A new study suggests artificial intelligence can predict premature death among people who have inflammatory bowel disease, also known as Crohn’s and colitis.
Researchers used machine learning to analyze data from more than 9,000 people in Ontario who had IBD and died between 2010 and 2020.
The study, published today in the Canadian Medical Association Journal, found half of those deaths were before age 75, which is considered a premature death.
People were most likely to have died prematurely if they had been diagnosed with certain other chronic conditions — including arthritis, high blood pressure, kidney failure, cancer and mental health issues — before age 61.
Senior author Dr. Eric Benchimol, a pediatric gastroenterologist, says people don’t usually die of inflammatory bowel disease itself.
He says the research findings show the need to watch for other chronic conditions in people with IBD and treat those as early as possible.
“I think what’s important here is that we need to be aware that it can’t be a gastroenterologist working in isolation anymore,” said Benchimol, who practices at Sick Kids Hospital and is also a senior scientist at the Institute for Clinical Evaluative Sciences (ICES), which provided the patient data for the study.
Get weekly health news
Receive the latest medical news and health information delivered to you every Sunday.
“We really need to be aware of all of these other things that are happening to the people with IBD, and try to get them a system of co-ordinated care.”
Crohn’s and colitis inflame the lining of the gastrointestinal tract and interfere with the body’s ability to digest food and absorb nutrients, according to Crohn’s and Colitis Canada. Instead of normal bowel movements, patients suffer from “feelings of urgency, diarrhea, and bloody stool.”
It’s believed to be caused by a combination of genetics and environmental factors that affect the gut microbiome, Benchimol said, noting that smoking and early use of antibiotics are two factors thought to be associated with the disease.
IBD is quite common in Canada and the rate of cases is growing, he said. By 2035, more than one in 100 people in the country are expected to have the disease.
Benchimol said the specific chronic diseases identified alongside IBD in the study “unfortunately weren’t that much of a surprise.”
“I specialize in children with IBD, and we see some of these diseases all the time. Like we see arthritis. We see mood and anxiety disorders in kids with IBD. And teens with IBD are much more likely to have anxiety and depression than their peers,” he said.
What’s surprising — and worrisome — is the association of those conditions with early death, he said, noting more research is needed to determine how or why those deaths happen.
In addition to AI, the researchers used traditional statistical models to determine the factors linked to premature death among people with IBD and the results were consistent, boosting confidence in the accuracy of machine learning, Benchimol said.
Machine learning can process much more data and find relationships between different factors better than humans can, he said, noting that future studies using AI will track people who are living with IBD and include additional clinical variables such as how bad their inflammation is and what medications they are taking.
Benchimol hopes AI will be a way to figure out not only how to prevent premature death among people who have the disease, but ultimately how to prevent it altogether.
“We’re going to end up using machine learning to try to understand what environmental risk factors might predispose people to getting IBD in the first place in a way that maybe we can’t think of with our own brains,” he said.
© 2025 The Canadian Press