Attrition through Multiple Stages of HIV Care in South Africa: A Challenge for Test-and-Treat

By  Dr. Matthew Fox  Dr. Kate Shearer  |  | 


Introduction: While momentum for test-and-treat strategies has been increasing in resource limited settings, if patients cannot be retained in HIV care continuously from the time of testing positive through long term adherence to antiretroviral therapy (ART), such strategies may fall well short of the expected gains. While numerous estimates of retention on ART exist, few cohorts have data on retention from testing positive through pre-ART and on ART care. Methods: We explored attrition (loss or death) at the Themba Lethu HIV clinic in Johannesburg, South Africa in 3 cohorts enrolled at HIV testing, pre-ART care initiation, and ART initiation. Loss was defined as ≥3 months late for the ART and pre-ART cohorts and ≥1 month late for the VCT cohort. Follow-up time accrued from the beginning of the stage until LTF, death, 12 months of follow up, or dataset closure on September 6, 2012. Results: Between March 2010 and March 2012 we enrolled 306 patients testing HIV+, 185 initiating pre-ART care, and 152 initiating ART. By Kaplan-Meier analysis, attrition through 12 months from HIV testing was 70.2%, from pre-ART initiation was 23.5% and from ART initiation was 11.8%. In rates per 100 person months, attrition was highest in the VCT cohort (14.7; 95%CI: 12.9-16.9) and lower in the pre-ART (1.8; 95%CI: 1.3-2.6) and ART cohorts (1.0; 95%CI: 0.65-1.7). Of the 306 patients enrolled at testing HIV+, 63.1% were lost, 6.9% died, 8.2% are in pre-ART care and 21.9% are on ART. Of the 185 patients enrolled at pre-ART care, 15.7% were lost, 2.2% died, 18.4% are still in pre-ART and 63.8% are on ART. Of the ART cohort, 8.6% were lost, 3.3% died and 88.2% are still on ART. Multiplying these parameters, we estimate only 28.4% of patients are retained from testing HIV+ through one year on ART. While predictors of attrition (see table) differed by stage, attrition rates differed by age and sex. First CD4 count (at testing HIV+) was predictive of attrition for pre-ART patients (>200 vs. <100, RR: 1.72; 95%CI: 0.59, 5.03) and baseline CD4 count was for ART patients (>200 vs. <100, RR: 0.57; 95% CI: 0.18, 1.81). Conclusions: Our findings from a well-resourced clinic demonstrate continual loss from all stages of HIV care and low overall retention. This poses a threat to the effectiveness of plans for test-and-treat strategies. Such projects must pay careful attention to maintain constant patient care if such strategies are to produce expected reductions in transmission.

Conference: Croi conference 2013, Atlanta, USA

Publication details