a panel of three independent epidemiologists known as the “C8 Science Panel.” There had been plaintiff attorney discussion about including a toxicologist and a clinician, but the final implementation agreement was for three epidemiologists. The sci-ence panel role was to evaluate summary data for associations to health out-comes and to make subsequent decisions about the presence or absence of “probable links” between exposure and disease. This work started with evalu-ation of the data warehouse created by Troy Young, database consultant to the project.

In addition to health reports, the science panel created an exposure model from the data.7It also planned and successfully reenrolled a substantial subset of the C8 participant population for longitudinal follow-up, investigating out-comes such as cancer.8Probable links eventually reported by the science panel (http://www.c8sciencepanel.org/panel.html) included pregnancy-induced hyper-tension, hypercholesterolemia, thyroid disease, kidney cancer, testicular cancer, and ulcerative colitis. People with the “probable link” conditions were able to sue DuPont for compensation without having to adjudicate the epidemiological basis each time; more than 3500 have done so.6While the adjudicated deliver-ables included “probable links,” this legal term of art is not amenable to eval-uation for its contribution to subsequent peer thinking. More than sixty peer-reviewed articles have been published based on data collected in the C8 Health Project, and the project is well-described in the literature.1,9 Brookmar also commissioned West Virginia University to conduct a random, anonymous survey addressing participant perceptions of satisfaction (the “Satisfaction Survey”) reported herein in order to inform future efforts for community inter-ventions such as health studies and medical monitoring.

The goal of this paper is to examine satisfaction/perceptions reported by C8 Health Project participants and to evaluate if there are differences according to major demographic variables. We also sought a surrogate measure to report on whether the data have been influential within the scientific community. Such information can be useful for planning future settlements, for consideration of projects that may involve medical monitoring, for future epidemiological

studies, as well as for the community engagement that can be engendered in such efforts.

An explanation for the ten-year delay from data gathering to analysis is worth mentioning. This represents another PFAS publication in a decade-long ongoing effort by one of the participating authors (Ducatman) to detail PFAS exposures and excretion, and now the social aspects of community perception. This report provides support from social scientists so that the initially envi-sioned description of community perception now adds a social predictor assess-ment and historical content.

Data and Methods

The satisfaction survey was developed in 2006 by the West Virginia University Department of Political Science and the Department of Community Medicine, with support from the Department of Statistics, under a contract with Brookmar, Inc. It was mailed in January 2007, approximately six months after the last C8 Health Project participant was enrolled, to a randomly selected sample of 1500 participating adults. Responses were anonymous. The West Virginia University institutional review board deemed the satisfaction survey and evaluation process exempt.

C8 survey questions were informed by results of three 2005 focus group meetings conducted by Salter & Co, an independent media consulting firm. The focus groups were held immediately after the class-action settlement was finalized and before the C8 Health Project team began its health survey work. Meetings were not heavily scripted and started with open-ended inquiries such as “What do you know about C8?” and “What do you know about the settle-ment and the planned C8 Health Project?” Participants’ responses revealed a strong concern for maintaining individual privacy, an anticipated concern for which extensive planning was underway. At town meetings, attended by hun-dreds of local residents, Brookmar personnel emphasized that their role was to administer the survey to qualified participants and to collect the data and put it into database format, but not to analyze the data or to provide follow-up healthcare. Data analysis would be the role of an independent scientific group (soon to be called the C8 Science Panel). Any healthcare needs detected would be between the participant and their personal clinical provider. Brookmar also held open meetings for local healthcare providers.

Based on the range of focus group and town meeting responses and inter-actions, satisfaction survey questions were also created to address the partici-pant intake experience, including fairness of the registration and documentation process, participant ability to complete the survey online versus in person at the intake sites, the participants’ reception and recall of their individual results, the participants’ evaluation of whether privacy had been conserved, as well as over-all perceptions of benefits to personal health and satisfaction with the overover-all C8

Health Survey process. The satisfaction survey also posed questions intended to evaluate participant perceptions about community aspects of the C8 Health Project process. For example, were there adequate public affairs efforts made calling attention to the opportunity to participate? Were participants aware of the presence of the independently appointed science panel and their specific role concerning epidemiologic analyses of the collected data?

