Accuracy of self-reported adherence to tuberculosis therapy among DOTS patients in Mumbai

Purpose: To compared self-reported adherence to DOTS therapy with urine rifampicin metabolite levels and medical records among patients in Mumbai, India. Methods: Study subjects (N=538) were randomly selected from the DOTS centers in Mumbai, India. Self-reported adherence was ascertained by interviews; unannounced home visits were conducted, and urine samples were collected for rifampicin metabolite testing using the n-butanol test. Information from medical records was abstracted for documented receipt of drugs from the DOTS centers. Results: Agreement between self-reported adherence and urine tests was very poor (kappa, 0.08); and between self-reports and medical records was moderate (kappa, 0.47). Receipt of drugs did not ensure adherence. Based on urine n-butanol test, 75% of patients were adherent. Physical appearance of urine for rifampicin excretion with the n-butanol extraction method indicated a high positive predictive value (95%). Conclusion: We recommend incorporating urine tests for various drug metabolites periodically in the DOTS program to ensure treatment adherence.