JMIR Publications released two feature stories in its News and Perspectives section on June 16. Shalini Kathuria Narang's "Can Humanlike Reasoning Be Replicated in Large Language Models for Clinical Decision-Making?" and Sara Novak's "How Health Care Workers Can Manage Digital Fatigue" address artificial intelligence capabilities in diagnostics and the exhaustion faced by medical professionals managing digital systems.
In her article, Narang discusses a recent study comparing the diagnostic reasoning of OpenAI's o1 model with physicians. The study found that the model matched or exceeded human performance across three stages of care: triage on arrival, first contact with a physician, and upon admission. The widest performance gap occurred at initial emergency room triage when information was most limited. Adam Rodman, hospitalist and one of the researchers on the study, said that these results validate the diagnostic performance of the models, but do not mean they are ready to be deployed independently.
Rodman said that while large language models excel at integrating text-based information, actual clinical practice relies heavily on nontext inputs such as visual and auditory cues gathered during physical examinations. He said that LLMs are excellent at synthesizing curated data or collecting verbal information but cannot replace a physician’s ability to physically examine a patient or integrate uncurated sources.
Narang writes that rather than replacing doctors, the future of AI in medicine requires collaborative integration and careful evaluation. The researchers say there is a need for prospective trials in real-world settings to evaluate newer multimodal models safely. Rodman suggests that one promising application is using this technology as a second opinion to catch diagnostic errors before they occur.
Novak explores digital fatigue among health care workers as clinical processes increasingly rely on digital tools. Her article highlights how automation has brought benefits but also led many providers to struggle with continuous demands from digital interfaces and redundant alerts. Citing expert sources including physician Hassan Bencheqroun and researchers Rachel Hoopsick and Audrey Hai, Novak says that administrative burdens can drive clinician burnout—especially when combined with time constraints imposed by fee-for-service health care systems.
Novak outlines strategies for both institutions and individuals to manage these risks. Experts urge healthcare systems to streamline workflows by reducing low-value prompts such as warnings for non-life-threatening allergies, eliminating redundant alerts, restructuring tasks into team-based systems, sharing responsibilities like inbox management, and preventing after-hours work accumulation. She writes that institutions should recognize digital tasks as part of daily workloads, while individuals can protect recovery time through scheduled breaks or delayed email deliveries during off hours. "Digital fatigue needs to be taken seriously," writes Novak, "like you would any other occupational risk."
