Changes in adherence to highly active antiretroviral therapy medications in the Multicenter AIDS Cohort Study*

OBJECTIVES: To characterize the determinants of changes in adherence to antiretroviral therapy and examine whether there are persistent lower adherers. DESIGN: A cohort study with repeated measurements. METHODS: Self-reported 100% adherence was defined as taking all doses and numbers of pills over a 4-day period as prescribed for current HIV medications. Independent predictors of changing adherence (< 100% to 100% and 100% to < 100%) were determined by logistic regression, correcting for correlated repeated measures for 597 HIV-positive men reporting the use of highly active antiretroviral therapy (HAART) between October 1998 and October 2000. RESULTS: Of the 942 visit-pairs with initial 100% adherence, 106 (11.3%) reduced adherence to less than 100%, and 836 (88.7%) remained 100% adherent at the next 6-month visit. No recent outpatient visits, younger age, depression, less than college educated, and later in calendar time predicted decreasing adherence. Among 186 visit-pairs starting with less than 100% adherence, 133 (71.5%) improved adherence to 100% and 53 (28.5%) remained less than 100% adherent at the next visit. The determinants of improving adherence included not being African-American, not using recreational drugs, and having had more than three HAART regimens. Lower adherence was not a random event; it was significantly correlated across visits within the individual. CONCLUSION: Characteristics associated with improving and lowering adherence differed and should be considered in developing interventions to enhance adherence and optimize effective therapies.