Researchers from New York City-based NYU Langone Health licensed OpenAI’s GPT-4 tool for clinician use in March 2023. Clinicians used the tool experimentally and adhered to privacy rules surrounding patient data, according to a July 31 news release from the health system.
Researchers had GPT-4 generate easier to read reports from 100 echocardiogram analyses written by physicians. Five board-certified echocardiographers then reviewed the GPT-4 reports for accuracy, relevance and understandability. Researchers asked non-clinicians to evaluate both sets of reports as well, the release said.
The echocardiographers agreed or strongly agreed that 73% of the AI-generated reports could be sent to patients as is, without needing human oversight. The study also noted that 16% of the AI-generated reports contained inaccuracies.
Non-clinician reviewers said 97% of AI-generated reports were easier to understand compared to the reports written by physicians, the release said.
The study authors said echocardiogram reports can be difficult for patients to interpret, leading to unnecessary health anxiety if patients see test results before their physicians.
“If dependable enough, AI tools could help clinicians explain results at the moment they are released,” Jacob Martin, MD, study author and cardiology fellow at NYU Langone, said in the release.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
