Chronic diseases such as Type 2 diabetes take a massive toll on Canadians, both on our health and on our health-care system. Since obesity and inactivity are major risk factors for diabetes, the most common intervention so far has been for doctors to counsel their high-risk patients to lose weight and exercise more. That’s good, but is that the most effective approach as diabetes rates continue to soar?
Drawing on machine learning for her research, Laura Rosella, an epidemiologist and associate professor at the University of Toronto’s Dalla Lana School of Public Health, suggests that broader community-based actions could prevent more cases and save more money than targeting individual patients.
When Rosella took the risk-prediction algorithm that she and her team developed – the Diabetes Population Risk Tool – and applied it to Statistics Canada’s health information on the population, a clear picture jumped out at her. Beyond obesity, influential risk factors to predict who would get diabetes include lack of access to physical activities, social isolation, food insecurity, low socioeconomic status and chronic stress. The data suggested that making investments to address these factors could prevent disease.
Public health departments had suspected as much, but this was the first time they had the evidence to support it. Rosella’s algorithm would now enable them to clarify which populations to target for prevention efforts, and to calculate the health and economic benefits in their own municipalities from various investments, such as improving neighbourhood walkability. This opens up a whole new way of looking at health care, says Rosella, who holds a Canada Research Chair in Population Health Analytics. “It’s not a tweak. It’s going to actually change the way we think about disease and care.”
(Written by Marcia Kaye)