Mount Sinai Hospital researchers are moving clinical deterioration models from research to practice, including the implementation of machine learning interventions designed to improve clinical care and patient outcomes. A study documenting this progress was recently published in the journal Critical Care Medicine.
Image credit: thinkhubstudio/Shutterstock.com
The study's main findings showed that hospitalized patients whose care teams received AI-generated alerts about deteriorating health conditions were 43% more likely to receive enhanced care. Additionally, these patients were significantly less likely to die, highlighting the potential benefits of integrating AI into patient monitoring and care processes.
We wanted to see whether rapid alerts created by AI and machine learning trained on different kinds of patient data could help reduce both how often patients need intensive care and their chances of dying in hospital..
Matthew A. Levin, MD, Professor of Anesthesiology, Perioperative and Pain Medicine, and Genetics and Genomic Sciences, Icahn Mount Sinai
Levin, who also serves as director of clinical data science at Mount Sinai Hospital, continued:Traditionally, we have relied on older manual methods, such as the Modified Early Warning Score (MEWS), to predict clinical deterioration. However, our study shows that an automated machine learning algorithm score that triggers an assessment by a provider outperforms these previous methods in accurately predicting this deterioration. Importantly, it could allow for earlier intervention, saving more lives.. “
The nonrandomized, prospective study, conducted at Mount Sinai Hospital in New York, involved 2,740 adult patients admitted to four medical-surgical units. Patients were split into two groups: one group received real-time alerts predicting potential deterioration, sent directly to an “emergency response team” of nurses, doctors or intensivists.
In the other group, alerts were generated but not distributed. In units where alerts were suppressed, patients who met standard deterioration criteria received emergency intervention from the emergency response team.
Further investigation of the intervention group revealed that patients:
- Less likely to die within 30 days
- They are more likely to be prescribed medication to help their heart and circulation, suggesting doctors are acting quickly.
“Our study shows that real-time alerts using machine learning can significantly improve patient outcomes,” said David L. Reich, MD, chancellor and professor at Mount Sinai Hospital and Mount Sinai Queens, and senior study author.
These models support accurate and timely clinical decision making, helping to dispatch the right team to the right patient at the right time. We see these as “augmented intelligence” tools, speeding up in-person clinical assessments by doctors and nurses and facilitating care that keeps patients safe. These are important steps towards our goal of becoming a learning health system..
Horace W. Goldsmith, Professor of Anesthesiology, Department of Artificial Intelligence and Human Health, Icahn Mount Sinai Hospital
The clinical deterioration algorithm study at Mount Sinai Hospital was terminated early due to the COVID-19 pandemic. Nevertheless, the algorithm was subsequently implemented in all step-down units within the hospital. These units accommodate patients who are stable enough to be discharged from the ICU but still require close monitoring and care, and serve as a transitional space before moving to the general hospital area.
The implementation involves a team of intensivists who screen the 15 patients with the highest predictive scores generated by the algorithm each day. These physicians then make treatment recommendations to the doctors and nurses in charge of each patient.
The algorithm becomes more sophisticated as it is continually retrained with data from an ever-growing number of patients, and ongoing evaluation by the intensive care team helps ensure its accuracy, effectively using reinforcement learning to improve the algorithm's predictive capabilities.
Additionally, this clinical deterioration algorithm is just one of 15 additional AI-based clinical decision support tools developed and deployed across the Mount Sinai Health System, further integrating advanced technology into the practice of medicine.
Unless noted, other authors on the paper are Arash Kia, MD, MS, Prem Timsina, PhD, Fu-yuan Cheng, MS, Kim-Anh-Nhi Nguyen, MS, Roopa Kohli-Seth, MD, Yuxia Ouyang, PhD, and Robert Freeman RN, MSN, NE-BC, all of the Icahn School of Medicine at Mount Sinai. Hung-Mo Lin, Sc, is affiliated with Yale University.
Journal References:
Matthew, Alabama, other(2024) Real-time machine learning alerts to prevent escalation of care: a non-randomized clustered pragmatic clinical trial. Intensive Care Medicine. doi.org/10.1097/CCM.0000000000006243
