Evaluation of a practical method to assess antiretroviral adherence in HIV-infected Thai children

The objective of this study was to evaluate a practical method to assess adherence to antiretroviral therapy by observing virological and immunological responses. We conducted a 12-month longitudinal cohort study of 162 HIV-infected Thai children. Adherence was assessed using 5 methods (self reporting calendar, records of missed doses, pill counts, physician assessment, and an interview questionnaire). CD4 count, percentage and viral load were performed at baseline and at 12 months. Mean adherence rates at 2, 6, and 12 months were 98, 100, and 99% by the calendar method; 98, 100, and 100% by recording missed doses; 96, 96, and 92% by pill count; and 90, 94, and 97% by physician assessment. Poor agreement (kappa < or = 0.1) was found among the methods. There was a statistically significant difference (p = 0.05) in virological response between participants with > or = 95% adherence (0.8 log10) and those with < 95% adherence (0.2 log10) when pill counts were used to assess adherence. In conclusion, despite poor agreement among these tools, a pill count appeared to be the only practical, validated method to differentiate the virological outcome between those who were fully and partially adhere to the treatment regimen.