How will machine learning impact clinical care? Study finds that AI-generated alerts can dramatically improve patient outcomes
The study, published in the journal Critical Care Medicine, found that hospitalized patients whose care teams received real-time AI alerts were 43% more likely to receive enhanced care and were significantly less likely to die.
Lead author Dr. Matthew A. Levin, professor of anesthesiology and genetics at Icahn Mount Sinai Hospital, highlighted the importance of this innovation, saying: “We wanted to see whether rapid alerts powered by AI and machine learning trained on different types of patient data could help reduce both how often patients need intensive care and their chances of dying in hospital.”
AI and Healthcare
Healthcare providers have relied on manual methods, such as the Modified Early Warning Score (MEWS), to predict patient deterioration, but this study shows that AI can outperform these older methods, allowing for earlier intervention and potentially saving more lives.
The study was a nonrandomized prospective study of 2,740 adult patients admitted to four medical-surgical units at Mount Sinai Hospital. Patients were divided into two groups: those who received real-time alerts about possible deterioration in their health, and those in which alerts were generated but not sent. Patients in units where alerts were suppressed received emergency intervention if they met standard criteria for deterioration.
Key findings from the intervention group included an increased likelihood of patients taking medications to support their heart and circulation, and a reduced 30-day mortality rate. Dr. David L. Reich, president of Mount Sinai Hospital and senior author, highlighted the benefits, saying, “We wanted to see whether rapid alerts from AI and machine learning trained on different types of patient data could help reduce both how often patients need intensive care and their likelihood of dying in the hospital.”
AI-based clinical decision making
Although the study was halted due to the COVID-19 pandemic, its success led to the algorithm being implemented in all of Mount Sinai's step-down units. These specialized units closely monitor patients who are stable but still require significant care, bridging the gap between intensive care units and general hospital areas.
Each day, an intensivist visits the 15 patients with the highest prediction scores and makes treatment recommendations to the attending physicians and nurses. This process allows the algorithm to be retrained on larger patient datasets over time, continually improving through reinforcement learning.
Beyond the clinical deterioration algorithm, Mount Sinai researchers developed and deployed 15 additional AI-based clinical decision support tools across the health system. The study, titled “Real-time Machine Learning Alerts to Prevent Escalation of Care: A Non-Randomized Clustered Pragmatic Clinical Trial,” involved several experts from Icahn Mount Sinai and Yale University.
Editor's Recommendation article
