JMIR Publications

ABSTRACT

The demands of modern healthcare place a significant burden on physicians, with clinical documentation often cited as a major contributor to workload and burnout. The rise of electronic health records has, in some ways, exacerbated this challenge, requiring physicians to navigate and synthesize extensive clinical data from multiple sources while maintaining adherence to legal and professional standards in clinical documentation. Physicians risk losing focus on direct patient care as a result. In this setting, artificial intelligence (AI) has the potential to enhance the quality of medical records, improve efficiency, and offload tasks from the physician so that they can focus on direct patient care. Although AI offers significant potential in these domains, challenges arise in balancing efficiency and the legal and regulatory requirements associated with healthcare records. Some regulatory bodies place emphasis on capturing the clinical narrative as opposed to reproduction of summarized clinical data into daily progress notes. Here we propose a near-term conceptual framework for language model processing the medical record that could be built into electronic medical systems to fulfill the clinical narrative and to chart clinical data automatically, thus reducing the burden of charting on the physician. This proposed framework preserves the human-based clinical impression while leveraging AI to help provide contextual information.