AI Shifts Doctor’s Role From Diagnoser to Navigator in Patient care
SAN FRANCISCO – Artificial intelligence is poised to reshape the practice of medicine, but not by replacing doctors, according to emerging insights. Instead of seeking definitive diagnoses from AI, physicians are finding value in utilizing the technology as a refined “wayfinding” tool – assisting with identifying patterns, surfacing overlooked details, and supporting complete patient care. This shift acknowledges the nuanced reality that medical cases often lack neat solutions and that effective care extends beyond simply identifying a disease.
The evolving role of AI in healthcare stems from the limitations of applying a purely diagnostic approach to complex medical scenarios. As one physician noted, sometimes the root cause of a patient’s ailment remains unclear even after treatment, yet the patient still improves. Asking an AI to pinpoint the exact diagnosis in such cases offers little practical benefit.
Gurpreet Dhaliwal, a physician at the University of California, San Francisco, described by the New York Times as “one of the most skillful clinical diagnosticians in practice,” frames the issue as “starting with the end.” He advocates for AI’s use in identifying trends in a patient’s medical journey and highlighting possibly missed facts.”Just as you have a Japanese-English dictionary in yoru desk doesn’t mean you’re fluent in Japanese,” Dhaliwal explained, emphasizing that AI should augment, not supplant, a doctor’s expertise.
Tools like OpenEvidence,which provides access to peer-reviewed medical literature and adheres to patient privacy regulations like HIPAA,are already being adopted by clinicians. Unlike some AI platforms,OpenEvidence prioritizes clarifying questions before offering potential insights,avoiding the risk of “hallucinations” – generating inaccurate or fabricated information. While platforms like ChatGPT are often blocked in hospital networks, tools like OpenEvidence demonstrate a practical path forward for integrating AI into patient care, focusing on support and information access rather than automated diagnosis.