BackgroundMost HIV treatment programs in resource-limited settings utilize multiple facilitators of adherence and retention in care but there is little data on the efficacy of these methods. We performed an observational cohort analysis of a treatment program in Kenya to assess which program components promote adherence and retention in HIV care in East Africa.MethodsPatients initiating ART at A.I.C. Kijabe Hospital were prospectively enrolled in an observational study. Kijabe has an intensive program to promote adherence and retention in care during the first 6 months of ART that incorporates the following facilitators: home visits by community health workers, community based support groups, pharmacy counseling, and unannounced pill counts by clinicians. The primary endpoint was time to treatment failure, defined as a detectable HIV-1 viral load; discontinuation of ART; death; or loss to follow-up. Time to treatment failure for each facilitator was calculated using Kaplan-Meier analysis. The relative effects of the facilitators were determined by the Cox Proportional Hazards Model.Results301 patients were enrolled. Time to treatment failure was longer in patients participating in support groups (448 days vs. 337 days, P<0.001), pharmacy counseling (480 days vs. 386 days, P = 0.002), pill counts (482 days vs. 189 days, P<0.001) and home visits (485 days vs. 426 days, P = 0.024). Better adherence was seen with support groups (89% vs. 82%, P = 0.05) and pill counts (89% vs. 75%, P = 0.02). Multivariate analysis using the Cox Model found significant reductions in risk of treatment failure associated with pill counts (HR = 0.19, P<0.001) and support groups (HR = 0.43, P = 0.003).ConclusionUnannounced pill counts by the clinician and community based support groups were associated with better long term treatment success and with better adherence.

BackgroundPill counts are often used to measure adherence to ART, but there is little data on how they affect adherence. We previously showed a bivariate relationship between clinicians counting pills and adherence in patients receiving HIV care in Kenya. We present a secondary analysis of the relationship between numbers of pill counts and clinical outcomes in resource limited settingsMethodsPatients initiating ART at Kijabe Hospital were monitored for the number of discretionary pill counts performed by their clinician in the first 6 months of ART. Subjects were followed for at least 1 year after enrollment. The number of clinician pill counts was correlated to ART adherence. The primary endpoints were time to treatment failure, defined as a detectable HIV-1 viral load, death; or loss to follow-up.ResultsClinician pill counts were done at 68% of clinic visits for 304 subjects. There was a positive correlation between the number of clinician pill counts and ART adherence (r = 0.21, p <0.001). Patients were divided into 3 groups (0 counts, 1 to 3 counts, 4 to 7 counts) and exhibited adherence of 76%, 84%, and 92%, respectively (p = 0.004). Time to treatment failure for these groups was 220 days, 438 days, and 497 days (P<0.01), respectively. Time to virologic failure in living patients remaining in the cohort was longer in those with more pill count (P =0.02). Multi-variate analysis adjusting for co-variates affecting time to treatment failure found that that clinician pill counts were associated with a decreased risk of treatment failure (HR = 0.69, p =0.04).ConclusionsThe number of clinician pill count performed was independently associated with better adherence and a decreased risk of treatment failure. The use of clinician pill counts should be further studied as an adherence promoter through a randomized clinical trial.

No abstract