OBJECTIVE: To compare the treatment outcomes and mortality in a rural community-based ART (CBART) program with a hospital-based ART program in the same district. METHODS: The study design was a non-randomized cohort study consisting of 185 persons living with HIV (PLWHIV) in the CBART cohort and 200 PLWHIV in the hospital cohort. Eligibility for both cohorts was: being HIV-infected and eligible for ART, being treatment naïve, age 18 years or older, and being a resident of Rwimi sub-county. The intervention consisted of a community-based program which included weekly home visits to patients by trained volunteers who delivered antiretroviral drugs (ARVs), monitored and supported adherence to treatment, and identified and reported adverse reactions and other clinical symptoms. Outcome variables were compared to patients in a hospital-based cohort who received the standard care delivered to all other HIV patients in the hospital. The main outcome measures were HIV-1 RNA viral load (VL), CD4 cell count and mortality after six months of treatment. RESULTS: Successful ART treatment outcome as measured by virological suppression (VL<400 copies/ml) in the CBART cohort were similar to those in the hospital-based cohort (90.1% vs 89.3%, p=0.47). The median CD4 cell count increased significantly in both cohorts (community-based cohort 159 cells/microl vs 145 cells/microl in the hospital-based cohort). Mortality was not significantly different in both cohorts (community-based cohort 11.9%, hospital-based cohort 9.0%). CONCLUSION: The findings show that outcomes of a CBART intervention in a rural area compare favorably to outcomes of hospital-based care. If the study results are sustainable over a longer time period, this model could be considered for ART roll-out to impoverished rural/remote populations in Uganda and elsewhere.