BackgroundCommunity-drug distribution point is a care model for stable patients in the community designed to make ART delivery more efficient for the health system and provide appropriate support to encourage long-term retention of patients. We examined program retention among ART program participants in rural Uganda, which has used a community-based distribution model of ART delivery since 2004.MethodsWe analyzed data of all patients >18 years who initiated ART in Jinja, Ugandan site of The AIDS Support Organization between January 1, 2004 and July 31, 2009. Participants attended clinic or outreach visits every 2–3 months and had CD4 cell counts measured every 6 months. Retention to care was defined as any patient with at least one visit in the 6 months before June 1, 2013. We then identified participants with at least one visit in the 6 months before June 1, 2013 and examined associations with mortality and lost-to-follow-up (LTFU). Participants with >4 years of follow up during August, 2012 to May, 2013 had viral load conducted, since no routine viral load testing was available.ResultsA total of 3345 participants began ART during 2004–2009. The median time on ART in June 2013 was 5.69 years. A total of 1335 (40 %) were residents of Jinja district and 2005 (60 %) resided in outlying districts. Of these, 2322 (69 %) were retained in care, 577 (17 %) died, 161 (5 %) transferred out and 285 (9 %) were LTFU. Factors associated with mortality or LTFU included male gender, [Adjusted Hazard Ratio (AHR) = 1.56; 95 % CI 1.28–1.9], CD4 cell count <50 cells/μL (AHR = 4.09; 95 % CI 3.13–5.36) or 50–199 cells/μL (AHR = 1.86; 95 % CI 1.46–2.37); ART initiation and WHO stages 3 (AHR = 1.35; 95 % CI 1.1–1.66) or 4 (AHR = 1.74; 95 % CI 1.23–2.45). Residence outside of Jinja district was not associated with mortality/LTFU (p value = 0.562). Of 870 participants who had VL tests, 756 (87 %) had VLs <50 copies/mL.ConclusionCommunity-based ART distribution systems can effectively mitigate the barriers to program retention and result in good rates of virologic suppression.

Introduction One-in-three men who have sex with men (MSM) in Uganda have never tested for HIV. Peer-driven HIV testing strategies could increase testing coverage among non-testers. We evaluated the yield of peer distributed HIV self-test kits compared with standard-of-care testing approaches in identifying undiagnosed HIV infection. Methods From June to August 2018, we conducted a pilot study of secondary distribution of HIV selftesting (HIVST) through MSM peer networks at The AIDS Support Organization (TASO) centres in Entebbe and Masaka. Peers were trained in HIVST use and basic HIV counselling. Each peer distributed 10 HIVST kits in one wave to MSM who had not tested in the previous six months. Participants who tested positive were linked by peers to HIV care. The primary outcome was the proportion of undiagnosed HIV infections. Data were analysed descriptively. Results A total of 297 participants were included in the analysis, of whom 150 received HIVST (intervention). The median age of HIVST recipients was 25 years (interquartile range [IQR], 22-28) compared to 28 years IQR (25-35) for 147 MSM tested using standard-of-care (SOC) strategies. One hundred forty-three MSM (95%) completed HIVST, of which 32% had never tested for HIV. A total of 12 participants were newly diagnosed with HIV infection: 8 in the peer HIVST group and 4 in the SOC group [5.6% vs 2.7%, respectively; P = 0.02]. All