Introduction: Cambodia suffers very high rates of tuberculosis (TB). In 1994, Hospital Direct Observed Treatment (DOT), as recommended by WHO, was introduced. In 2002 ambulatory or health center DOT (HC-DOT) was implemented, requiring patients to come to the HC to take medicine with HC staff during the intensive phase. In 2005, community-based DOT (CDOT), using health volunteers in their own community, was implemented in some areas. This study compares CDOT and HC-DOT in terms of treatment success, case notification and acceptability. Methods: This thesis includes a systematic review and meta-analysis of studies comparing CDOT to HC-DOT and Family-DOT; a controlled before-after quantitative study of the introduction of CDOT (1067 participants in 48 HCs with all TB forms were surveyed at the baseline in 2002/2003, as were 1297 in 2006/2007 two years after its introduction); and a qualitative study assessing its strengths, weaknesses and acceptability to HC staff, patients and CDOT watchers. Results: The meta-analysis from the systematic review shows that CDOT was at least as effective as HC-DOT (RR=1.05, Cl: 1.01-1.10), and treatment of smear-positive patients treated by CDOT is clearly better than for family-DOT (OR=1.33, Cl: 1.20-1.48) The quantitative study shows that in Cambodia the treatment success rates of patients treated with CDOT and HC-DOT are similar (OR=1.26, 95% Cl: 0.56-2.82). In the CDOT areas, the death rate among all TB patients was significantly reduced after its implementation (OR=0.37, 95%CI: 0.15-0.92) Most participants in the qualitative study felt CDOT was helpful particularly for very sick patients, and because it reduced travel costs and allowed patients to continue working. There was significant involvement of family members in both approaches. Conclusion: CDOT is as effective as HC-DOT for treatment success and case finding. Both approaches are well accepted. There remain problems which need to be addressed in order to improve outcomes