Health Services/Population Health

31. Predicting Medication Adherence and Hospital Readmission Rate in Chronic Disease Patients: An AI Strategy

Introduction: For chronic diseases such as high blood pressure, heart failure, cancer, diabetes, and depression, more than half of these patients with prescribed medications fail to comply or refill doses as agreed with their doctor. Non-adherence to medication has contributed to increased healthcare utilization by chronic disease patients, high hospital readmission rates, and preventable medical costs globally.

Methods: We explored the publicly available Medication Possession Ratio (MPR) data which looks at how often a person refills their prescription for over 600 patients from males, females, teenagers, adults, and the elderly within a 4-year period of 2016 to 2020. Data consisted of filled prescriptions, medication type, dosage form, and hospital admission. We predicted the risk of medication nonadherence and hospital readmission for the first 3 months after a patient fills their prescription.

Results: Our data generated probability-based risk scores based on patients’ diagnosis codes. 33% of patients showed nonadherence due to procrastination and inattention. 39% simply forgot to complete their doses, while 9% procrastinated getting refills and 20% did not renew their prescription on time. Retrospective data from express scripts showed that 69% skipped medication and failed to complete dosages simply because of inattention and forgetfulness. 16% due to cost implication and 15% due to clinical concerns of medication not working.

Impact: This research hints at the tremendous amount of work that needs to be done in order to improve medication nonadherence in chronic disease patients to further reduce hospital readmission rate, and most importantly drive health care costs to the barest minimum.

Organization: Saint Louis University

Okoye V, Emmanuel O, Okoye G