Background. Large-scale, government-based antiretroviral therapy (ART) programs in rural areas of resource-poor countries remain largely unevaluated. Methods. We conducted a retrospective review of all patients receiving ART (n ¼ 399) to assess survival and retention in care and a prospective evaluation of patients on ART for at least 6 months (n ¼ 175). We used 3-day self-report to measure adherence. Results. The probability (95% confidence interval [CI]) of surviving and remaining in care was 0.76 (0.72, 0.81) at 1 year. Men and patients with advanced disease were more likely to die or be lost to follow-up. At baseline, 149 (85%) reported 100% adherence. Nonadherence was associated with lack of suppression of viral replication (odds ratio [OR] ¼ 4.5; 95% CI: 1.8, 11.5). Missing a scheduled clinic visit and lack of disclosure of HIV status were associated with nonadherence. Conclusion. Viral suppression was high, but counseling to include HIV disclosure to family and keeping scheduled clinic appointments may improve long-term adherence and treatment outcomes.