Using electronic methods of adherence monitoring and therapeutic drug monitoring (TDM) to eliminate discordance between antiretroviral adherence and virological failure

Abstract

Background: Adherence to antiretroviral therapy (ART) is critical: only 70% achieve viral suppression at a year. Current adherence methodologies, with slow reaction to missed dosing, inadequately predict virological outcomes. Ideal adherence methods would be cheap, easy to use, and allow rapid response to missed doses to improve outcomes. We explored ideal adherence monitoring methodology for a large public sector ART clinic in Cape Town. Methods: We designed a randomised controlled study for ART-naïve individuals to determine whether text messaging after a missed dose would improve adherence recorded by an electronic adherence monitoring device (EAMD), reduce treatment interruptions or impact on virological outcome (using regression modelling). Five other measures of adherence were captured prospectively during the study: selfrecall (SR), clinic-based pill count (CPC), pharmacy refill data (PR-average or PR-gaps) and efavirenz concentration. The predictive value of each adherence methodology on virological and HIV-1 resistance outcomes was compared by calculating the area under the receiver operating characteristic curve, from logistic regression models. The impact of efavirenz concentration and CYP2B6 metaboliser genotype data on failure was examined using Cox proportion hazard modelling; and the most predictive lower limit for EFV concentration was determined. Antiretroviral cohort and pharmacy refill data were compared, using simple statistics, to determine which provided the best method of determining those retained in care.