The first goal of this paper was to explore the existence of clusters of EBRBs in a large sample of school-age children across Europe. Five reliable EBRBs clusters showing good stability were identified for both boys and girls. Meaningful clusters were found while correlations between the EBRBs were low, showing that low correlations do not exclude co-occurrence of health-related behavioral indicators within certain groups. One of the major findings is that none of the clusters showed marked healthy or unhealthy trends for all the included EBRBs. This fact implies that health-related behaviors do not always discriminate in the same direction. Children with specific healthy habits are not necessarily predisposed to be involved in other specific healthy behaviors. Similar results were found in previously published studies [37, 38]. Cluster prevalence was not equal for each subgroup; the most prevalent were patterns characterized by low physical activity.
The results pertaining to the second goal – the characterization of the cluster solutions by parental education level and correlates – revealed that the cluster distribution was significantly different according to parental education level. It was clear that parental education was relevant for children’s EBRBs patterns. Children from higher educated parents were more likely to be allocated in the active pattern cluster, while low activity/sedentary pattern and sedentary sugared drink consumers were more prevalent among children from lower educated parents. It is noteworthy that low activity/sedentary pattern and sedentary sugared drink consumers pattern combined unhealthy levels in more than one of the assessed EBRB. These results suggest that children from lower educated parents are not only more likely to engage in less healthy behaviors [39, 40], but also more prevalent in clusters combining more unhealthy lifestyles. These results suggest the need to specifically address the relevance of physical activity and sedentary behaviors in obesity prevention strategies focusing on lower educated parents and their children.
The results show that children (both boys and girls) in the active cluster had below average screen time levels, and those children in the sedentary clusters (sedentary sugared drinks consumers for girls, sedentary pattern for boys) had also below average physical activity levels. This may suggest that there is some displacement between sedentary behavior and physical activity, and vice versa, although earlier studies suggest being careful with this displacement theory, as this displacement mechanism seems not to be universal across countries [41, 42].
The cluster solutions included two groups mainly characterized by their sleep duration scores (Long sleepers inactive pattern and short sleepers inactive pattern). Even if these groups showed an association with parental education, sleep did appear to differ in a stronger way according to country with more long sleepers in northern countries and more short sleepers in southern-east countries, in line with previous studies . This distribution seems to reflect an important country-specific cultural influence on sleep duration in children.
The cluster solutions showed large gender differences. These differences were reflected not only in the clustering itself (with a gender specific cluster, namely the low activity/sedentary pattern cluster) but also in the behavioral levels in the rest of the clusters. One of the main differences was found for sugared drinks intake, much higher for boys compared to girls, in agreement with previous literature [43, 44].
Additionally, cluster solutions were also characterized by BMI, being the short sleepers inactive pattern cluster the one with the highest proportion of overweight and obese boys and girls. In general, previous research showed that short sleep duration is associated with higher risk of childhood obesity [26, 45]. However, our findings should be interpreted with caution, taking into account the disproportionate country-specific representation in the short sleepers inactive cluster (i.e. Greek sample).
Finally, it has to be kept in mind that this study is subject to some limitations. First, this concerns a cross-sectional study providing evidence for associations but not causation. Further, data on dietary, physical activity and sedentary behaviors were based on self-reports, and thus possibly biased. However, the measures showed good test-retest reliability and construct validity . Additionally, for some behaviors both 24-h recall and frequency questions were included, showing similar results. When considering sedentary behaviors it is also important to note that some sedentary activities, like reading or studying were not included in the present study. Is therefore possible that questionnaires did not reflect the real, total time spent in sedentary behaviors.
Strengths of the present study include the large multinational sample from different regions across Europe and the standardized data collection protocol across the different countries. The use of cluster analysis for the assessment of EBRBs in relation with socio-demographic variables allowed us to reflect a more ecological view of the actual children’s behaviors and their socio demographic correlates. The described clusters in the present analysis showed a good stability and could therefore be seen as representative clusters for European children. It could be interesting for future research to examine the effectiveness of tailored obesity prevention strategies focusing on the most prevalent combinations of unhealthy EBRBs, with special attention to subgroups at higher risk, like children of lower educated parents.