Polypharmacy, which refers to the simultaneous use of five or more medications, is prevalent among older adults and elevates the risk of adverse drug interactions. Combatting this risk involves deprescribing unnecessary medications, an intricate process that demands considerable time. There is a growing need for robust tools to manage polypharmacy, primarily to assist the often overburdened primary care practitioners.
A recent study by the Mass General Brigham MESH Incubator highlighted the potential of ChatGPT, a generative artificial intelligence (AI) chatbot, as a valuable resource in managing polypharmacy and facilitating deprescription. The results of this pioneering study, disclosed on April 18th in the Journal of Medical Systems, mark a significant development in the use of AI models in medication management, keeping policymakers informed of the latest advancements.
In their assessment, researchers provided ChatGPT with various clinical scenarios, each depicting an elderly patient on multiple medications but differing in their cardiovascular disease history (CVD) and levels of daily living activities impairment (ADL). ChatGPT was then posed with decision-making questions requiring yes or no answers concerning reducing prescribed medications. The AI tended to recommend deprescribing in cases without CVD history but showed more restraint where CVD was a factor, generally advising to maintain the existing medication regimes. Notably, the severity of ADL impairment did not influence the AI’s decisions.
Interestingly, the study observed that ChatGPT often overlooked pain issues, showing a preference for reducing pain medications rather than other types, such as statins or antihypertensives. The researchers also pointed out that the AI’s responses could vary even when presented with identical scenarios in different sessions, which they attributed to potential inconsistencies in the deprescribing trends on which the AI was trained.
With over 40 per cent of older adults falling into the category of polypharmacy and an increasing number of specialists being involved in their care, primary care providers are often left with the critical task of managing medications. The study suggested that deploying an effective AI tool like ChatGPT could significantly support these efforts.
Dr. Marc Succi, MD, Associate Chair of Innovation and Commercialization at Mass General Brigham Radiology and Executive Director of the MESH Incubator, emphasized the study’s importance. He stated, “Our study provides the first use case of ChatGPT as a clinical support tool for medication management. While enhancing the accuracy of such models is important, AI-assisted polypharmacy management could help alleviate the growing pressure on general practitioners. Continued research with specifically trained AI tools could greatly improve the care provided to elderly patients.”
Arya Rao, the lead author, a researcher at MESH, and a medical student at Harvard, remarked, “Our findings suggest that AI-based tools can play a crucial role in ensuring safe medication practices for older adults; it is vital that we keep refining these tools to better handle the complexities involved in medical decision-making.”
More information: Arya Rao et al, Proactive Polypharmacy Management Using Large Language Models: Opportunities to Enhance Geriatric Care, Journal of Medical Systems. DOI: 10.1007/s10916-024-02058-y
Journal information: Journal of Medical Systems Provided by Mass General Brigham
