BACKGROUND: Differentiated Models of HIV care (DMOC) aim to improve health care efficiency in the context of ART scale up. This study in Kwazulu Natal, South Africa, evaluates effectiveness of four DMOC models: counselor-led adherence clubs, community ART groups (CAG), fast lane spaced appointment (FLSA), and community pick up points (PuP). We compared retention between DMOC models, and retention on ART between DMOC patients and patients in normal care at 10 MSF supported clinics.
METHODS: Retrospective cohort study among patients enrolled into DMOC between 1/1/2012 and 31/12/2018, using routine ART data. DMOC was offered to patients >=18 years old, on ART for >= 12 months, and virologically suppressed. Kaplan-Meier techniques were applied. DMOC patients were followed from enrollment till return to clinic care, death or loss-to-follow up (LTFU). Outcomes for ART retention were LTFU or death. Censoring occurred in case of transfer out or database closure on 5/12/2019.
RESULTS: 9501 were enrolled into DMOC; median age was 39 years (IQR 32-48), 2353 (24.8%) were males, time on ART prior enrollment was 3.2 years (IQR 2-5); 1561 (18.1%) returned to clinic care, 657 (6.9%) were LTFU, 109 (1.1%) died. DMOC retention was 89.1%, 82.5%, 77.8% at 12, 24 and 36 months. DMOC retention was lower for CAGs (P=0.006) (Table 1). ART retention was 96.6%, 93.2%, 90.2% at 12, 24, and 36 months and did not differ across DMOC types (P=0.09), but was significantly higher compared to clinics (P<0.0001). VL completion was sub-optimal in all models; VL suppression at <400 copies/ml was lower amongst patients at clinics and CAGs (P<0.0001).

CONCLUSIONS: Concurrent implementation of alternative DMOC approaches in a large ART program is feasible, and achieves high long-term retention on ART. In order to address patients' changing needs robust monitoring is essential to assess clinical status, movement between models of care, and return to clinic care.