Observational study of continuity of HIV care following release from correctional facilities in South Africa

BACKGROUND: We sought to describe linkage to care, ART continuity, and factors associated with linkage to care among people with HIV following release from incarceration in South Africa. METHODS: We conducted a study of South African correctional service community reentrants who were receiving ART at the time of release. The study was implemented in three of 46 correctional service management areas. Participants were enrolled prior to corrections release and followed up to 90 days post-release to obtain self-reported linkage to care status and number of days of ART provided at corrections release. Clinic electronic and paper charts were sought and abstracted to verify self-reported linkage to care. Log-binomial regression, adjusted for facility, was used to identify associations with post-release linkage to care (self-reported and verified). We sought to specifically assess for associations with HIV diagnosis during index incarceration, ART initiation during index incarceration, and duration of incarceration. RESULTS: From May 2014 to December 2016, 554 inmates met eligibility and 516 (93%) consented to participate; 391 were released on ART, 40 of whom were excluded from analysis post-release. Of the remaining 351, 301 (86%) were men and the median age was 35 years (interquartile range 30, 40). Linkage to care was self-reported by 227 (64%) and linkage to care could be verified for 121 (34%). At most, 47% of participants had no lapse in ART supply. Initiating ART during the index incarceration showed a trend toward increased self-reported post-release linkage to care. Age > 35 years was associated with increased verified linkage to care while HIV diagnosis outside of a correctional setting and ART initiation during the index incarceration showed trends toward association with increased verified linkage to care. DISCUSSION: The results of our study are the first description of retention in care following correctional facility release from an African setting and indicate high levels of attrition during the transition from correctional facility to community care. Initiating ART within a correctional facility did not impair post-release linkage to care.