Factors associated with attrition of African Americans during the recruitment phase of a clinical trial examining adherence among individuals with rheumatoid arthritis

Objective. To examine factors contributing to the loss of potential minority participants in a study of medication adherence among rheumatoid arthritis patients. Methods. Chi-square statistics were used to test for differences in refusal/ineligibility by race and site at 4 screening points (initial eligibility review, initial patient contact, adherence screening, and informed consent). Differences in criteria-specific risks for exclusion at initial eligibility review were examined across 4 sites by race. Odds ratios (95% confidence intervals) were estimated if differences were significant. Multivariate logistic regression was used to examine sociodemographic factors associated with the probability for ineligibility at the initial eligibility review. Stated reasons for refusal were qualitatively quantified. Results. A greater percentage of African Americans were lost at every screening point when compared with whites, but only the difference at the initial eligibility review was statistically significant. Conclusion. Factors associated with attrition included selection of area medical sites, research design issues, comorbid conditions, alcohol abuse, and being younger, unmarried, African American, and male.