|
In this JAMA+ AI Conversations, JAMA and JAMA+ AI Associate Editor Yulin Hswen, ScD, MPH, spoke with Stanford’s Nigam Shah, MBBS, PhD, about the challenges clinicians face when navigating electronic health records and how large language models could transform that experience.
Dr Shah describes current EHR use as a kind of "excavation," where clinicians must sift through fragmented, specialist-generated data without effective summarization tools, making information retrieval slow and inefficient. He described "ChatEHR," an LLM-powered interface that is designed to query and summarize a patient’s complete health timeline within a secure, HIPAA-compliant environment.
Together, they consider how tools like ChatEHR could help close both the information gap and the evidence gap, advancing a more connected, learning-oriented health system.
Listen now on Spotify | Apple Podcasts | YouTube | JAMA.com.
Editor’s Picks in this week’s JAMA+ AI:
- A machine learning model using intraoperative molecular imaging accurately estimates lung nodule malignancy in real time, achieving high AUCs (0.865–0.893) and 93.8% sensitivity with 100% specificity in validation. This approach enables rapid “optical biopsy,” improving surgical efficiency and potentially outcomes while reducing reliance on frozen sections. (JAMA Network Open)
- Expert dermatologists (>10 years’ experience) slightly outperformed AI in multiclass accuracy (74.2% vs. 72.2%), while AI matched mid-level clinicians and exceeded novices. AI showed higher specificity in binary tasks, but top sensitivity remained with the most experienced clinicians, supporting its role as a decision-support tool, and not a replacement for expert judgment, in skin cancer diagnosis. (JAMA Dermatology)
- A Viewpoint identifies risks of using AI-driven ambient scribes for billing, including increased health care spending, greater upcoding, and long-term threats to clinician satisfaction and patient trust. The authors call for registries, stronger auditing, and reimbursement adjustments to mitigate financial and ethical risks. (JAMA Health Forum)
|
