Toronto, Canada — Nearly half of all inflammatory bowel disease (IBD) patients who died in Canada passed away prematurely, according to a new study that also harnesses artificial intelligence to predict early mortality risks with unprecedented precision. Published in the Canadian Medical Association Journal (CMAJ), the study adds a new layer of urgency to efforts addressing gaps in chronic disease management.
Using machine learning models trained on nationwide health data, researchers identified patterns of premature death in IBD patients — those who died before reaching age 75 — and found that 47% of IBD-related deaths in the cohort were classified as premature. The findings, experts say, could reshape how healthcare providers assess risks and allocate resources to patients who need them most.
IBD, which primarily includes Crohn’s disease and ulcerative colitis, is a lifelong autoimmune condition marked by inflammation of the gastrointestinal tract. Affecting over 270,000 Canadians and more than 10 million people globally, IBD can result in severe complications such as bowel perforation, malnutrition, and colorectal cancer.
While survival rates have improved over recent decades, the new study highlights that these gains are unevenly distributed.
“The fact that nearly half of IBD deaths occur prematurely, even in a country with universal healthcare, signals deep systemic issues,” said Dr. Sanjay Murthy, a gastroenterologist at The Ottawa Hospital and senior author of the study. “AI can help us pinpoint which patients are most at risk of falling through the cracks.”
Machine learning tools allowed the research team to analyze a complex web of variables, from comorbidities such as cardiovascular disease to socioeconomic status, medication use, and health service utilization. The models successfully identified high-risk profiles that traditional statistical methods might overlook.
“What’s novel here is the ability of AI to detect nonlinear patterns — things like how combinations of risk factors can accelerate mortality risk,” Dr. Murthy explained. “It’s like having a second set of eyes capable of processing hundreds of data points simultaneously.”
AI models are increasingly being used across medicine, but their application in IBD-specific mortality prediction is a relatively new frontier. This study underscores their promise not just for forecasting, but for informing earlier, targeted interventions.
Beyond clinical variables, the study surfaces deeper inequities linked to where patients live and their socioeconomic standing.
Patients residing in rural areas and lower-income neighborhoods were more likely to experience premature death from IBD complications. Limited access to gastroenterology specialists, delays in receiving advanced therapies, and reduced availability of preventive care services all contributed to heightened risk.
“Geography shouldn’t dictate survival, but in many cases, it does,” said Dr. Jennifer Jones, a Toronto-based public health researcher. “We see patients in rural areas waiting months longer for colonoscopies or struggling to afford travel to larger centers for treatment.”
The disparities revealed by the study align with broader trends observed in Canadian healthcare, where rural populations often face reduced access to specialist care, despite universal health coverage.
Canada’s findings mirror international trends, illustrating that premature mortality among IBD patients is a global concern. A 2023 report from the British Society of Gastroenterology (BSG) found that approximately 40% of IBD-related deaths in the United Kingdom occur before age 75, with rural areas in Northern England and Scotland faring worse than urban centers like London.
IBD Premature Mortality Rates
- Canada: 47%
- United Kingdom: ~40%
- Australia: 35-45%
- United States: ~42% (higher in rural areas)
Similarly, data from Australia’s National Health and Medical Research Council report premature IBD mortality rates of 35-45%, with Indigenous populations facing the greatest risks. In the United States, studies have shown that IBD patients from rural counties or lower socioeconomic backgrounds are significantly more likely to experience early death, reflecting persistent healthcare access barriers.
“What’s striking is the consistency of this pattern, regardless of healthcare system,” said Dr. Anya Patel, a health equity expert at Stanford University. “Whether you have a private, mixed, or public system, structural inequities continue to define health outcomes.”
While AI-driven models offer a promising way to identify high-risk patients, experts caution that technology alone will not solve systemic issues.
“AI is a powerful diagnostic tool, but we still need boots-on-the-ground solutions — more gastroenterologists in rural areas, better patient follow-up systems, and culturally sensitive care for marginalized populations,” Dr. Patel noted.
In Canada, gastroenterologists are unevenly distributed, with rural and remote regions such as parts of Northern Ontario and the Prairies suffering acute specialist shortages. The Canadian Association of Gastroenterology has long advocated for stronger regional referral networks and telehealth expansion, but gaps remain.
The team behind the CMAJ study is now working to validate their AI models in broader and more diverse populations, including those outside of Canada. The ultimate aim is to embed AI-based risk scoring into routine IBD care, where it could assist clinicians in identifying patients for earlier specialist referral, tailored treatment, or enhanced monitoring.
“Imagine being able to sit down with a patient and tell them, with data, that they are in a high-risk group — and then have a clear plan of action to reduce that risk,” said Dr. Murthy. “That’s the potential we’re chasing.”
While AI’s potential to save lives is clear, the study’s findings also serve as a wake-up call for healthcare policymakers, particularly as chronic disease rates surge in aging populations worldwide.
“We must ensure that advancements in technology don’t widen the equity gap,” Dr. Jones added. “AI is only as effective as the health system it’s plugged into.”
For now, patients like Maria Lopez, a 51-year-old IBD patient living in rural Manitoba, can only hope that predictive models translate into real-world solutions. “I’ve learned to manage flare-ups, but I’ve also had years where specialist visits felt out of reach,” Lopez said. “If AI can help others get help sooner, that’s something to celebrate.”
As AI’s role in medicine expands, the challenge will be to ensure that those most in need — often the hardest to reach — are the first to benefit.
