New “AI Doctor” Predicts Hospital Readmission and Other Health Outcomes
A new artificial intelligence (AI) tool developed by NYU Grossman School of Medicine researchers can read physicians’ notes and accurately estimate patients’ risk of death, length of hospital stay, and other important outcomes related to their care. The tool, NYUTron, is a large language model (LLM) that can “learn” from text without needing specially formatted data, making it ideal for use in healthcare, where patient records are written in creative, individualized language.
In a study published in Nature, the research team showed that NYUTron could predict 80% of those readmitted to the hospital within a month, a roughly 5% improvement over a standard, non-LLM computer model requiring reformatting medical data.
NYUTron could also identify 85% of those who died in the hospital (a 7% improvement over standard methods) and estimated 79% of patients’ actual length of stay (a 12% improvement over the standard model). The tool also successfully assessed the likelihood of additional conditions accompanying a primary disease (comorbidity index) and the chances of insurance denial.
The researchers say that NYUTron could be used to help physicians make better decisions about patient care and to identify patients who are at risk of adverse outcomes. The tool could also be used to develop new healthcare policies and procedures.
NYUTron is still in the early stages of development, but the researchers are optimistic about its potential to improve healthcare.
New “AI Doctor” Predicts Hospital Readmission and Other Health Outcomes (Jiang, Nature, 6/7).
