Long-term retention and predictors of attrition for key populations receiving antiretroviral treatment through community-based ART in Benue State Nigeria: A retrospective cohort study

Background Key populations (KP) are disproportionately infected with HIV and experience barriers to HIV care. KP include men who have sex with men (MSM), female sex workers (FSW), persons who inject drugs (PWID) and transgender people (TG). We implemented three different approaches to the delivery of community-based antiretroviral therapy for KP (KP-CBART) in Benue State Nigeria, including One Stop Shop clinics (OSS), community drop-in-centres (DIC), and outreach venues. OSS are community-based health facilities serving KP only. DIC are small facilities led by lay healthcare providers and supported by an outreach team. Outreach venues are places in the community served by the outreach team. We studied long-term attrition of KP and virological non-suppression. Method This is a retrospective cohort study of KP living with HIV (KPLHIV) starting ART between 2016 and 2019 in 3 0SS, 2 DIC and 8 outreach venues. Attrition included lost to follow-up (LTFU) and death. A viral load >1000 copies/mL showed viral non-suppression. Survival analysis was used to assess retention on ART. Cox regression and Firth logistic regression were used to assess risk factors for attrition and virological non-suppression respectively. Result Of 3495 KPLHIV initiated on ART in KP-CBART, 51.8% (n = 1812) were enrolled in OSS, 28.1% (n = 982) in DIC, and 20.1% (n = 701) through outreach venues. The majority of participants were FSW—54.2% (n = 1896), while 29.8% (n = 1040), 15.8% (n = 551) and 0.2% (n = 8) were MSM, PWID, and TG respectively. The overall retention in the programme was 63.5%, 55.4%, 51.2%, and 46.7% at 1 year, 2 years, 3 years, and 4 years on ART. Of 1650 with attrition, 2.5% (n = 41) died and others were LTFU. Once adjusted for other factors (age, sex, place of residence, year of ART enrollment, WHO clinical stage, type of KP group, and KP-CBART approach), KP-CBART approach did not predict attrition. MSM were at a higher risk of attrition (vs FSW; adjusted hazard ratio (aHR) 1.27; 95%CI: 1.14–1.42). Of 3495 patients, 48.4% (n = 1691) had a viral load test. Of those, 97.8% (n = 1654) were virally suppressed. Conclusion Although long-term retention in care is low, the virological suppression was optimal for KP on ART and retained in community-based ART care. However, viral load testing coverage was sub-optimal. Future research should explore the perspectives of clients on reasons for LTFU and how to adapt approach to CBART to meet individual client needs.