How well does self-reported adherence fare compared to therapeutic drug monitoring in HAART?

Objective. The aim of the study was to determine how well self-reported adherence fares compared to therapeutic drug monitoring in monitoring HAART adherence. Methods. We administered a validated self-reported adherence (SRA) questionnaire to 925 HIV patients on HAART in a large Malaysian hospital from 2010 to 11. We also performed Therapeutic Drug monitoring (TDM) by concurrently collecting and testing blood samples for Efavirenz, Nevirapine and Lamivudine using Liquid Chromatography/Mass Spectrometry. We compared the SRA against the TDM results. Sensitivity, specificity, positive (PPV) and negative predictive (NPV) and diagnostic accuracy values were computed for each drug. Results. Self-reported adherence (SRA) over-estimates adherence by between 6 and 10 percentage points compared to therapeutic drug monitoring (TDM). SRA is highly sensitive with sensitivity exceeding 0.90 but is not very specific (0.56-0.63). PPV for SRA ranged between 0.76 (Lamivudine) and 0.84 (Efavirenz) while NPV ranged between 0.78 (Lamivudine) and 0.81 (Efavirenz). Overall diagnostic accuracy ranged between 0.76 (Lamivudine) and 0.84 (Nevirapine). Conclusion. Self-reported adherence is a surprisingly accurate instrument for measuring HAART adherence compared to TDM and can be reliably used in practice in resource-poor settings. (C) 2013 Elsevier Inc. All rights reserved.