High school students who have a large weekly allowance, friends who smoke and low levels of physical activity are more likely to use multiple substances over time. Conversely, being older, being Black and eating breakfast daily were factors associated with a smaller chance of transitioning to multiple use.
These conclusions were reached by a team of researchers at the University of Waterloo who used artificial intelligence to analyze a large, complex public health dataset—a novel way to approach public health analysis.
The study used machine learning instead of traditional statistical methods, allowing researchers to thoroughly assess multiple factors related to alcohol and other substance use patterns and transitions among Canadian high-school students over three time periods between 2016-19.
“Machine learning has advantages over traditional statistical methods,” said Helen Chen, a public health professor at…