A multihospital study by the New York-based Mount Sinai Health System suggests that AI can predict which emergency department patients will require admission hours before traditional methods.

Researchers tested a machine learning model trained on data from more than 1 million prior visits across seven hospitals. Over a two-month period, they compared the model’s predictions with triage assessments from more than 500 emergency nurses, reviewing nearly 50,000 patient encounters.

The model performed consistently across urban and suburban hospital settings and proved to be a strong stand-alone predictor. Combining its output with nurse assessments did not significantly improve accuracy, the team reported.
“Emergency department overcrowding and boarding have become a national crisis,” said Jonathan Nover, MBA, RN, vice president of nursing and emergency services at Mount Sinai and lead author of the study. “We developed a tool to forecast admissions needs before an order is placed, offering insights that could fundamentally improve how hospitals manage patient flow.”

Dr. Eyal Klang, chief of generative AI at the Icahn School of Medicine at Mount Sinai and co-corresponding senior author, said the model was designed to capture patterns in patient data that could guide earlier admission planning. “The strength of this approach is its ability to turn complex data into timely, actionable insights for clinical teams,” he said.

The next phase will integrate the system into real-time clinical workflows to assess its impact on boarding times, patient flow, and operational efficiency.

The findings were published July 9 in Mayo Clinic Proceedings: Digital Health. The study was supported by the Clinical and Translational Science Awards program, the National Center for Advancing Translational Sciences, and the National Institutes of Health.

Artificial Intelligence Homepage