To prepare potential respondents, researchers distributed (via website and news outlets) a prenotification that a random sample would receive a participant satisfaction survey that would include a return stamped envelope for those selected to participate. The satisfaction survey questions were pilot tested on students and reviewed for literacy level. Questionnaires were mailed in January 2007 to 1500 adult C8 Health Survey participants with unique addresses who were selected by a randomization sequence based on a random number gener-ator using JMP software (SAS Institute, Cary, NC).

Satisfaction survey data were analyzed using Stata software. Simple linear regression analyses were run with each of the following as independent varia-bles: age, income, gender, education level, test site, and whether or not the respondent was a parent. These linear regressions were performed for all survey questions. Multivariate analysis investigated how age, income, education, and gender influenced respondents’ perception of the following when considered together: Brookmar’s public awareness campaign, the benefit (or absence of benefit) of the class-action lawsuit, overall experience with the study, the sur-vey’s importance to overall health, confidence that the science panel would find a link between C8 and human illness, awareness of the science panel, and belief that C8 contributed to family illnesses. The uniformly positive responses do not suggest an influence of multiple analyses.

Results

Descriptive Statistics

Response rate for the satisfaction survey among a random sample of C8 Project participants was 38.2 percent (573/1500), comparable to the 33 percent response reported by Giannini et al.10 for PFAS serum concentration feedback in University of Cincinnati studies. Some respondents skipped some questions; Table 1 indicates the subtotal of classifiable responses to selected questions. Median age range was the age-group forty-six to fifty-five years, influenced by the exclusion of minor participants. The majority of respondents were women (58.2 percent), compared to 52.5 percent of female adult participants in the C8 Health Project. Most survey participants (82.8 percent) had not graduated from college, comparable to 87.2 percent of participants in the entire C8 Health Project who had not graduated from college.1

Table 1. Demographic Characteristics of Satisfaction Survey Respondents, n¼ 573. Frequency % Age (years) 25 and under 34 6.1 26–35 69 12.4 36–45 80 14.4 46–55 136 24.5 56–65 118 21.2 66–75 76 13.7 76 and over 43 7.7

Did not answer 17

n (classifiable) 556

Total 573

Education

Less than a high school diploma 51 9.1

High school diploma or GED 228 40.9

Some college or associate degree 183 32.8

Bachelor’s degree 57 10.2

Advanced or graduate degree 39 7.0

Did not answer 15

n (classifiable) 558

Total 573

Gender

Male 236 41.8

Female 328 58.2

Did not answer 9

n (classifiable) 564 Total 573 Income Less than $10,000 37 7.2 $10,000–$19,999 66 12.8 $20,000–$29,999 63 12.2 $30,000–$39,999 63 12.2 $40,000–$49,999 49 9.5 $50,000–$59,999 48 9.3 $60,000–$69,999 40 7.8 $70,000–$79,999 31 6.0 $80,000–$89,999 27 5.2 $90,000–$99,999 15 2.9 $100,000 or more 33 6.4

Not sure or don’t know 44 8.6

Did not answer 57

n (classifiable)¼ 516

Respondents indicated they were satisfied with the monitoring study on var-ious levels (Table 2). Most (89.6 percent) had positive perceptions of the fairness of the eligibility process including scanning/archiving documents in order to prove eligibility, although 9.1 percent of respondents also reported the need to make more than one visit to prove eligibility. Most (78.6 percent) considered the C8 Health Project survey “very easy” or “somewhat easy” to complete, and 74.4 percent reported they completed it online. Most (65.7 percent) considered it very or somewhat easy to schedule their visit. Only 1.4 percent reported the experience of having their blood sampled to be “below average” or “poor.” Most (97.5 percent) respondents recalled receiving a mailed report of personal PFAS serum concentration and clinical laboratory values. Most (71.9 percent) respondents were confident that their privacy had been preserved, and 25.8 per-cent reported they “did not know” if the survey results preserved their privacy. Most (92.7 percent) respondents characterized their overall experience as either “excellent” or “good,” and most (96.2 percent) considered the project process to be “very important,” “important,” or “moderately important” to their health.

Concerning community aspects of the experience, most (90.7 percent) rated Brookmar, Inc., which created and conducted the C8 Health Survey, as excellent or good (Table 2). Although most (82.6 percent) respondents indicated aware-ness of the C8 Science Panel of independent epidemiologists, there was less awareness (51.1 percent) of the science panel’s specific upcoming role of evalu-ating probable links to diseases.

