New York City-based NYU Langone Health developed an AI model that predicts with 88% accuracy whether patients will require skilled nursing care after hospital discharge.
The two-part tool analyzes admission notes to identify seven key risk factors. It then generates brief summaries — 94% shorter than the original notes — for a second model to evaluate post-discharge needs, according to a Jan. 28 news release from the health system.
In the study, published Jan. 9 in npj Health Systems, researchers reviewed EHRs for 4,000 patients admitted to NYU Langone’s general medicine services. They found that patients who received a high-risk score from the model were 13.5 times more likely to be independently flagged by a nurse case manager as needing skilled nursing care.
The study also noted that about 15% of NYU Langone patients are discharged to skilled nursing facilities.
