Background

Medication adherence is a critical aspect of healthcare and is an ongoing issue that fails to meet measures for optimal therapeutic benefits.1 However, the World Health Organization (WHO) estimates the medication adherence value of chronic medications in developed countries around 50%.2 With the technology we have today, electronic modes of intervention are rapidly becoming a promising method of increasing medication adherence. Studies have shown that indirect methods of intervention have similar efficacy to direct, in-person models.3 Additionally, strong correlations have been particularly demonstrated between mobile phone interventions and increased medication adherence and health outcomes.4-7

Objective

Analyze claims data and determine whether plan-initiated text message refill reminders increase prescription refill rates of chronic medications among Medicare patients.

Methods

This will be a historical cohort study that will collect and analyze data from October 23, 2018 to April 23, 2019 for a total study period of six months. The study population will be selected from a cohort of members enrolled under Magellan Health’s Medicare Prescription D Plan (PDP). The primary outcome of this study will be mean refill rates. The secondary outcomes were PDC and proportion of patients with PDC values ≥ 80%. All outcomes were stratified into three disease states (i.e., hyperlipidemia (HLD), Type 2 diabetes (T2DM), and hypertension (HTN)) the patients were taking their chronic medications for. The two-sided 95% confidence intervals were utilized to determine statistical significance. The one-way ANOVA tests were used to detect statistical significance at an alpha of 0.05 between the three arms. The generalized linear model (GLM) and the GLMSELECT procedure were utilized to identify important predictors.

Results

The opt-in group was associated with 2.1% (95% CI, 1.1 – 3.1%) higher mean PDC for HLD medications compared to the no-text group. The opt-out group was associated with 3.2% (95% CI, 0.1 – 6.3%) higher mean PDC than the no-text group for T2DM medications. Additionally, both the opt-in and opt-out groups were associated with higher mean PDC for HTN medications compared to the no-text group, 2.8% (95% CI, 1.8 – 3.9%) and 3.3% (95% CI, 1.2 – 5.4%) respectively. Finally, the GLM and GLMSELECT models identified LIS status, the region of Arizona, and the region of New York to be the most significant confounding predictors.

Conclusions

This study demonstrates that weekly to monthly plan-initiated text-based medication refill reminders are not significantly associated with an increase in refill rates; however, it may be associated with an increase in PDC values for HTN and HLD medications. Future studies should focus on controlling for potentially significant confounders like LIS status, geographical location, and other medication adherence interventions they may be already receiving.