Do behavioural risks cluster among college students in Chandigarh, India? Novel insights from a latent class analysis.

BACKGROUND: Youth is a critical phase in life, as behavioural risks in youth can have profound health impacts throughout the life course. Coexistence of behavioural risks is especially hazardous, and is important to address in public health interventions. Given the gaps in existing research on this issue, we aimed to determine the prevalence of multiple behavioural risks and identify their clustering among young adults in a prominent North Indian city. METHODS: We collected data from young adults aged 18-22 years in a representative survey conducted across six colleges, assessing injury risks, victimisation, suicide, substance use, diet and physical activity using a self-administered questionnaire. We determined the prevalences of individual risks. We then identified risk clustering using latent class analysis, with gender as a covariate. We selected the most appropriate latent class model based on fitted probabilities, likelihood ratio tests, entropy, Akaike information criterion and Bayesian information criterion. RESULTS: Of 752 participating students with median age 19 years, 64.4% identified as female. Latent class analysis identified four classes: multiple risks (8.8%), smoking and alcohol-related risks (5.4%), only dietary and physical activity risks (73.7%), and victimisation and injury risks (12.1%). Unhealthy diet and insufficient physical activity were the most common risks, and occurred uniformly across classes. Men were more likely than women to belong to higher-risk classes. CONCLUSION: Policymakers working for youth health must consider the interconnectedness of behavioural risks and their gender differentials, to simultaneously address multiple risk factors such as violence, unsafe sex and substance use.