Abstract
BACKGROUND: Improved medication adherence, represented as an increase in the proportion of days covered (PDC), to chronic medications is associated with better patient outcomes, yet effective strategies to improve adherence are often resource intensive. OBJECTIVE: To quantify the impact of a pharmacist-supported electronic outreach initiative on medication adherence measures and to qualitatively evaluate patient engagement with and response to electronic messaging. METHODS: This retrospective cohort evaluation used mixed methods to assess the impact of a population health quality improvement program to address medication adherence for Medicare Advantage enrollees. The intervention was performed between January 1, 2023, and December 31, 2023, by population health teams supporting the University of Colorado Primary Care practices. The teams reviewed insurer-provided lists to identify patients late to refill a medication and sent an electronic health message, a mailed letter, or a phone call conveying concern for adherence. Patient responses requiring clinical intervention or education were triaged to clinical pharmacists for management per their clinical discretion. The proportion of Medicare Advantage enrollees classified as adherent, defined as PDC value of 0.8 or higher, was compared before (2022) and after implementation of the population-based outreach intervention for the 2023 plan year. Qualitative methods were used to evaluate patient response to electronic messages. The text of all patient replies to electronic messages was thematically analyzed and categorized. RESULTS: The proportion of patients classified as adherent to diabetes medications was higher in the postintervention group (87.5%) compared with the preintervention group (83.4%, P = 0.021), whereas the adherence rates in the post- vs prehypertension (89.3% vs 88.7%, P = 0.517) and cholesterol (89.4% vs 89.2%, P = 0.721) groups were not significantly different. The population health teams sent 1,593 electronic health record messages, 1,185 (74.4%) of which were opened, and patients responded to 516 (32.4%). The most common patient response was patients self-reporting being adherent (306, 59.3%); fewer patients admitted to some degree of nonadherence (111, 21.5%). An equal number of patients reported appreciation for (111, 21.5%) and confusion about or irritation with the outreach (111, 21.5%). CONCLUSIONS: The University of Colorado Medicine's population health initiative provided mixed results on medication adherence metrics. Electronic health record messaging provided insight into ways to improve the intervention to better engage and assist patients.