Bivariate Analysis

Given the overall very positive assessment of so many components of the research experience, major differences across categories of respondents are not expected. Nevertheless, bivariate relationships can uncover useful hypotheses about demographic influences. Supplementary Tables S1 to S4 (available on the journal website) show coefficients only when statistically significant. Location of field sites (Table S1) showed no effect of proximity to the DuPont facility in respondents’ perception that Brookmar’s work successfully brought public awareness of the study (85 to 93 percent agreeing). Nor was there a geographic difference in residents’ awareness of the science panel, with 79 to 86 percent being aware without regard to location. Residents from the most distant test site, Point Pleasant, were slightly more likely to think that the science panel could detect “probable links” to human illness conditions (64 percent vs. 47 to 54 percent from other sites, p¼ .024), however (Table S1).

In advance of published studies from the dataset, most residents responded that they did not know if PFOA exposure had affected their health or the health of their family (Table 1), and there were only small differences among sites for affirmative responses (range 16 to 19 percent) despite large differences in the degree of water contamination.

Table 2. Aggregate Responses. Rating Frequency Percentage (%) Public awareness of C8 Health Survey Excellent 247 43.3 Good 254 44.5 Average 49 8.6 Below average 4 0.7 Poor 4 0.7

Not sure or don’t know 13 2.3

Did not answer 2

n¼ 573

Fairness of proving survey eligibility

Very fair 408 71.4

Somewhat fair 101 17.8

Neither fair nor unfair 36 6.3

Somewhat unfair 17 3.0

Very unfair 6 1.1

Did not answer 5

n¼ 573

Successful first attempt to prove eligibility

Yes 512 90.9

No 51 9.1

Did not answer 10

n¼ 573 Survey ease of completion Very easy 237 41.7 Somewhat easy 210 36.9 Average 94 16.5 Somewhat difficult 26 4.6 Very difficult 2 0.4

Did not answer 4

n¼ 573

Method of

survey completion

Online (using a computer by myself) 322 56.4

Online (using a computer with some help)

103 18.0

Paper copy by myself 122 21.4

Paper copy with some help 23 4.0

Never filled out a survey 1 0.2

Did not answer 2

n¼ 573 Ease of scheduling survey visit Very easy 246 43.0 Somewhat easy 130 22.7 Average 79 13.8 Somewhat difficult 94 16.4 Very difficult 23 4.0

Did not answer 1

n¼ 573

Table 2. Continued. Rating Frequency Percentage (%) Experience of giving blood sample Excellent 398 69.6 Good 142 24.8 Average 24 4.2 Below average 7 1.2 Poor 1 0.2

Did not answer 1

n¼ 572 Recalled receiving mailed results report Yes 555 97.5 No 12 2.1 I don’t remember 2 0.4

Did not answer 4

n¼ 573

Confidence that priva-cy was preserved

Yes 410 71.9

No 13 2.3

Not sure or don’t know 147 25.8

Did not answer 3

n¼ 573

Brookmar Inc.’s per-formance compared to expectations

Much better than I expected 221 39.0

Somewhat better than I expected 164 28.9

What I expected 171 30.2

Somewhat worse than I expected 8 1.4

Much worse than I expected 2 0.4

Don’t know or not sure 1 0.2

Did not answer 6

n¼ 573

Benefit to health Very important 244 44.1

Important 223 40.3

Moderately important 65 11.8

A little important 12 2.2

Not at all important 8 1.5

Don’t know/not sure 1 0.2

Did not answer 20

n¼ 573

Overall experience Excellent 273 47.9

Good 250 43.9

Average 42 7.4

Below average 4 0.7

Poor 1 0.2

Did not answer 3

n¼ 573

Impact of having children. There was a statistically significant difference (Table S2; p¼ .026) between how respondents with and without children estimated the survey’s overall importance for their health. Compared to those without chil-dren, respondents with children were less likely to consider the overall survey experience as either “very important” or “important” to their health (78.2 per-cent vs. 87.1 perper-cent, respectively). While statistically significant, the actual dif-ference is not large.

Education. Awareness of the science panel was higher among respondents who had education after high school (87 to 91 percent), compared to high school or less (77 to 78 percent), p¼ .017 (Table S3). Respondents who had attended at least some college were least likely to predict that the science panel would find a link between PFOA exposure and human illness (p¼ .018). Level of education did not predict a respondent’s belief whether exposure had a personal impact on their health/their family’s health, or how respondents valued the C8 Health Project (Table S3). However, those with high school or less education were more likely to think the C8 Health Project will help the community.

