UB Pharmacy Professor Develops AI Tool to Predict Hospitalizations in Cardiac Patients
A pharmacist-turned-innovator at the University at Buffalo has created an artificial intelligence model designed to identify cardiac patients most at risk of hospitalization. The effort, led by a UB professor with a background in pharmacotherapy and healthcare data, aims to transform how clinicians intervene before patients deteriorate and end up back in the hospital.
The Challenge: Readmissions and Preventable Hospitalizations
Hospital readmissions among cardiac patients remain a persistent challenge for healthcare systems, driving costs and often signaling gaps in outpatient management. Discharged patients frequently return with conditions that could have been managed with timely, proactive care. The UB project seeks to shift care from reactive to proactive, using data rather than gut feeling to flag those at highest risk.
How the AI Model Works
The model combines demographic data, medical history, comorbidities, medication use, and real-time clinical indicators to predict 30-day and 90-day hospitalization risk. By analyzing patterns in electronic health records and pharmacy data, the tool provides clinicians with a risk score and key contributing factors for each patient. This enables targeted interventions—such as medication reconciliation, patient education, remote monitoring, and early follow-up appointments.
Key Data Points
- Recent hospitalizations and emergency department visits
- Cardiovascular risk factors like hypertension, diabetes, and hyperlipidemia
- Medication adherence indicators and potential drug interactions
- Social determinants of health that may affect access to care
Clinical Impact and Potential Benefits
For clinicians, the AI model offers a risk stratification tool that can be integrated into existing workflows. By focusing attention on high-risk patients, healthcare teams can implement preventive strategies, such as closer follow-up, home-based care checks, and timely therapy adjustments. Reducing preventable hospitalizations not only improves patient outcomes but also lowers the burden on hospital systems and caregivers.
Safety, Ethics, and Validation
The UB team emphasizes that the model is a decision-support tool, not a replacement for clinician judgment. Rigorous validation, ongoing monitoring for bias, and secure handling of patient data are central to the development process. The researchers plan prospective studies to confirm the model’s accuracy across diverse patient populations and care settings.
Collaboration and Future Directions
Developed within UB’s robust academic environment, the AI project brings together pharmacists, clinicians, data scientists, and IT specialists. The team envisions expanding the model to incorporate wearable data, such as heart rate and activity levels, and to adapt the approach for other high-risk conditions beyond cardiac care. Partnerships with healthcare systems could pave the way for wider deployment in the near term.
Why This Matters Now
As healthcare systems grapple with rising costs and an emphasis on value-based care, predictive analytics offers a practical path to prevent deterioration. Early identification of at-risk cardiac patients aligns with broader goals of improving outcomes, enhancing quality of life, and supporting families who manage chronic conditions.
About the Research Team
The project is led by a UB pharmacy professor whose expertise spans pharmacotherapy, health economics, and healthcare analytics. The team combines clinical insight with advanced data science to translate complex information into actionable patient care plans.
What’s Next for the AI Initiative
The researchers plan to publish detailed findings, release performance metrics, and share implementation guidelines for health systems interested in adopting similar predictive tools. With continued refinement and ethical safeguards, the model could become a standard component of cardiac care management, helping clinicians anticipate needs and prevent avoidable hospitalizations.
Conclusion
By applying AI to the problem of hospitalization risk in at-risk cardiac patients, UB is contributing to a future where proactive, data-informed interventions reduce readmissions and improve patient outcomes. The initiative exemplifies how academia can translate theoretical analytics into tangible benefits for patients and healthcare providers alike.
