Benefits of ART simplification on adherence, clinical and economic outcomes

Methods The unit analysis was a person-regimen of 2 NRTI (NNRTI, PI/r) lasting ]90 days. Adherence was measured by the proportion of days covered using pharmacy refill data. The Wilcoxon rank-sum test was used to compare CNA, PNA and total non-adherence (TNA) in the following groups: STR vs its components (EFV TDF/FTC) and non-FDC vs FDC vs STR. The level of adherence per group and the impact of PNA and CNA on virological failure (VF) were estimated using multivariate regression panel data models (MVRPDM). Control variables included age, gender, hepatitis co-infection, cumulative comorbidity score, CD4, viral load, regimen duration and calendar year. MVRPDM we also used to access the impact of PNA-free regimens on the probability of at least one hospitalization and on annualized hospitalization ART costs. The analysis was performed in Stata 11†.