Table 2. Continued. Rating Frequency Percentage (%) Awareness of Science Panel Yes 464 82.6 No 98 17.4

Did not answer 11

n¼ 573

Awareness of science panel role: epidemi-ology and linking C8 with illness

Yes 247 51.1

No 53 11.0

Not sure or don’t know 183 37.9

Did not answer 90

n¼ 573

Personal benefit of class settlement

Agree strongly 125 22.2

Agree somewhat 217 38.5

Neither agree nor disagree 152 27.0

Disagree somewhat 45 8.0

Disagree strongly 23 4.1

Don’t know or not sure 1 0.2

Did not answer 10

n¼ 573

C8 has caused any health problem for you and your family

Yes 99 17.4

No 138 24.2

Not sure or don’t know 333 58.4

Did not answer 3

Age. The percentage of respondents who answered that PFOA had caused health problems for them and their family was higher among the younger age groups: age under thirty (27 percent), age thirty-one to sixty (20 percent), and 10 percent for those over sixty years (Table S4; p¼ .003). Age was modestly asso-ciated with respondents’ confidence that the science panel could find a link between C8 and human illness, with a range of 45 to 67 percent. The majority (84 to 94 percent) of respondents from all age categories rated Brookmar’s efforts to bring public awareness to the C8 Health Project as above average, suggesting the efforts were subjectively successful across all age groups.

Most respondents considered that the C8 Health Project would eventually benefit the people who live or had lived in the affected water districts. Interestingly, respondents aged twenty-one and younger were substantially more likely (80 percent vs. 64 percent in all other age groups) to expressly see the project’s benefits (Table S4; p¼ .026), but a number of participants aged twenty-one and younger is constrained by the survey design (n¼ 8).

Income. Respondents’ reported income level had no impact on responses to this satisfaction survey (results not shown).

Multivariate Analysis

Given the homogeneity of responses, we did not expect anything significant to come from multivariate analysis. In a regression model with age, income, and education as dependent variables, only age was significant, and the overall R2 was low (R2¼ .037), even though significant.

Discussion and Conclusion

Overall, respondents expressed strongly affirmative opinions of the project pro-cess, of the entity Brookmar that conducted the survey, and of the individual exposure and clinical laboratory data that they received by mail. There were many areas with strongly positive responses and few areas with more than a small number of negative responses. That 74.4 percent reported they were able to complete the C8 Health Survey online is an interesting result for 2005–2006 in a Central Appalachian region consisting of small metropolitan and rural areas. Participants were very likely to believe that the C8 Health Project was effective in its intake and could contribute to both their personal health and community health. The vast majority of participants recalled receiving their personal labo-ratory data by mail. Our finding that 82.6 percent of respondents knew of the C8 Science Panel of independent epidemiologists, yet only 51.1 percent knew of the panel’s specific role of evaluating probable links to diseases, is not surprising since the satisfaction survey was conducted in January 2007, and the first peer-reviewed reports by the C8 Science Panel had not yet been published, and

“probable links” for six conditions (as well as null findings for many other conditions) had not yet been posted.

A short list of precedent responses from surveyed medical monitoring populations provides context for our findings. Gerhardstein and Brown11 inter-viewed an opportunity sample (N¼ 35) of community member participants from a Fernald Medical Monitoring Project (n¼ 9179). During production years, the Fernald, Ohio, nuclear weapons site released more than one hundred tons of uranium dust into the air and seventy tons into a nearby river. The ensuing liti-gation settlement included persons who lived or worked full time within five miles of the Fernald plant for a continuous two-year period, between 1952 and 1984. In interviews, Fernald class members generally agreed that long-term medical mon-itoring was an appropriate and positive response to their community situation: more than half (62 percent; n¼ 21) of respondents considered medical monitoring the most important component of the settlement, while 94 percent (n¼ 33) ranked it among the top two. Furthermore, 86 percent (n¼ 30) of respondents thought that the monitoring was a good use of settlement money, and 82 percent (n¼ 28) indicated that they would rather have settlement funds be used for medical mon-itoring than paid out to litigants.11The researchers concluded that “The Fernald experience suggests that medical monitoring in the short-term may cause a spike in perception of risk, but in the long-term provides reassurance and peace of mind.” The survey of C8 Health Project participants presented in this paper con-trasted to this Fernald survey because it was administered much nearer to the start of the project. Like the smaller Fernald survey, the C8 Health Survey data were affirmative for perceived personal and community value.

