A gender-stratified, multilevel latent class assessment of chronic disease risk behaviours' association with Body Mass Index among youth in the COMPASS study

This paper sought to examine chronic disease risk behaviour latent classes and their association with Body Mass Index (BMI), assessing for gender differences. Participants were youth (n = 116,086; grades 9-12) enrolled in the COMPASS study (Ontario, Canada) during 2013, 2014, 2015. Multilevel latent class analysis was used to identify underlying, homogenous classes of youths' engagement in physical activity, smoking, binge drinking and marijuana use. Adjusted multilevel models regressed BMI on the latent classes controlling for ethnicity and grade. Three latent classes were identified: active experimenters (ACE), inactive clean youth (INC) and inactive substance users (INSU). This study found that gender differences are apparent in chronic disease risk behaviour latent classes and their association with BMI. INC males (OR = 0.85, 95% CI = 0.78, 0.93) were associated with a lower odds of overweight/obesity relative to active males who experimented with substance use. As for females, the class with the highest proportion of youth using substances were associated with higher odds (Females: OR = 1.2, 95% CI = 1.1, 1.4) of overweight/obesity relative to their active experimenting peers. As such, youth in latent classes with substance use are associated with higher BMI and weight status. Successful interventions may include school policies/programs that limit screen time use, as they were seen to have a positive effect on PA engagement and including social-influences approaches for substance use. Future research and interventions should be gender-specific as our results show that different latent classes are associated with obesity across genders.