A University at Buffalo pharmacy professor has developed an artificial intelligence model that predicts hospital readmission within 90 days for cardiac patients with 95% accuracy, according to findings published in the December 2025 issue of British Medical Journal Health and Care Informatics.
Arinze Nkemdirim Okere, PharmD, MBA, clinical professor and head of the Divisions of Outcomes and Practice Advancement at UB’s School of Pharmacy and Pharmaceutical Sciences, created the machine-learning model between July 2021 and December 2022. Okere, who joined UB in September 2025, was inspired by his experience as a hospital pharmacist in Tallahassee, Fla., where he observed that discharged patients frequently returned to the hospital for preventable issues. “The reasons behind hospital readmissions are multifactorial,” Okere says. “One big factor is medication adherence issues. This is especially true among patients with cardiovascular disease.”
Study Design and Findings
The research team, which included Md. Mohaimenul Islam, research assistant professor in UB’s Department of Pharmacy Practice, recruited more than 1,300 adults from community pharmacies, outpatient clinics and social media platforms. All participants had at least one cardiovascular risk factor such as high blood pressure, high cholesterol or Type 2 diabetes. Of the participants, 35% reported at least one hospitalization and 10.4% reported a 90-day readmission. The AI model identified heart disease, multiple medications, race/ethnicity, employment and insurance status as the most influential predictors of 90-day readmission.
Medication-Related Risk Factors
Common reasons patients return to hospitals include medication allergies, missed doses, adverse drug interactions and medication misunderstanding. Okere noted that some patients prescribed heart-failure medications weren’t taking them, particularly in underserved communities lacking primary care physicians. Poor medication reconciliation during care transitions can also lead to omissions, duplications or inappropriate continuation of medications. Patients often misunderstand medication instructions, discontinuing treatments once they feel better rather than maintaining long-term adherence.
Future Implementation
Okere plans to collaborate with Buffalo-area hospitals to implement the AI system in clinical practice. “We are hoping that physicians, and even the nursing assistants and triage nurses, can use this AI system to quickly flag a patient who might be at risk of being readmitted,” he said.
This article is an AI-assisted summary. All facts and figures are drawn from the original report: https://www.buffalo.edu/news/releases/2026/01/Okere-hospitalization-study.html