Researchers at Mount Sinai Hospital are moving clinical deterioration models from the research phase to practical application. This involves deploying a machine learning intervention designed to enhance clinical care and patient outcomes. The study documenting this advancement was recently published in the journal Critical Care Medicine.
Image Credit: thinkhubstudio/Shutterstock.com
The main findings of the study indicated that hospitalized patients whose care teams received AI-generated alerts about adverse changes in their health were 43 percent more likely to have their care escalated. Additionally, these patients were significantly less likely to die, underscoring the potential benefits of integrating AI into patient monitoring and care processes.
We wanted to see if quick alerts made by AI and machine learning, trained on many different types of patient data, could help reduce both how often patients need intensive care and their chances of dying in the hospital.
Matthew A. Levin, MD, Professor, Anesthesiology, Perioperative and Pain Medicine, and Genetics and Genomic Sciences, Icahn Mount Sinai
Levin, who is also the 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 automated machine learning algorithm scores that trigger evaluation by the provider can outperform these earlier methods in accurately predicting this decline. Importantly, it allows for earlier intervention, which could save more lives.”
The non-randomized, prospective study conducted at the Mount Sinai Hospital in New York involved 2,740 adult patients admitted to four medical-surgical units. The patients were divided into two groups: one received real-time alerts predicting the likelihood of deterioration, which were sent directly to their nurses, physicians, or a "rapid response team" of intensive care specialists.
The other group had alerts generated but not delivered. In the units where alerts were suppressed, patients who met standard deterioration criteria still received urgent interventions from the rapid response team.
Further research in the intervention group revealed that patients:
- Were less likely to pass away within 30 days
- More likely to receive medications to support the heart and circulation, suggesting that doctors were acting quickly
David L. Reich, MD, President, Professor, and Senior Study Author at the Mount Sinai Hospital and Mount Sinai Queens, said, “Our research shows that real-time alerts using machine learning can substantially improve patient outcomes.”
These models are accurate and timely aids to clinical decision-making that help us bring the right team to the right patient at the right time. We think of these as ‘augmented intelligence’ tools that speed in-person clinical evaluations by our physicians and nurses and prompt the treatments that keep our patients safer. These are key steps toward the goal of becoming a learning health system.
Horace W. Goldsmith, Professor, Anesthesiology, Department of Artificial Intelligence and Human Health, Icahn Mount Sinai
The study on the clinical deterioration algorithm at Mount Sinai Hospital was terminated early due to the COVID-19 pandemic. Despite this, the algorithm has since been implemented across all step-down units within the hospital. These units cater to patients who are stable enough to leave the ICU but still require close monitoring and care, serving as a transitional space before moving to a general hospital area.
The implementation involves a team of intensive care physicians who review the 15 patients with the highest prediction scores generated by the algorithm daily. These physicians then make treatment recommendations to the doctors and nurses responsible for each patient.
As the algorithm is continually retrained with data from a growing number of patients, it becomes more refined. The ongoing assessments by the intensive care team help ensure its accuracy, effectively using reinforcement learning to improve the algorithm's predictive capabilities.
Moreover, this clinical deterioration algorithm is just one of 15 additional AI-based clinical decision support tools that have been developed and deployed throughout the Mount Sinai Health System, further integrating advanced technology into healthcare practices.
The other authors of the paper from Icahn School of Medicine at Mount Sinai, except as noted, include Arash Kia, MD, MSc; 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. Hung-Mo Lin, ScD, is affiliated with Yale University.
Journal Reference:
Mathew, A. L., et al. (2024) Real-Time Machine Learning Alerts to Prevent Escalation of Care: A Nonrandomized Clustered Pragmatic Clinical Trial. Critical Care Medicine. doi.org/10.1097/CCM.0000000000006243