Giannini et al.10mailed satisfaction surveys to participants of three cohorts who had received PFAS serum concentrations results in University of Cincinnati studies. Those cohorts were not originally designed as PFOA studies; two groups were part of the Breast Cancer and the Environment Research Program (a multisite project funded by National Institute of Environmental Health Sciences and National Cancer Institute), and one was an extension of the Fernald study discussed above. These findings provide a more direct basis for comparison of our satisfaction findings because they dealt with PFOA and involved people living farther downstream of the very prominent DuPont con-tamination site. The main satisfaction outcome for the three cohorts in Giannini’s study was participants’ responses to receiving report back of their personal exposure levels. The overwhelming majority (96 percent) were pleased that researchers had sent the report, 86 percent found that the results were easy to understand, and 87 percent responded that the enclosed fact sheet was helpful in answering questions. This high degree of satisfaction was attributed to exten-sive community engagement in these cohorts: Breast Cancer and the Environment Research Program mandates a community engagement core involving breast cancer activists and study participants, and the Fernald

cohort has long had its own institutional review board to ensure that research initiatives are useful to members of the class.

Responses to the satisfaction survey for the C8 Health Project were, generally as, or even more, affirmative than these comparable findings, but not uniformly so. Some participants expressed uncertainty about the preservation of personal privacy, with the 71.9 percent expressing confidence that privacy would be pre-served and the remainder either uncertain (25.8 percent) or confident that it would not be preserved (2.3 percent). Focus groups had identified preservation of privacy as the most common concern of potential participants a year earlier, particularly for participants who worked for or had relatives who worked for DuPont, the dominant employer in the area. The survey data suggest that this concern was resolved a year later for most but not all who did participate. From a community perspective, the biggest area of uncertainty was about the role of the science panel, with 51.1 percent of respondents confident that they under-stood the science panel role, and the remainder mostly uncertain. We do not know to what degree the subsequent reports of “probable links” to disease conditions or the substantial and still emerging peer publication record has altered those early prepublication perceptions.

Bivariate analysis revealed few differences, despite the multiple comparisons. Having children was not relevant for belief that PFOA causes disease. While people with post-high school education were more likely than others to be aware of the science panel, and more likely to predict that it would find links between PFOA and disease, they were less convinced that PFOA personally affected them and their families. Younger and less educated people were more likely to expect the C8 Health Project to be of benefit to themselves or the community and to think PFOA caused health problems for them and their families. Income did not predict anything. Almost all of these stratification differences are subtle, statistically significant but not based on large differences in percent satisfaction/ dissatisfaction responses, because the primary outcome was widespread and strong satisfaction with many aspects of the C8 study.

The timing of the effort is important to understanding the results of this satisfaction survey. The reported perceptions were about six months after initial enrollment was complete and therefore six to eighteen months after individual biomarker of exposure and clinical testing results were known to participants. In this setting, it is not surprising that 97.5 percent were able to recall seeing their results, but fewer had an understanding of what the science panel was about to do. Further, demographic descriptions of the population, associations of PFOA exposure with laboratory findings such as total and LDL cholesterol in adults and children, with “liver functions” such as alanine aminotransferase, and with uric acid, as well as perturbations in thyroid function, were just starting to become visible in summary online graphics hosted at West Virginia University but were not yet reported in peer-reviewed literature.

There are clear limitations of this survey research report. The 38.2 percent survey response rate is a potential limitation. While it is substantially higher than most mailed survey response rates,12and while the respondents are shown in this paper to be demographically similar to the participants in the larger survey, the response rate still leaves open the possibility of selection bias based on either positive or negative views of the C8 Health Project. In addition, those who moved without a forwarding address are not represented. We do not know what the respondents would conclude about the same experience more than a decade later, and this is the primary study limitation. This paper is about the perceptions of the project prior to the substantial productivity and scientific influence of the 38 highly cited peer papers (as of February 2019) by the three members of the C8 Science Panel. Our data suggest that participants mostly could not envision the substantial scientific influence of the work of these three epidemiologists. It is unclear if anyone could have envisioned the extent and influence of the science panel’s findings.

