Molecular data, essential for
genomics research, can be captured more efficiently in large-scale, population-based biobanks of genetic material rather than by individual studies. Biobanks also offer improved quality and reliability of genetic samples, and access through automated sample retrieval. However, it is challenging to adequately inform participants of the broad nature of the research and participation risks and benefits. In addition, recent studies suggest concerns about data sharing1
and return of research results,2,3
or future research topics (eg, stereotypical or potentially stigmatizing traits).4
We evaluated the interest in participating in a biobank and reasons for nonparticipation.
This research was part of the development of a biobank at Kaiser Permanente Northwest (KPNW), an integrated healthcare delivery system in Oregon and Washington. Eventually, we plan to invite all adult members to contribute blood samples that are left-over after routine clinical visits. We developed written recruitment materials at an eighth grade reading level that were vetted by a communications team, focus groups of KPNW members, KPNW's Institutional Review Board (IRB), and external advisory boards. We then surveyed 500 KPNW members to assess response to the materials. Research staff approached KPNW members in waiting rooms, and asked them to review the materials then complete a brief survey on demographics, willingness to participate, reasons for nonparticipation, and knowledge of the written materials. Most questions were categorical, although 2 open-ended questions asked about reasons for nonparticipation and concerns or questions about the biobank. The overall response rate was 62%. SAS Release 9 (SAS Institute, Cary, NC) was used for all statistical analysis including descriptive statistics, χ2
-tests, and logistic regression for multivariate modeling. Qualitative responses were summarized using theme analysis to identify the most salient concerns regarding participation in the biobank.5
KPNW's IRB approved this study.
>485,000 members, which is about 17% of the area's population. Medicare members represent about 12% of KPNW's membership. Members over 65 represent 11.5% of total membership, 2% are below 200% of the federal poverty level, and 13% are minorities. Survey respondents were broadly representative of the overall adult membership including 39% male, 20% with a high school education or less, 16% minorities, and 50% over 49 years old.
Overall, 69% agreed to contribute to the biobank, 20% were not sure, and 11% did not agree (). Respondents who were more likely to agree to contribute were males (75% males vs. 64% females, P
0.02), older (75% highest age group vs. 62% lowest age group, P
0.12), more affluent (77% highest income group vs. 58% lowest income group, P
0.02), and white (71% vs. 56% of minorities, P
0.002). However, in a multivariate model to adjust for confounding factors, only gender remained in the model after backward variable selection.
Characteristics of Survey Respondents by Agreement to Contribute to Biobank
In consideration of the NIH data sharing policy for genome-wide association studies (NIH Guide Notice: NOT-OD-07-008), we asked respondents if they would agree to allow researchers to post [their] information in a US government database that can be broadly shared with researchers worldwide to maximize the chances of new discoveries. Of those who would contribute to the biobank, 82% would also agree to have their information posted in a US government database.
The primary reason for nonparticipation was concern about confidentiality of health-related and genetic information. Respondents who said they were not sure or would not agree to the biobank were asked “what is the main reason you would not give your left-over blood.” The reasons for not agreeing to contribute to the biobank included concerns about confidentiality/privacy, lack of interest, confusion with a blood bank (eg, “I already give blood to the red cross”), concerns about sample use (eg, “I don't agree with all research, so I wouldn't agree to blindly participate in research”), lack of trust (eg, “I do not trust HMOs to protect my information and if you found something wrong, I would be afraid you would withhold care because of a known preexisting condition”), and religious or philosophical reasons. Respondents who said they were not sure cited all of the same concerns listed above. However, many respondents in this category also said they would like more information (eg, “feel that I need a little more info[rmation] before deciding,” “Need to think over what I read”).
In addition to the qualitative responses, we asked whether 5 specific concerns affected the decision not to give their residual blood: (1) security and confidentiality of their health information, (2) security and confidentiality of their genetic information, (3) results for their sample will not be shared or returned to them, (4) the potential to develop commercial products, and (5) disapproval of future research studies that could be conducted (respondents could select more than one). The most frequent issues were concern about security and confidentiality of health (73%) and genetic (61%) information, followed by disapproval of future research studies (40%). Concerns about commercialization and returning research results were not frequently selected (26% and 28%, respectively) as primary reasons for nonparticipation in this biobank.
Overall, 74% of respondents correctly answered
≥6 of 9 questions about the information presented in the recruitment materials, indicating a high level of comprehension. Respondents who were white (77% vs. 58% of minorities, P
0.004), college graduates (83% vs. 52% of high school education or less, P
0.0001), and more affluent (86% highest income group vs. 58% lowest income group, P
0.0001) answered more questions correctly. Respondents who agreed to contribute to the biobank were more likely to correctly answer the knowledge questions (71% vs. 51% who did not agree or 56% who were not sure).
Biobanks will benefit genomic research most if they broadly represent the underlying population. Our findings indicate that in a healthcare setting, data security and confidentiality are the primary reasons for nonparticipation. Our findings that women and minorities are less willing to participate in genetic research are consistent with studies on similar topics,6–9
some of which have investigated possible solutions.10
Overall, many people will provide broad consent for population-based research.9,11–13
Although our contact method is necessarily brief given the magnitude and broad scope of potential participation, recruitment could be improved by providing additional opportunities14,15
to learn about the biobank beyond the written materials presented in this survey.