Differentiated service delivery for people using second‐line antiretroviral therapy: clinical outcomes from a retrospective cohort study in KwaZulu‐Natal, South Africa

INTRODUCTION: Evidence is needed to guide the inclusion of broader groups of people living with HIV (PLHIV) in differentiated service delivery (DSD) programmes. We assessed treatment outcomes among PLHIV on second‐line regimens in a community antiretroviral therapy (ART) delivery programme, compared to those who remained at clinics. METHODS: Using data from 61 public clinics, we did a retrospective cohort study among PLHIV receiving second‐line ART following rollout of the Centralized Chronic Medicines Dispensing and Distribution (CCMDD) programme in KwaZulu‐Natal, South Africa. We included PLHIV from the timepoint when they were first eligible, though not necessarily referred, for community ART within CCMDD and followed them for 18 months. We used multivariable logistic regression to compare 12‐month attrition and viraemia between clients referred for community ART and those remaining in clinic care. RESULTS: Among 209,744 PLHIV aged ≥ 18 years who collected ART between October 2016 and December 2018, 7511 (3.6%) received second‐line ART. Of these, 2575 (34.3%) were eligible for community ART. The median age was 39.0 years (interquartile range 34.0–45.0) and 1670 (64.9%) were women. Five hundred and eighty‐four (22.7%) were referred for community ART within 6 months of meeting eligibility criteria. Overall, 4.5% [95% confidence interval (CI) 3.0–6.6%] in community ART and 4.4% (95% CI 3.5–5.4%) in clinic care experienced attrition at 12 months post eligibility for community ART. Two thousand one hundred and thirty‐eight (83.0%) had a viral load recorded 6–18 months after becoming eligible, and of these, 10.3% (95% CI 7.7–13.3%) in community ART and 11.3% (95% CI 9.8–12.9%) in clinic care had viraemia > 200 copies/ml. In separate regressions adjusted for age, gender, district, time on second‐line ART, nucleoside reverse transcriptase inhibitor backbone and year of eligibility, no differences in the odds of attrition [adjusted odds ratio (aOR) 1.02, 95% CI 0.71–1.47] or viraemia (aOR 0.91, 95% CI 0.64–1.29) were observed between those in community ART and those remaining in clinic care. CONCLUSIONS: We found good outcomes among PLHIV who were stable on second‐line regimens and referred for community ART. Efforts to expand DSD access among this group should be prioritized.