Patterns of engagement in care during clients’ first 12 months after HIV treatment initiation in South Africa: A retrospective cohort analysis using routinely collected data

By  Dr Mhairi Maskew  Mariet Benade  Dr. Amy Huber  Dr Sophie Pascoe  Dr. Linda Sande  Lufuno Malala  Musa Manganye  Professor Sydney Rosen  |  | 


Retention on antiretroviral therapy (ART) during the early treatment period is one of the most serious challenges facing HIV programs, but the timing and patterns of early disengagement from care remain poorly understood. We describe patterns of engagement in HIV care during the first year after treatment initiation. We analysed retrospective datasets of routinely collected electronic medical register (EMR) data for ≥18-year-old clients who initiated ART at public sector clinics in South Africa after 01/01/2018 and had ≥14 months of potential follow-up. Using scheduled visit dates, we characterized engagement in care as continuous (no treatment interruption), cyclical (at least one visit >28 days late with a return visit observed) or disengaged (visit not attended and no evidence of return). We report 6- and 12-month patterns of retention in care and viral suppression. Among 35,830 participants (65% female, median age 33), in months 0–6, 59% were continuously in care, 14% had engaged cyclically, 11% had transferred to another facility, 1% had died, and 16% had disengaged from care at the initiating facility. Among disengagers in the first 6 months, 58% did not return after their initiation visit. By 12 months after initiation, the overall proportion disengaged was 23%, 45% were classified as continuously engaged in months 7–12, and only 38% of the cohort had maintained continuous engagement at both the 6- and 12-month endpoints. Participants who were cyclically engaged in months 0–6 were nearly twice as likely to disengage in months 7–12 as were continuous engagers in months 0–6 (relative risk 1.76, 95% CI:1.61–1.91) and were more likely to have an unsuppressed viral load by 12 months on ART (RR = 1.28; 95% CI1.13–1.44). The needs of continuous and cyclical engagers and those disengaging at different timepoints may vary and require different interventions or models of care.

Publication details

PLOS Glob Public Health