It is known that physical activity levels (PA levels) decline during adolescence, but there is a lack of knowledge on possible predictors of changes in PA levels in this period of life. This study aimed to prospectively investigate the relationship between sociodemographic and behavioral factors (predictors), PA levels and changes in PA levels in older adolescents from Bosnia and Herzegovina. The sample comprised 872 participants (404 females) tested at baseline (16 years of age) and at follow-up (18 years of age). Predictors were sociodemographic characteristics (age, gender, socioeconomic status, urban/rural residence, paternal and maternal education level) and variables of substance misuse (consumption of cigarettes, alcohol and illicit drugs). The PA level, as measured by the Physical Activity Questionnaire for Adolescents (PAQ-A), was observed as a criterion. Boys had higher PAQ-A scores than girls at baseline and follow-up. Paternal education levels were correlated with PAQ-A scores at baseline (Spearman’s R: 0.18, 0.15 and 0.14, p < 0.05, for the total sample, females and males, respectively) and at follow-up (Spearman’s R: 0.12, p < 0.01 for the total sample). Logistic regression, which was used to calculate changes in PA levels between baseline and follow-up as a binomial criterion (PA decline vs. PA incline), evidenced a higher likelihood of PA incline in adolescents whose mothers were more educated (OR: 1.29, 95% CI: 1.05–1.60) and who live in urban communities (OR: 1.56, 95% CI: 1.16–2.10). The consumption of illicit drugs at baseline was evidenced as a factor contributing to the lower likelihood of PA incline (OR: 0.36, 95% CI: 0.14–0.92). The negative relationship between illicit drug consumption and PA decline could be a result of a large number of children who quit competitive sports in this period of life. In achieving appropriate PA-levels, special attention should be placed on children whose mothers are not highly educated, who live in rural communities, and who report the consumption of illicit drugs. The results highlighted the importance of studying correlates of PA levels and changes in PA levels during adolescence.

Physical activity, body mass, and dietary habits are known to be important determinants of overall health status, but there is an evident lack of studies that examine these issues specifically in preschool children. The aim of this study was to identify associations that may exist between adhering to the Mediterranean diet (MD), levels of physical activity (PA), and body composition indices in apparently healthy preschool children from southern Croatia. Participants were 5- to 6-year-old preschoolers from the Mediterranean part of the country (the Split-Dalmatia County; n = 260, 126 females). Adherence to the MD was observed by the Mediterranean Diet Quality Index (KIDMED), PA level was evaluated by the Preschool-age Children’s Physical Activity Questionnaire (Pre-PAQ), and responses were collected from the parents. The participants’ waist circumferences (in cm), waist-to-hip ratios, and body mass index (in kg/m2, and in a z-score calculated relative to the normative value for age and sex) were used as indicators of body composition. All children were of the same age and tested over a one-month period of the same year as a part of the regular examination undertaken before attending elementary school. With only 6% of the children having a low KIDMED score, adherence to the MD was high. MD adherence was higher in girls (Chi-square = 15.31, p < 0.01) and children who live on the coast of the Adriatic Sea (Chi-square = 18.51, p < 0.01). A mixed effects logistic regression (with kindergarten as random factor) identified sedentary activity to be negatively associated with MD adherence (OR per point: 0.65, 95% CI: 0.44–0.91). High adherence to the MD in the studied sample may be attributed to regulated feeding in kindergarten. Considering that most Croatian elementary schools do not provide food to their students, MD adherence should be investigated later in life and also in other parts of the country where the MD is culturally less prevalent.

The prevalence of illicit drug misuse, including cannabis, in Croatian touristic regions is alarming. This study aimed to identify the prevalence of cannabis consumption (CC), to identify associations between sociodemographic and sport factors and CC, and to evaluate the predictors of CC initiation in adolescents residing in touristic regions. This study enrolled 644 adolescents from two touristic regions in Croatia (Split-Dalmatia and Dubrovnik-Neretva County) who were tested at baseline (16 years of age) and follow-up (18 years of age). The study instrument consisted of questions focused on predictors (age, gender (male, female), place of residence (urban or rural environment), familial social status, and different sport-related factors) and CC outcome. The results indicated a high prevalence of cannabis consumption (>30% of adolescents consumed cannabis), with a higher prevalence in males, and adolescents from rural communities. The prevalence of CC increased by 10% during the study period, with no significant differences between genders in trajectories of changes. Quitting sports was a risk factor for CC at baseline and follow-up. Better sport competitive results (odds ratio (OR): 0.80, 95% confidence interval (CI): 0.65–0.96) and familial social status (socioeconomic status: OR: 0.66, 95% CI: 0.39–0.91; maternal education: OR: 0.65, 95% CI: 0.48–0.88) were associated with lower likelihood of CC at baseline. The adolescents who reported better sport competitive results were at increased risk for initiation of CC during the course of the study (OR: 1.40, 95% CI: 1.03–2.01). The protective effects of sports at baseline were most likely related to various factors that prevent the consumption of substances in youth athletes (i.e., commitment to results, adult supervision); with the end of active participation, adolescent athletes are at high risk for CC initiation.