Prior antiretroviral therapy exposure among clients presenting for HIV treatment initiation in South Africa: an exploratory mixed-methods study using multiple indicators of exposure

Background The era of universal treatment for HIV has seen high rates of disengagement from antiretroviral therapy (ART) programs and re-engagement after interruptions, with modeled estimates of non-naive initiators >50% in many places. Most re-engagers are reluctant to admit prior antiretroviral exposure, and non-self-reported data on proportions of re-initiators are scarce. We synthesized data from multiple sources to explore the proportion of people who present for initiation with evidence of prior ART use in South Africa. Methods We enrolled a sequential sample of adults presenting to initiate ART or to re-initiate ART after an interruption >3 months and collected 1) self-reported previous treatment experience; 2) electronic medical record (EMR) evidence of prior ART clinic visits; 3) baseline blood tests for metabolites of tenofovir diphosphate; and 4) laboratory records indicating prior ART-related tests. Interviews were conducted with a sub-sample of clients who self-reported no prior ART use but had evidence of metabolites. Results Among 89 enrolled participants (median age 32.5, 62% female), 16 (18%) self-reported previously taking ART >3 months prior to enrolment. An additional 33 (45%) who did not self-report prior exposure had EMR or laboratory evidence of prior ART use, for a total of 49 (55%) clients with known prior treatment exposure at initiation. Sensitivity of self-report was 40%, EMR 43%, metabolite testing 45%, and laboratory records 73%. Interviewees (n=11) reported opting to present as naive because they perceived that disclosure of prior disengagement would cause delays accessing treatment, require additional documentation, and elicit negative responses from healthcare workers. Study limitations included short duration of metabolite detectability (90 days), inability to link individuals within the EMR to discern ART experience at other facilities, and lack of baseline viral load testing. Conclusions At least 55% of clients initiating ART in South Africa have prior treatment experience, but only a third of re-initiators voluntarily reveal this. Laboratory records, which reflect long-term experience, yielded the most accurate results for ascertaining prior treatment exposure. As numbers re-engaging in HIV care after a treatment interruption increase, understanding reluctance to self-report ART experience and exploring opportunities to overcome barriers are critical for preventing repeated interruptions.