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

• Published in: Arthritis and rheumatism 2004-06, Vol.51 (3), p.422-428
• Main Authors: Dunbar‐Jacob, Jacqueline, Holmes, Janice L., Sereika, Susan, Kwoh, C. Kent, Burke, Lora E., Starz, Terence W., McCall, Maura, Foley, Susan M.
• Format: Article
• Language: English
• Subjects: Adult; African American; African Americans - statistics & numerical data; Aged; Aged, 80 and over; Antirheumatic Agents - therapeutic use; Arthritis, Rheumatoid - drug therapy; Arthritis, Rheumatoid - psychology; Arthritis, Rheumatoid - therapy; Attrition; Biological and medical sciences; Chi-Square Distribution; Clinical Trials as Topic - statistics & numerical data; Cognitive Therapy; Diseases of the osteoarticular system; European Continental Ancestry Group - statistics & numerical data; Female; Humans; Inflammatory joint diseases; Informed Consent; Logistic Models; Male; Medical sciences; Middle Aged; Minority; Odds Ratio; Patient Compliance; Patient Dropouts - statistics & numerical data; Personnel Selection; Recruitment; Rheumatoid arthritis; Sex Ratio
• ISSN: 0004-3591; 0893-7524; 1529-0131; 1529-0123
• DOI: 10.1002/art.20411

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.