Clustering of Health-Related Behaviors and its Relationship with Individual and Context Factors in Portuguese Adolescents: Results from a Cross-Sectional Study

Background: Health behaviours are shaped early in life, laying the foundations of future health. Unhealthy behaviours tend to occur in complex specific patterns and are influenced by individual and context factors.

Objective: We aimed to identify and characterize patterns of health-related behaviours among Portuguese adolescents and correlate them with individual and context factors.

Methods: This study was based in Portuguese 2009/10 survey of Health Beahviour in School Aged Children Study, comprising 4036 adolescents. Individuals were partitioned into groups using two step cluster analysis based on 12 variables regarding diet, physical activity, screen use and substance use. The association between cluster membership and individual, family, school and peer factors was calculated based on crude and adjusted odds ratio using a multinominal regression.

Results: The median age was 13,6 and 54% were of the female gender. Overweight and obesity were highly prevalent (25%), as of sedentarism, with only 13% exercising daily. 78% lived with both parents and 59% had medium-low affluent families. We identified four distinct behavioural clusters: “Active screen users” (31,1%), “Substance users” (13,3%), “Healthy” (26,4%) and “Sedentary low fruit and vegetable eaters” (29,1%). Sociodemographics varied across clusters, showing a positive and significant association between male gender, poor family communication, violent behaviours, poor academic performance and school attachment and the “Substance users” and “Active screen users” clusters and between a low socioeconomic status and the “Sendentary low fruit and vegetable eaters” cluster. We found no association between nutritional status and cluster membership.

Conclusion: Cluster analysis identified several health compromising behaviour patterns. The understanding of these behavioural clusters and its relationship with individual and context factors is of extreme use to Public Health, allowing tailored health-promoting interventions. Further research is needed to understand how cluster membership evolves over time and its influence on nutritional status.