Hospitals turn to AI chatbots to sift through sprawling patient records
A Stanford physician used the AI tool ChatEHR to locate a prior cancer diagnosis that clarified a puzzling lymph-node biopsy, highlighting how health systems are deploying large-language-model chatbots to navigate complex electronic records.
A team of six pathologists failed to identify the source of a cancer detected in a lymph-node biopsy, even after 70 staining attempts. Seeking help, a Stanford physician turned to ChatEHR, a generative-AI chatbot designed to query electronic health records. After a brief exchange, the system revealed that the patient had previously been diagnosed with sarcomatoid squamous cell carcinoma at a different institution, a fact that fully accounted for the current findings.
The doctor lauded the tool, saying its success proved its value. This episode reflects a growing trend among health systems to roll out AI-powered chatbots—both homegrown and from vendors—to summarize and search increasingly bulky patient records. While diagnostic assistance is a clear benefit, providers anticipate broader applications for routine care. The article appears exclusively for one outlet+ subscribers, with additional analysis available behind a paywall.
Why it matters
AI chatbots can quickly surface critical medical history, helping clinicians make faster, more accurate decisions.
In this story
