SETTING: Kilimanjaro Region, northern Tanzania. OBJECTIVE: To assess the effect of the introduction of the patient-centred tuberculosis treatment (PCT) approach-which allows tuberculosis (TB) patients to choose between community and facility-based directly observed treatment (DOT)-on treatment outcomes, and to analyse factors that contribute to opting for community DOT. DESIGN: Retrospective analysis of treatment outcomes of TB patients registered in the Kilimanjaro Region in 2007, differentiating between patients under community vs. facility-based DOT and taking into account demographic factors, disease classification, TB diagnosis and human immunodeficiency virus (HIV) status. RESULTS: Data from 2769 TB patients were analysed. Treatment success rates were respectively 81% and 70% in patients under community vs. facility-based DOT (P < 0.001). Cure rates were respectively 73% and 72% in smear-positive pulmonary TB patients under community vs. facility-based DOT (P = 0.62). Women, children, patients residing in districts other than Hai, patients with newly diagnosed TB and patients with smear-negative pulmonary TB were most likely to be under community DOT. CONCLUSION: The PCT approach was shown to be effective in terms of treatment outcomes. Treatment success rates were higher in patients who opted for community DOT than in patients who chose facility-based DOT (all cases), and were similar in smear-positive pulmonary TB patients under community or facility-based DOT