Deploying and evaluating a machine learning intervention to improve clinical care and patient outcomes is vital in advancing clinical deterioration models from theory to practice, as highlighted in a recent editorial and study published in Critical Care Medicine. The study, conducted at The Mount Sinai Hospital in New York, demonstrated that patients who received AI-generated alerts signaling adverse changes in their health were 43 percent more likely to have their care escalated, leading to significantly lower mortality rates. Lead study author Matthew A. Levin emphasized the importance of utilizing automated machine learning algorithms to predict clinical decline accurately, allowing for earlier intervention and potentially saving more lives compared to traditional methods like the Modified Early Warning Score (MEWS).
The non-randomized, prospective study included 2,740 adult patients admitted to medical-surgical units at The Mount Sinai Hospital. Patients were divided into two groups: one receiving real-time alerts based on the predicted likelihood of deterioration, and another group with alerts created but not sent. Patients in the intervention group were more likely to receive early interventions to support heart and circulation and had a lower 30-day mortality rate. Senior study author David L. Reich highlighted the effectiveness of real-time alerts in aiding clinical decision-making and optimizing patient outcomes, emphasizing the role of these tools as “augmented intelligence” to expedite evaluations and prompt appropriate treatments.
Due to the COVID-19 pandemic, the study was terminated early, but the algorithm has been implemented in all stepdown units within The Mount Sinai Hospital. A stepdown unit serves as an intermediary level of care between the ICU and general hospital areas, ensuring close monitoring for stable patients. Intensive care physicians visit the top 15 patients with the highest prediction scores daily, providing treatment recommendations to the care team. As the algorithm continues to be refined through reinforcement learning and evaluation by intensive care physicians, it becomes more accurate in predicting patient deterioration and facilitating timely interventions.
In addition to the clinical deterioration algorithm, The Mount Sinai researchers have developed and deployed 15 AI-based clinical decision support tools across the Mount Sinai Health System. These tools aim to enhance clinical decision-making and optimize patient care in various medical settings. The research team’s efforts culminated in the paper titled “Real-Time Machine Learning Alerts to Prevent Escalation of Care: A Nonrandomized Clustered Pragmatic Clinical Trial,” underscoring the importance of leveraging artificial intelligence in healthcare to improve patient outcomes and streamline clinical workflows. Collaborative efforts in implementing AI technologies in healthcare settings can potentially revolutionize patient care delivery and contribute to better overall outcomes for patients.
