Abstract
INTRODUCTION: Adherence to prescribed inhaled controller medication is a determinant of asthma health outcomes. Traditional methods for assessing medication adherence (MA) can be challenging in real-world clinical settings. A new behavioral science approach presents opportunities to develop a novel MA assessment tool that also allows the prediction of acute asthma exacerbations. The current study aims to evaluate MA among adults with asthma based on their prescription collection behavior and its relationship with subsequent exacerbations. METHOD: This retrospective study was conducted on Asian adults with clinically diagnosed asthma and managed in public primary care clinics in Singapore from 2016 to 2023. Clinical data of patients, including socio-demographical, clinical (including Asthma Control Test scores), and prescription records were thoroughly examined to determine MA. The participants were stratified into the Full Collection (FC) group for those collections of prescribed asthma medication within a week; Partial Collection (PC) group for partial medication collection; No Collection (NC) group for no dispensation record within 1 year of the prescription date. The Proportion of Days Covered (PDC), defined as the proportion of days in which a patient gets access to the medication was determined to correlate with the prescription collection method. Multiple stepwise logistic regression was used to assess MA with rescue therapy (RT) occurrence as indicators of acute asthma exacerbations. RESULTS: In this study, complete records of 13,482 patients were analyzed. The patients were categorized into three groups: FC (23.2%), PC (72.9%), and NC (3.9%) groups. Those who had PC or NC were more likely to have RT in the following year (19.5% and 9.4%, respectively), compared to FC (5.2%) group. Patients with RT demonstrated higher oral steroid dispensed compared to those without RT (mean (SD) of 319.7 (273.7) vs. 143.6 (175.8) compared to patients without RT (0.78 [0.26] vs. 0.81 [0.29]). Logistic regression analysis revealed that both patients from PC and NC groups had a greater probability of experiencing RT in the following year (partial: 2.364 (1.964-2.847), p < 0.001); no collection: 2.030 (1.318-3.127), p = 0.001). Lower minimum ACT scores (0.317 vs. 1.0) were noted for the PC group and an increase in minimal ACT score of 0.167 for every unit increase in PDC was also observed. CONCLUSION: Patients in the FC group exhibited higher MA and were less likely to receive RT due to their asthma exacerbations in subsequent years.