Countries are considering whether to offer antiretroviral therapy (ART) to all HIV patients regardless of CD4 count. Clinical trials have found modest health benefits to early ART. However, these trials may underestimate the benefits. By seeking to minimize attrition, they fail to investigate an important behavioral pathway through which deferred ART eligibility may affect health in real world settings: non-retention among patients not yet eligible for therapy. We address this critical gap by assessing the effect of immediate (vs. deferred) ART eligibility on retention in HIV care in rural South Africa.
All patients (n=11,307) presenting to the public sector ART program in Hlabisa sub-district with a first CD4 count between 12 August 2011 and 31 December 2012 were included in the analysis. Patients were eligible for immediate ART if CD4<350 cells/μL; patients not yet eligible for ART were referred to pre-ART care and instructed to return every 6 months for CD4 monitoring. Because of measurement error in the CD4 laboratory assay, assignment to immediate versus deferred ART was effectively random near the threshold. We used a regression discontinuity design to recover causal effects. We assessed the effect of immediate eligibility on retention in HIV care at 12 months, as measured by the presence of a clinic visit, lab test, or ART start date in the interval 6 to 18 months (intent-to-treat effect). In addition, we assessed the causal effect of eligibility on retention in the subgroup of patients whose treatment uptake was determined by their CD4 count (complier causal effects).
Immediate eligibility increased 12-month retention from 32% to 50% (intent-to-treat effect: 18% points; 95%CI 11-23; p<0.001) among patients with first CD4 counts close to the 350-cell threshold. Having an eligible CD4 count increased the probability of initiating ART within six months from 18% to 43% (25% points; 95%CI 20-31; p<0.001). In patients whose uptake of ART was determined by the value of their CD4 count, having an eligible CD4 count increased 12-month retention from 21% to 91% (complier effect: 70% points; 95%CI 42-98; p < 0.001).
Deferred ART eligibility resulted in dramatically lower retention in HIV care among otherwise similar patients who just missed the cutoff for immediate eligibility. The results suggest that clinical trials may underestimate the benefits of early ART and, consequently, the clinical and population health benefits of eliminating CD4 initiation criteria.