Patterns of clustering of six health-compromising behaviours in Saudi adolescents

Abstract

Background: Clustering of multiple health-compromising behaviours is associated with an increased risk of various chronic diseases. There are few studies on patterns of clustering of multiple health-compromising behaviours in adolescents. Therefore, the aim of this study is to assess how six health-compromising behaviours, namely, low fruit consumption, high sweet consumption, less frequent tooth brushing, low physical activity, physical fighting and smoking, cluster among Saudi male adolescents.

Methods: A representative stratified cluster random sample of 1,335 Saudi Arabian male adolescents living in Riyadh city answered a questionnaire on health-related behaviours. Hierarchical Agglomerative Cluster Analysis (HACA) was used to identify cluster solutions of the six health-compromising behaviours.

Results: HACA suggested two broad and stable clusters for the six health-compromising behaviours. The first cluster included low fruit consumption, less frequent tooth brushing and low physical activity. The second cluster included high sweets consumption, smoking and physical fighting.

Conclusions: The six health-compromising behaviours clustered into two conceptually distinct clusters among Saudi Arabian male adolescents, one reflecting non-adherence to preventive behaviours and the second undertaking of risk behaviours. Clustering of health behaviours has important implications for health promotion.

Figure 1

Tree diagram of hierarchical agglomerative cluster analysis of the six health-related behaviours. The dendrogram provides a visual presentation of the distance at which clusters are combined. It is read from left to right and the vertical lines show joined clusters. The position of the vertical line on the scale indicates the distance at which clusters are joined. Variables with smaller distance have higher homogeneity and they are combined with a vertical line linking them in a cluster, while the variables with larger distance indicate the least homogenous clusters. The distances (agglomeration coefficients) displayed in the top of the plot are rescaled (by default) to fall into a range of 1 to 25.