Artificial Intelligence in the Management of Polypharmacy Among Older Adults: A Scoping Review

The advent of artificial intelligence (AI) presents an opportunity to enhance the management of multiple medications for patients. Despite the growing body of research on the clinical impact and accuracy of AI in healthcare, current literature is limited in addressing AI's role in medication management for older adults. The goal of this scoping review was to assess the extent and applicability of AI applications in managing multiple pharmacological therapies among adults aged 50 and older. Relevant articles were searched in EMBASE, Ovid MEDLINE, and Web of Science databases. Search phrases included "polypharmacy/polytherapy", "artificial intelligence in healthcare", "machine learning/language learning", and "elderly populations". The search initially identified 58 citations. After a systemized and rigorous screening process, 12 articles were further screened for eligibility and critically appraised for bias and appropriateness. Of those, a total of five articles were retained for the final analysis. They focused on the application of AI and web-based applications to reduce inappropriate medication use and drug interactions, how machine learning and audio-based activity recognition systems could improve medication adherence among older adults, and recognizable patterns between multimorbidity and polypharmacy in elderly populations with chronic illnesses. The main findings of this review suggested that AI tools have demonstrated efficiency and accuracy in eliminating drug-drug interactions, assisting in the detection of potentially inappropriate medications (PIMs), and identifying patterns of multimorbidity due to polypharmacy in older adults. Moreover, AI tools were considered easy to use and helped enhance medication adherence. Results suggest that AI has potential in managing polypharmacy, particularly in enhancing medication safety, improving adherence, and predicting risk factors for medication-related errors in older adults. Future investigations are needed that focus on AI's long-term effect on patient outcomes, its role in personalized pharmacotherapy, and ongoing challenges related to algorithmic transparency, bias mitigation, and regulatory oversight. As AI continues to evolve, its integration into healthcare would benefit from methodical evaluation, stringent oversight, and interdisciplinary collaboration to ensure its safe and effective deployment in patient care.