There are some important take-home messages for those conducting future settlement agreements, environmental health studies, and medical monitoring projects. First, the 74.4 percent of responders who reported filling out the survey online in 2005–2006, in a relatively poor and rural area, is impressive and suggests that online surveys of this type are effective. The satisfaction survey methods relied on a mailed survey, so it is possible that our results may even understate the percent who completed the survey online. It should be noted that C8 Health Project surveys completed online were reverified with participants by a project nurse on-site. This was done at the time of official project enrollment and associated blood specimen drawing, a step that took approximately thirty minutes per participant, but was less time-consuming than filling out the survey in person. Any edits were made by the participant in the presence of the nurse, and the edits were done online for automated uploads. This step may have increased participant confidence in their online submissions. In contemporary news stories, part of the high online penetrance has been informally attributed to community efforts, with neighbors helping neighbors to get into the online survey and with community libraries pitching in to provide access. Although we do not have direct data, it is reasonable to believe that online capability to complete environmental health surveys and medical monitoring participation could be higher in U.S. venues today, especially if similar supports are in place, and especially for PFAS because of the growing national awareness of PFAS contamination.

A second take-home message is not unique to our data. Our data and the small number of other environmental health monitoring projects that have been inves-tigated with subsequent participant data show a high degree of satisfaction with and/or belief in the value of participation in the project. Nonetheless, knowing is better than assuming, and our follow-up survey is larger and more comprehensive than previous efforts. The finding of participant satisfaction does not endorse complacence. We continue to advocate for the information that postactivity par-ticipant surveys can add to future health survey and medical monitoring projects.

A third lesson from the C8 Health Project, supported by the participant reactions, is that local resources can create and successfully conduct health survey projects that are of clear value to the community. If summary interpretation is also a part of the process, then expert analysis such as that provided by the C8 Science Panel is also crucial. Much more is now known about the human exposure/outcome associations for PFOA specifically, and PFAS in general, than was known in 2007, when this survey was performed. A random follow-up survey that reflected on the community perception of the impact of the subsequent clearly influential research would add to our understanding of the project’s local impact.

Perhaps the most important take-home point is that settlement agreements can be crafted to make a difference in affected communities and beyond. From the perspective of additions to the peer-review literature and of community satisfac-tion with the process, the C8 Health Project has been an affirmative model of what can be achieved by settlement agreements. As Wones et al.13note, the Fernald program was likely the first and largest ongoing program related to class-action litigation from environmental contamination, and it led to strong health benefits. The C8 Project is larger and contains provisions for ongoing medical monitoring in specific circumstances but is no longer an ongoing cohort generating new data, although reenrollment is both possible and scientifically desirable. Ongoing research from data previously obtained does continue.

It remains to be seen if the next steps relating to PFAS-exposed populations can be as affirmative. In the immediate region, there are other less affected water systems that might now be considered for intervention based on evolving views of policy which are in turn based on science generated since 2005–2006, including science from the C8 Health Project. More broadly, it is clear that other PFAS exposure populations will be enrolled in the future, including some U.S. popula-tions affected by drinking water contamination at U.S. Department of Defense sites. The overall “research-right-to-know” framework for environmental health studies calls for democratic, participatory report back in a community-based participatory research model.14 The C8 Health Project represents a valuable model for such procommunity environmental health medical monitoring, and it also provided valuable new science.

Acknowledgments

The authors thank Grace Poudrier for help in editing the article and Melanie Smith for help in formatting.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Alan Ducatman was coinvestigator for the 2007 satisfaction survey. In addition, he has provided consultation to a plaintiff class seeking medical monitoring. No other authors have any conflicting interest.

Funding

The author(s) received no financial support for the research, authorship, and/or publi-cation of this article.

ORCID iD

Phil Brown https://orcid.org/0000-0001-7627-5228

Supplemental Material

Supplemental material for this article is available online.

References

1. Frisbee SJ, Brooks AP Jr, Maher A, et al. The C8 Health Project: design, methods, and participants. Environ Health Perspect 2009; 117: 1873–1882.

2. Richter L, Cordner A and Brown P. Non-stick science: sixty years of research and (in)action on fluorinated compounds. Soc Stud Sci 2018; 45: 691–714.

