Introduction In adolescents several psychosocial factors contribute to non-adherence. This study was undertaken to ascertain overall adherence of HIV infected adolescents to either boosted atazanavir (ATV/rtv) or lopinavir (LPV/rtv). The study was also conducted to compare self-report/pill counting adherence measurement techniques against the MEMS cap device technique.

Methods This was a prospective cohort study in adolescents (aged 12 to 23) receiving either boosted atazanavir (ATV/rtv) or lopinavir (LPV/rtv) as part of an antiretroviral regimen. This study was conducted at Newlands Clinic, Harare, Zimbabwe. Participants received a MEMS cap to assess adherence to the protease inhibitor (PI). During subsequent clinic visits, MEMS caps would have data downloaded before medicine was refilled. Participants were blinded to the results of the MEMS cap result. At each subsequent visit, a pill count and a self-report was also conducted to measure adherence. Viral load measurements were also recorded against the adherence data.

Results Fifty-two participants with a median age of 18 (range 12–23) years participated in the study with 53.8% being female. Utilising a pill count to assess adherence, 45 (86.5%) participants had a greater than 95% adherence to their PI regimen. However using the MEMS cap only, 4 (7.7%) participants had a greater than 95% adherence. Twenty-three of the 52 participants had a viral load greater than 50 (median = 21,228 cells/ml; range = 52–1,884,215) with a median adherence level of 100% (range = 93–100%) as determined by a pill count and a median adherence level of 41% (range = 3–100%) as determined by the MEMS cap.

Conclusion Pill counts and self-reported adherence overestimated adherence in adolescent patients on PIs as part of an antiretroviral regimen. Pill dumping phenomenon was observed in participants with high viral loads and greater than 95% adherence when assessed by pill count.

Disclosure of interest statement The authors have no conflict of interests to declare.