Abstract
Adherence to antiretroviral therapy (ART) is essential for the effective management of HIV, which includes keeping the HIV viral load undetectable. This study aimed to determine whether certain ART medications are more "forgiving" of poor adherence in achieving viral suppression. We identified subgroups of ART medication usage and determined the extent to which ART adherence is associated with viral suppression across those subgroups. Data came from claims and clinical records (2017-2019) of 3,552 HIV-positive adult members of a Medicaid managed care plan. Pharmacy fill data were examined to characterize ART medications using latent class analysis (LCA), which captures the complexity of real-world ART usage (i.e., multiple medications, ART switching). LCA yielded five ART medication patterns over three years, mostly characterized by recent medications and formulations of ART, though they varied in number of tablets and in medication class. Mixed effects logistic regression models were estimated to determine whether odds of viral suppression differed by ART adherence level. After adjusting for covariates, those with at least 90% adherence (i.e., 90 to < 95%) did not significantly differ from those with 95% adherence or greater in terms of viral suppression, which corroborates existing clinical recommendations. These findings can inform provider-patient communication for people with HIV, especially those who have difficulty maintaining adherence. This includes those experiencing unstable housing, mental health conditions, or substance use.