A groundbreaking study conducted at St. Michael’s Hospital in Toronto revealed that an artificial intelligence tool significantly reduced unexpected deaths among hospitalized patients. The AI system, called Chartwatch, was developed by the Unity Health AI team and implemented in October 2020 after years of rigorous testing. The technology monitored patients’ vital signs and lab results, making hourly predictions about their likelihood of deterioration.
The research, published in the Canadian Medical Association Journal, compared over 13,000 admissions to St. Michael’s general internal medicine ward with thousands of admissions to other subspecialty units. The results showed a remarkable 26% decrease in unexpected deaths among patients in the unit using Chartwatch.
The Birth of Chartwatch
The Unity Health AI team began developing Chartwatch in 2017. They focused on creating a tool that could predict deaths or serious illness, areas where machine learning could make a significant impact. The technology underwent several years of development and testing before its deployment in October 2020. Chartwatch’s primary goal was to enhance patient care by providing early warnings of potential health deterioration.
How Chartwatch Works
Chartwatch monitored about 100 inputs from a patient’s medical record, including vital signs and lab test results. The AI system worked in the background alongside clinical teams, analyzing changes in medical records every hour. It made dynamic predictions about the likelihood of a patient’s condition worsening, potentially requiring intensive care, or even facing the risk of death. This continuous monitoring allowed for timely interventions by medical staff.
The Study’s Scope and Methodology
The research team examined over 13,000 admissions to St. Michael’s general internal medicine ward, an 84-bed unit caring for complex patients. They compared the impact of Chartwatch in this unit to thousands of admissions in other subspecialty units without the AI tool. The study period lasted for a year and a half, providing a substantial dataset for analysis. This comparative approach allowed researchers to isolate the effects of Chartwatch on patient outcomes.
Striking Results: A 26% Drop in Unexpected Deaths
The study’s findings were remarkable, showing a 26% reduction in unexpected deaths among hospitalized patients in the unit using Chartwatch. This significant decrease was not observed in other units of the hospital that did not use the AI tool. The research team, led by Dr. Amol Verma, considered this disparity a promising sign of Chartwatch’s effectiveness. The results suggested that the AI system was indeed saving lives.
Enhancing, Not Replacing, Human Care
Chartwatch was designed to complement, not replace, the judgment of healthcare professionals. The AI tool provided early warnings, allowing doctors and nurses to intervene sooner in critical situations. In some cases, these interventions involved escalating treatment to save lives. In others, it allowed for timely palliative care when patients couldn’t be rescued. The system worked alongside clinical teams, enhancing their ability to provide optimal care.
Addressing Healthcare Challenges
The implementation of Chartwatch came at a crucial time when Canada’s healthcare system faced significant challenges, including staff shortages. Dr. Verma emphasized that AI tools like Chartwatch had immense potential to supplement traditional bedside care. The technology helped combat these shortages by providing continuous monitoring and early warnings. This support allowed healthcare professionals to focus their attention where it was most needed.
A Pioneer in AI Healthcare Implementation
Chartwatch stood out as one of the first AI technologies implemented in a clinical setting in Canada. Unlike many AI healthcare initiatives still in the research phase, Chartwatch was actively helping care for patients daily at St. Michael’s Hospital. Dr. Verma highlighted the significance of this real-world application, noting that few AI technologies had reached this level of practical implementation in Canadian hospitals.
Expanding Chartwatch’s Reach
The Unity Health team expressed hope for wider implementation of Chartwatch in the future. They planned to expand its use within their Toronto-based hospital network and beyond. Dr. Muhammad Mamdani, vice president of data science at Unity Health Toronto, highlighted the role of GEMINI, Canada’s largest hospital data-sharing network for research and analytics. This network provided opportunities to test Chartwatch and other AI tools in various clinical settings across multiple hospitals.
Collaborative Efforts in AI Healthcare
The development and implementation of Chartwatch exemplified the power of collaboration in healthcare innovation. The project involved input from various healthcare professionals, data scientists, and researchers. Over 30 hospitals across Ontario were working together through the GEMINI network, sharing data and insights. This collaborative approach set the groundwork for deploying AI tools beyond individual institutions, potentially revolutionizing patient care on a broader scale.
The Future of AI in Patient Care
The success of Chartwatch opened up new possibilities for AI integration in healthcare. The technology demonstrated its ability to work alongside human healthcare providers, enhancing their capabilities rather than replacing them. As AI tools continued to evolve, they held the potential to address various challenges in healthcare, from early disease detection to resource allocation. The Chartwatch study provided a valuable blueprint for future AI implementations in clinical settings.
17 Epic “I’ll Do It Myself” Moments That Changed History
17 Epic “I’ll Do It Myself” Moments That Changed History
13 Household Problems Solved with WD-40
13 Household Problems Solved with WD-40
How Facebook’s News Blackout Reshaped Canadian Info-Sharing
