Directly observed treatment is associated with reduced default among foreign tuberculosis patients in Thailand

SETTING: Thailand's Tuberculosis (TB) Active Surveillance Network in four provinces in Thailand. OBJECTIVE: As treatment default is common in mobile and foreign populations, we evaluated risk factors for default among non-Thai TB patients in Thailand. DESIGN: Observational cohort study using TB program data. Analysis was restricted to patients with an outcome categorized as cured, completed, failure or default. We used multivariate analysis to identify factors associated with default, including propensity score analysis, to adjust for factors associated with receiving directly observed treatment (DOT). RESULTS: During October 2004-September 2006, we recorded data for 14359 TB patients, of whom 995 (7%) were non-Thais. Of the 791 patients analyzed, 313 (40%) defaulted. In multivariate analysis, age >= 45 years (RR 1.47, 95%CI 1.25-1.74), mobility (RR 2.36, 95%CI 1.77-3.14) and lack of DOT (RR 2.29, 95%CI 1.453.61) were found to be significantly associated with default among non-Thais. When controlling for propensity to be assigned DOT, the risk of default remained increased in those not assigned DOT (RR 1.99, 95%CI 1.03-3.85). CONCLUSION: In non-Thai TB patients, DOT was the only modifiable factor associated with default. Using DOT may help improve TB treatment outcomes in non-Thai TB patients.