Objective: To determine whether patients receiving directly observed treatment (DOT) had lower all-cause mortality than those treated with self-administered treatment (SAT) and to identify factors associated with mortality among tuberculosis (TB) patients.
Design: All TB patients in Taipei, Taiwan, diagnosed between 2006 and 2008 were included in a retrospective cohort study.
Results: Among 3624 TB patients, 45.5% received DOT, which was disproportionately offered to older patients and those with more underlying illness and severe TB disease. After controlling for patients' sociodemographic factors, clinical findings and underlying comorbidities, the odds of death were 40% lower (aOR 0.60, 95%CI 0.5-0.8) among patients treated with DOT than those on SAT. After adjusting for DOT, independent predictors of death included non-Taiwan birth, increasing age, male, unemployment, end-stage renal disease requiring dialysis, malignancy, acid-fast bacilli smear positivity and pleural effusion.
Conclusion: DOT was associated with lower all-cause mortality after controlling for confounding factors. DOT should be expanded in Taiwan to improve critical treatment outcomes among TB patients.