Background:
Retention of participants is a critical component of ensuring the scientific goals of longitudinal research studies. Differential rates of attrition for ethnic minority participants can be particularly problematic in the fields of health and cardiovascular research, where ethnic minorities are shown to have disproportionately higher rates of both cardiovascular disease and risk factors for heart disease such as diabetes, hypertension, and obesity. The ability to explore the underlying causes of these differences is adversely affected when attrition in a study occurs at a higher rate among the ethnic minority subject participants. Understanding and preventing the causes of subject drop-out to improve retention among all ethnic groups is therefore a vital endeavor of any longitudinal research or cohort study.
Methods:
We analyzed data on ten-year retention rates of Caucasian, Chinese, Hispanic, and African-American participants in the Multi-Ethnic Study of Atherosclerosis (MESA). 6814 participants were recruited into the study in 2000, and 5,865 participants were still alive for the fifth in-person examination ten years later. Of these, 4651 participants returned for this examination. Logistic regression was used to examine the association between retention in the study, race/ethnicity and various baseline demographic characteristics, including age, gender, marital status, income, employment, education, language, place of birth, health insurance status and overall health status as measured by a 10-year predicted cardiovascular disease rate.
Results:
Racial differences in retention were observed, with Chinese, African-American, and Hispanic participants having 30-40% lower odds of being retained than Caucasian participants. However, after adjusting for demographic variables, these differences were primarily explained by indicators of socioeconomic status. Higher income, higher education, employment status, availability of health insurance and health status were significantly associated with ten-year retention in the study across all racial/ethnic groups. Marital status, gender, age, and birthplace (US vs non-US) were not associated with retention.
Conclusions:
Although minority participants were retained at lower rates in MESA, this can be primarily explained by differences in socioeconomic status and health status. Individuals with higher SES indicators may have life circumstances making participation in an examination taking much of a full day more plausible. Future studies should consider how these findings may inform developing support services or incentives which make follow-up participation in clinical research more persuasive for these individuals.
Recruiting ethnic minority participants to a clinical trial: a qualitative study
