Home / Health / AI Drafts Patient Summaries: Doctors Prefer It
AI Drafts Patient Summaries: Doctors Prefer It
3 Nov
Summary
- AI-generated patient discharge summaries were preferred by doctors 95% of the time.
- AI summaries scored higher in overall quality, readability, accuracy, and completeness.
- Physicians still need to review and verify AI-generated discharge drafts.
Artificial intelligence, specifically large language models (LLMs), shows promise in enhancing the efficiency and accuracy of hospital patient discharge summaries. A study published in npj Digital Medicine explored LLM use for complex hospital stays, where summarizing extensive medical information can be burdensome for providers.
Clinicians helped tailor the AI approach to ensure it organized patient records and extracted core information for discharge summaries within minutes. Researchers compared AI-generated and human-authored summaries for 60 hospitalizations, finding that physician reviewers preferred the AI versions in 95% of cases.
These AI-generated summaries received higher scores for overall quality, readability, accuracy, and completeness, with fewer missing clinical details. Primary care physicians also reported finding the AI summaries easier to understand, explain, and more useful for ongoing care.
Despite the advancements, the study acknowledged that both AI and human-authored summaries contained errors. However, no significant difference was found in the potential severity or harm associated with these errors. The researchers stress that physician oversight remains crucial for any AI-generated discharge documentation.
