Predicting Adverse Events in Senior Care Settings Using Multi Source EHR and Operations Data: A Responsible Al Pipeline
  • Author(s): Nicholas Donkor; Munashe Naphtali Mupa; Zainab Mugenyi; Kwame Ofori Boakye; Hilton Hatitye Chisora
  • Paper ID: 1712619
  • Page: 580-589
  • Published Date: 08-12-2025
  • Published In: Iconic Research And Engineering Journals
  • Publisher: IRE Journals
  • e-ISSN: 2456-8880
  • Volume/Issue: Volume 9 Issue 6 December-2025
Abstract

Falls, unplanned transfers, and readmissions are some of the adverse events threatening patient safety in senior care due to the complex comorbidities and aging populations. Artificial intelligence (AI) provides the opportunities to combine electronic health record (EHR) and operational information, followed by proactive prevention. However, disjointed systems and cloudy algorithms are obstacles to adoption and ethical issues on fairness and accountability. The present research formulated a responsible AI pipeline that incorporates the vitals, medication history, nurse notes, and staffing data to forecast falls and unexpected transfers. The model was based on explainability and fairness, which was achieved through the use of machine learning, natural language processing, and bias auditing (Peng, 2025; Kalu-Mba et al., 2025). The results measured by validation using a stepped-wedge design included falls per 1000 resident days and 30-day readmissions. The findings showed a better predictive accuracy (AUC>0.85), understandable SHAP-generated insights, and nurse-actionable dashboards. In general, the strategy improved patient safety, operational efficiency, and clinical decision-making in senior care environments.

Citations

IRE Journals:
Nicholas Donkor, Munashe Naphtali Mupa, Zainab Mugenyi, Kwame Ofori Boakye, Hilton Hatitye Chisora "Predicting Adverse Events in Senior Care Settings Using Multi Source EHR and Operations Data: A Responsible Al Pipeline" Iconic Research And Engineering Journals Volume 9 Issue 6 2025 Page 580-589

IEEE:
Nicholas Donkor, Munashe Naphtali Mupa, Zainab Mugenyi, Kwame Ofori Boakye, Hilton Hatitye Chisora "Predicting Adverse Events in Senior Care Settings Using Multi Source EHR and Operations Data: A Responsible Al Pipeline" Iconic Research And Engineering Journals, 9(6)