Could AI be used to assess fracture risk for older adults?
That’s the question at the heart of research being done by PhD student Edward Chu, who is exploring how machine learning could predict fractures in older adults with osteoporosis.
Currently, when health-care practitioners are trying to determine whether someone has osteoporosis – low bone mass and deterioration of bone tissue that is a factor in the likelihood of fractures – they’ll use a DXA scan, a medical imaging test that measures bone density.
Depending on the condition of your bones, future fracture risk is then predicted through questionnaires, which draw on age, sex and other clinical factors – but not the DXA scan results.
Chu wants to change that, so that DXA scans are taken into account when predicting fracture risk – making estimates of risk more specific to an individual.
“What’s really being overlooked is… they’re not really looking at your images,” Chu said. “They’re taking some of factors and giving you a generalized score for that.”
“We want to take those images, where you can actually see the structure of your bone, and where the most intensities are in terms of your bone material, and bring that into a clinical application where we actually do take into account the DXA scan.”
This approach could provide a far more comprehensive tool for predicting the fracture risk if you have osteoporosis – a disease that affects more than two million Canadians aged 40 and up.
Osteoporosis is a significant degenerative disease that increases the risk of fractures, and it’s more common in women than men – about 80 per cent of those living with diagnosed osteoporosis are women.
“It’s essential that we have some safeguard against falls in populations with osteoporosis,” Chu said. Falls are the leading cause of injury for older adults, and have long-term effects on their health and well-being.
“After you have a fall, it could be very debilitating,” he said. You might not be able to do daily tasks, and the chance of getting injured in a second fall becomes significantly higher.
“When you’re over the age of 65, you’re very concerned with mobility in general. You lose your independence when you lose your mobility,” he said. “We want to make sure we can have a preventative measure for that.”
But “it’s very hard for a person to gauge their bone strength… so we want to be able to do that for you,” Chu said.
Another advantage of DXA scans is that they’re low in power compared to typical X-rays. “It’s a lot easier on your body to get the scans, which is beneficial for vulnerable populations,” Chu said. “We want to decrease the amount of exposure that you get.”
It’s important to have an accurate assessment of fracture risk, as this can help guide recommendations from health-care practitioners. Based on the level of risk, they might suggest, for example, exercising to improve strength and balance, changing certain medicines to reduce dizziness or drowsiness, or treating other conditions that could increase fracture risk.
As he’s working to better assess future fractures, Chu said it’s a huge advantage to do this work with the interdisciplinary team at McMaster.
“My research involves the health science field, the engineering field, the computer science field, the public health field,” he said. “It’s very hard to juggle all those.”
But “I have the best resources in Canada at McMaster to reach out to each of these disciplines.”
Chu recently received the 2025 Labarge Mobility PhD Scholarship from the McMaster Institute for Research on Aging (MIRA) and the Larbarge Centre for Mobility in Aging, as well as a $3,000 Targeted Scholarship from the Municipal Retirees Organization Ontario (MROO).