A new AI model helps hospitals identify patients who will need skilled nursing care after discharge, improving planning and reducing stress for patients and caregivers
Researchers at NYU Langone Health developed an AI tool that forecasts which patients may require skilled nursing care after discharge, potentially improving planning and reducing stress, according to a study in npj Health Systems.
The study addresses skilled nursing facilities, which provide short-term, intensive care and rehabilitation services for patients recovering from an illness or surgery. According to the authors, about 15% of patients from NYU Langone are discharged to skilled nursing facilities.
How the AI tool works
The researchers analysed electronic health records of 4,000 patients admitted to general medicine services at NYU Langone. They focused on the history and physical admission notes, which contain data on a patient’s health, functional ability, and social situation.
The researchers developed a generative AI model that reads each admission note, identifies and extracts information on seven specific risk factors, and organises them into a concise ‘AI Risk Snapshot’, summarising the relevant details for discharge planning. The researchers then tested nine different AI models to see which could best predict a patient’s discharge destination. They found that a model using short, AI-generated summaries of doctor notes was more accurate than models relying on the full, lengthy notes.
To check the AI tool’s accuracy, the researchers compared its predictions to those of human experts and found strong alignment with AI’s risk scores. The researchers demonstrated that the AI tool predicts, with 88% accuracy, whether patients will need skilled nursing care after their hospital stays.
“Our two-step approach acts like a fast, careful reader, turning a complex medical note into a simple summary of what matters most for discharge planning,” said senior study author Yindalon Aphinyanaphongs, MD, PhD, director of operational data science and machine learning for NYU Langone, and a research professor in the Departments of Population Health and Medicine at NYU Grossman School of Medicine.
“Our next step is to test this model in a real-world clinical setting to see if it helps our care teams plan discharges more effectively across all patients,” said first author William R. Small, MD, a clinical assistant professor in the Department of Medicine. “We will also monitor the system to ensure it is fair and safe and helps to improve patient care.”
Why is predicting discharge destination important?
Early care teams play a critical role in identifying hospital inpatients who are likely to require ongoing support at skilled nursing facilities (SNFs) after discharge. By making these determinations early in a patient’s hospital stay, care teams can proactively coordinate with SNFs, arrange transportation, prepare necessary documentation, and communicate with patients and their families. This facilitates a smoother transition, reduces the risk of discharge delays, and ensures that patients receive appropriate care as soon as they leave the hospital.
