0
0
A team of researchers, led by Dr. Amol Verma, a clinician-scientist of Indian origin, has developed a novel artificial intelligence (AI)-based system that could significantly reduce the risk of unexpected deaths among hospitalised patients. The system, called CHARTWatch, acts as an early warning mechanism, identifying patients at high risk of deteriorating health, thus allowing healthcare workers to take timely preventive measures.
Tackling Rapid Health Deterioration
Rapid deterioration in hospitalised patients is one of the leading causes of unplanned admissions to intensive care units (ICUs). CHARTWatch has been designed to alert healthcare teams about patients whose health is declining, thus preventing the need for ICU transfers and reducing mortality risks.
“AI tools are increasingly being used in medicine, and it’s essential that they are carefully evaluated to ensure their safety and efficacy,” said Dr. Verma, lead author of the study and clinician-scientist at St. Michael’s Hospital, Unity Health Toronto, Canada. “Our findings suggest that AI-based early warning systems are promising for reducing unexpected deaths in hospitals.”
CHARTWatch’s Impact on Patient Outcomes
The effectiveness of CHARTWatch was evaluated in a study of 13,649 patients aged 55-80 years admitted to general internal medicine (GIM) wards. Of these, 9,626 patients were in the pre-intervention group, while 4,023 received the benefits of CHARTWatch. In addition, 8,470 patients admitted to subspecialty units without the AI tool served as a control group.
CHARTWatch’s success was attributed to its real-time alerts to clinicians and regular updates for the healthcare team. It facilitated twice-daily emails to nursing staff, as well as daily emails to palliative care teams, ensuring consistent communication. Importantly, the AI system prompted care pathways for high-risk patients, encouraging close monitoring by nurses and more frequent reassessments by physicians.
Enhanced Communication and Care Pathways
The AI-driven system fostered improved collaboration between healthcare teams. Nurses and physicians were prompted to engage in more frequent and effective communication about the patients flagged as high-risk by CHARTWatch. This led to faster interventions and better care management, significantly contributing to reducing patient mortality.
“The AI system can support nurses and doctors in providing high-quality care,” Dr. Verma added, emphasising the potential of AI in enhancing patient outcomes through early detection and intervention.
Evaluating the Complex AI Deployment
Dr. Muhammad Mamdani, co-author and director at the University of Toronto, highlighted the importance of evaluating the real-world outcomes of deploying AI tools like CHARTWatch. “Understanding the real-world impacts of this promising technology is important,” said Mamdani. The study, published in the Canadian Medical Association Journal (CMAJ), showcases how integrating AI into healthcare can transform patient management, especially for those at high risk of health deterioration.
As AI technology continues to advance, tools like CHARTWatch could become standard in hospital settings, helping clinicians prevent complications, reduce unplanned ICU admissions, and ultimately save lives. This AI system exemplifies the future of personalised medicine, where healthcare workers are empowered to make informed decisions with the support of real-time data analytics.