3. Olsen GW, Butenhoff JL and Zobel LR. Perfluoroalkyl chemicals and human fetal development: an epidemiologic review with clinical and toxicological perspectives. Reprod Toxicol2009; 27: 212–230.

4. Olsen GW and Zobel LR. Assessment of lipid, hepatic, and thyroid parameters with serum perfluorooctanoate (PFOA) concentrations in fluorochemical production workers. Int Arch Occup Environ Health 2007; 81: 231–246.

5. Lyons C. Stain-resistant, nonstick, waterproof, and lethal: the hidden dangers of C8. Westport, CT: Greenwood Publishing, 2007.

6. Judge M, Brown P, Brody JG, et al. The exposure experience: participant responses to a biomonitoring study of perfluorooctanoic acid (PFOA). J Health Soc Behav 2016; 57: 333–350.

7. Shin H, Vieira VM, Ryan PB, et al. Retrospective exposure estimation and predicted versus observed serum perfluorooctanoic acid concentrations for participants in the C8 Health Project. Environ Health Perspect 2011; 119: 1760–1765.

8. Barry V, Winquist A and Steenland K. Perfluorooctanoic acid (PFOA) exposures and incident cancers among adults living near a chemical plant. Environ Health Perspect2013; 121: 1313–1318.

9. Steenland K, Jin C, MacNeil J, et al. Predictors of PFOA levels in a community surrounding a chemical plant. Environ Health Perspect 2009; 117: 1083–1088. 10. Giannini C, Herrick R, Buckholz JM, et al. Comprehension and perceptions of study

participants upon receiving perfluoroalkyl substance exposure biomarker results. Int

J Hyg and Environ Health2018; 221: 1040–1046.

11. Gerhardstein B and Brown P. The benefits of community medical monitoring at nuclear weapons production sites: lessons from Fernald. Envtl L Rep 2005; XXXV: 10530–10538.

12. Sinclair M, O’Toole J, Malawaraarachchi M, et al. Comparison of response rates and cost-effectiveness for a community-based survey: postal, internet and telephone modes with generic or personalised recruitment approaches. BMC Med Res Methodol2012; 12: 132.

13. Wones RG, Pinney SM, Buckholz J, et al. Medical monitoring: a beneficial remedy for residents living near an environmental hazard site. J Occup Environ Med 2009; 51: 1374–1383.

14. Morello-Frosch R, Brody JG, Brown P, et al. Toxic ignorance’ and the right-to-know: assessing strategies for biomonitoring results communication in a survey of scientists and study participants. Environ Health. 2009; 8: 6.

Author Biographies

Clare Malone graduated from Northeastern University in Boston, MA with a bachelor’s degree in sociology and anthropology. She worked as a research assistant for the Social Science Environmental Health Research Institute at Northeastern University.

Gu¨lnaz C¸ıg, MD, completed the Public Health Specialist Program in Istanbul University – Cerrahpasa, Faculty of Medicine, Department of Public Health. During her training, she was also selected as visiting scholar at West Virginia University Occupational and Environmental Health Department. She works as a public health specialist in Erzurum District Health Directorate in Turkey. Phil Brown is University Distinguished Professor of sociology and health science at Northeastern University, where he directs the Social Science Environmental Health Research Institute, the National Institute of Environmental Health Sciences T-32 training program, “Transdisciplinary Training at the Intersection of Environmental Health and Social Science,” the Community Outreach and Translation Core of Northeastern’s Children’s Environmental Health Center (CRECE), and the Research Translation Core and Community Engagement Core of Northeastern’s Superfund Research Program (PROTECT). He is the author of No Safe Place: Toxic Waste, Leukemia, and Community Action and Toxic Exposures: Contested Illnesses and the Environmental Health Movement and the coeditor of Social Movements in Health, and Contested Illnesses: Citizens, Science and Health Social Movements. He studies biomonitoring and household exposure and reporting back data to participants, chemical policy and health effects of flame retardants and PFAS (https://pfasproject.com/), and health social movements.

Alan Ducatman, MD, MS, is professor emeritus in the West Virginia University School of Public Health and a clinical consultant to industry, labor, nonprofit organizations, and government. His research focuses on environmental disease and disease prevention, including health communications to affected popula-tions. He has published actively concerning clinical population aspects of expo-sure to perfluoroalkyl substances.