Data Mining to Help Prevent Future Brain Injuries
U of T researchers used a novel data mining approach to uncover the most common medical problems affecting people in the time before they suffered a traumatic brain injury (TBI). The results can be used to better predict and prevent these life-altering incidents.
Lead author Tatyana Mollayeva, a post-doctoral research fellow in the Acquired Brain Injury Lab at the University of Toronto and the Toronto Rehabilitation Institute-UHN, worked with a team that found some interesting patterns in health status that distinguish people with TBI from those who entered the healthcare system for reasons other than brain injury.
“I was amazed at the numbers of people that had environmental exposure to gases and fumes, electrical currents, sharp objects, machinery and the cold in the five years before their traumatic brain injury,” Mollayeva says. “Our results bring attention to the link between exposure to occupational and other environmental hazards and subsequent critical injuries.”
Neurotoxicity from prescription drugs was another important link to brain injury, says Mollayeva. “Neurotoxic effects can mimic TBI symptoms, especially in the milder spectrum of severity, which adds complexity to the diagnosis of TBI.”
Professor Michael Escobar, a biostatistician at the Dalla Lana School of Public Health, and researchers from the Faculty of Medicine and University Health Network, examined the health records of almost 240,000 Ontario residents who suffered brain injuries, including those involved in falls, assaults and motor vehicle accidents. They sliced and diced the data, studying different time periods, including the five years before injury. Those studied were individually matched based by age, sex, income level, and place of residence.
Research shows two of the most common groups to suffer serious brain injury are older adults who fall and young people, who are at an overall greater risk of injury because of heightened risk-taking behaviours. The research bolstered a long-held hypothesis that the factors associated with injury are multiple, complex, and interdependent.
“TBI is a curious injury because it’s very heterogeneous,” notes Escobar. “It affects elderly women who fall, young hockey players, all kind of people who get into car crashes. So we need a lot of data, and very sophisticated means of analysing it to help identify the range of people at risk. In this study, we used an interesting variety of tools, from psychometrics to classical epidemiology to genomics to machine learning.”
The team’s results were published in Nature’s open-access journal Scientific Reportson April 3. The researchers hope their findings – and big data tools – contribute to the use of precision medicine to identify and intervene those at high risk for a traumatic brain injury.
This study was funded by the National Institutes of Health (PI. A. Colantonio).
