‘AI doctor’ reads radiology reports, other physician notes to predict patient outcomes
Scientists at NYU Langone Health have created an artificial intelligence computer program that can predict patients’ outcomes based on radiology reports and other inputs.
Their large language model, dubbed NYUTron, was trained to make its assessments using unaltered text from electronic health records. It was able to predict 80% of patients who were readmitted, a 5% improvement over a standard approach requiring reformatted medical data, according to a study published June 7 in Nature [1].
“These results demonstrate that large language models make the development of ‘smart hospitals’ not only a possibility, but a reality,” senior study author Eric K. Oermann, MD, an assistant professor in NYU’s Department of Radiology, said in an announcement. “Since NYUTron reads information taken directly from the electronic health record, its predictive models can be easily built and quickly implemented through the healthcare system.”
Scientists trained the system with millions of clinical notes from the system’s EHR, representing 336,000 patients treated at NYU Langone between 2011-2020. The final 4.1 billion-word language cloud included radiology reports, patient progress notes, discharge instructions and any other record written by a physician.
NYUTron was able to identify 85% of those who died in the hospital (a 7% improvement over standard approaches) and the precise length of stay for 79% of patients (a 12% improvement). It additionally assessed likelihood of comorbidities accompanying the primary disease, along with odds of an insurance denial.
Oermann and colleagues emphasized that NYUTron is not intended to replace provider’s decision-making, only to support it. Future investigations could explore the AI system’s ability to extract billing codes, predict infections or identify the right medication to order.
“Our findings highlight the potential for using large language models to guide physicians about patient care,” study author Lavender Jiang, a doctoral student at the NYU Center for Data Science, said in the announcement.
Read much more at the link below.
Marty Stempniak has covered healthcare since 2012, with his byline appearing in the American Hospital Association's member magazine, Modern Healthcare and McKnight's. Prior to that, he wrote about village government and local business for his hometown newspaper in Oak Park, Illinois. He won a Peter Lisagor and Gold EXCEL awards in 2017 for his coverage of the opioid epidemic.