Using data collected for a large genetic study of nicotine dependence, we traced subjects from random selection in the community through participation in a telephone interview and finally recruitment into the genetic study. This comprehensive design enabled us to evaluate which steps posed barriers in the recruitment of African Americans into genetic studies. First, we found considerable difficulty in locating subjects who lived in zip codes with higher proportions of African Americans because of changes of address or phone number. However, with a correct phone number, these subjects were more likely to answer the phone and more likely to participate in the telephone interview. Finally, when identified as eligible, African Americans were more likely to participate in the genetic study than European Americans. This suggests that the willingness to participate in genetic studies is not a significant barrier. In contrast, our data suggest that the strongest barrier is the establishment of initial contact. These results are consistent with findings in nongenetic studies, where locating minority subjects is the critical barrier to obtaining their participation (8
). This suggests that strategies developed in medical and survey research would also be effective for the recruitment of ethnic minorities into genetic studies (22
). In particular, targeting geographic regions (e.g., census blocks) with a high prevalence of minority households or otherwise oversampling minorities from a known sampling frame will increase contact rates among minorities at a given level of effort/cost in community-based surveys. Additionally, extending the number of telephone contact attempts, sending out additional mailings, and extending the data collection period have proven to be effective techniques for overcoming low contact rates and increasing recruitment of minorities for research, though at a higher cost (18
To clarify the discrepancy between our findings and the literature highlighting lower participation of African Americans in genetic studies, we examined the published participation rates in the published genetic studies (11
). Two of these 7 studies were population-based, as ours was, and the remaining 5 were other medical studies that requested that the person participate in the genetic component. None of these studies ascertained substance dependence or other psychiatric illness. Although all published genetic studies had significantly lower participation rates (P
< 0.05) among African Americans as compared with European Americans (11
), the magnitude of the difference between participation rates was small relative to the sample size of the individual studies (). Therefore, the hypothesis that willingness to participate does not seem to be a barrier in genetic studies is supported by previously published literature.
Participation Rates in Published Genetic Studies, by Ethnicity, 2000–2008
A limitation of our study is that we did not know the precise demographic characteristics of the unscreened subjects. We reduced this limitation by comparing the zip codes of the unscreened subjects and the reasons for lack of screening with demographic data on subjects aged 25–44 years from US Census 2000. Although there was an association between the proportion of people we failed to contact and the proportion of African Americans in a zip code, this association was highly confounded by poverty and other socioeconomic variables. Therefore, we were unable to separate the contributions of ethnicity and poverty at the step prior to contact. Interestingly, during community recruitment for a study involving an invasive test, low-income European Americans were found to be the most difficult to contact (24
). This suggests that poverty, rather than ethnicity, may be the dominant force behind the difficulty with locating and contacting subjects.
Subjects may have passively avoided participation in our study by either refusing the telephone screening or declining to answer the phone. Interestingly, avoiding contact was inversely correlated with proportion of African Americans in the zip code: The highest rates of both refusing the telephone screening and not answering the phone were in the zip codes with the lowest proportion of African Americans.
As we expand genetic databases to include more minority populations, it is important to understand potential barriers to minority recruitment. It is widely believed that minority populations in the United States, particularly African Americans, have a distrust of genetic research that leads to lower participation rates (7
). In contrast, we found that eligible African Americans participate in genetic studies as frequently as European Americans. The major barrier to minority recruitment in our study was the fact that African Americans were more difficult to locate in the community. Notably, because of decreased access of African Americans to health care in comparison with European Americans (25
), recruitment of African Americans at clinical centers is likely to be less representative than community-based recruitment. Although addressing feelings of distrust of medical and genetic research in minorities may remain important generally, our study suggests that increasing the participation of African Americans in genetic databases may be achieved by increasing efforts to locate and contact them.