A ground-breaking multi-hospital study by Mount Sinai Health System reveals that AI can predict hospital admissions hours in advance, promising improved patient care and efficiency in emergency departments.
A recent study conducted by the Mount Sinai Health System reveals that artificial intelligence can help emergency department (ED) teams better anticipate which patients will need hospital admission. The AI model achieved this feat hours earlier than current methods, significantly improving patient care and reducing overcrowding and “boarding,” a scenario where admitted patients remain in the ED due to a lack of available beds.
This pioneering research, published in the journal Mayo Clinic Proceedings: Digital Health, marks one of the most extensive evaluations of AI in emergency settings to date.
The study involved over 500 ED nurses across seven hospitals and analyzed data from more than 1 million patient visits to train a machine learning model.
“Emergency department overcrowding and boarding have become a national crisis, affecting everything from patient outcomes to financial performance. Industries like airlines and hotels use bookings to forecast demand and plan. In the ED, we don’t have reservations. Could you imagine airlines and hotels without reservations, solely forecasting and planning from historical trends? Welcome to health care,” lead author Jonathan Nover, a vice president of nursing and emergency services at Mount Sinai Health System, said in a news release.
“Our goal was to see if AI, combined with input from our nurses, could help hasten admission planning, a reservation of sorts. We developed a tool to forecast admissions needs before an order is placed, offering insights that could fundamentally improve how hospitals manage patient flow, leading to better outcomes,” he added.
The research assessed nearly 50,000 patient visits across Mount Sinai’s urban and suburban hospitals, showing that the AI model performed consistently well in these varied settings. Surprisingly, the combination of human and machine predictions did not significantly improve accuracy, indicating the AI system alone was a strong predictor.
“We wanted to design a model that doesn’t just perform well in theory but can actually support decision-making on the front lines of care,” added co-corresponding senior author Eyal Klang, chief of generative AI in the Windreich Department of Artificial Intelligence and Human Health at the Icahn School of Medicine at Mount Sinai. “By training the algorithm on more than a million patient visits, we aimed to capture meaningful patterns that could help anticipate admissions earlier than traditional methods. The strength of this approach is its ability to turn complex data into timely, actionable insights for clinical teams — freeing them up to focus less on logistics and more on delivering the personal, compassionate care that only humans can provide.”
While the pilot study was restricted to a two-month period and within one health system, the team hopes these findings will spark further live clinical testing.
The next phase involves integrating the AI model into real-time workflows, monitoring outcomes such as reduced boarding times, improved patient flow and operational efficiency.
“We were encouraged to see that AI could stand on its own in making complex predictions. But just as important, this study highlights the vital role of our nurses — more than 500 participated directly — demonstrating how human expertise and machine learning can work hand in hand to reimagine care delivery,” added co-corresponding senior author Robbie Freeman, chief digital transformation officer at Mount Sinai Health System. “This tool isn’t about replacing clinicians; it’s about supporting them. By predicting admissions earlier, we can give care teams the time they need to plan, coordinate and ultimately provide better, more compassionate care. It’s inspiring to see AI emerge not as a futuristic idea but as a practical, real-world solution shaped by the people delivering care every day.”
Source: Mount Sinai