Being lonely or using substances with friends? A cross-sectional study of Hungarian adolescents’ health risk behaviours

Table 1 presents the socio-demographic, health risk behaviour, social behavioural and social psychological characteristics for the secondary school student sample. Cronbach’s alpha values of the used social psychological variables in our sample were also included.

The sample contains data from 14 to 18-year-old students (mean = 16.21, SD = 1.13; female = 66.9 %, male = 33.0 %) representing three types of high school trainings (grammar = 51.2 %, modern = 42.9 %, technical = 5.9 %). 33.1 % of the sample smokes cigarette and 67.3 % drinks alcohol. These rates are in parallel with results of previous studies from the Hungarian high school student population. According to the ESPAD study from 2011, 37 % of Hungarian adolescents smoked and 61 % drank alcohol during the past 30 days [31]. Circa half (45.4 %) of students considered that their best friend smokes. This ratio is 71.3 % in the case of best friend’s alcohol use. Majority of adolescents considered that only minority or none of their friends smoked or drank alcohol. The response rates were high in the case of questions related to smoking and alcohol use (93.8 %–98.5 %).

Beyond descriptive statistics, binary logistic regression analyses were applied to study associations between the variables. Interrelationships between social behavioural and social psychological factors and health behaviours are presented in Table 2. Statistically significant relationships were found between both examined health risk behaviours (smoking, drinking and binge drinking) and social variables. The rates of smokers, drinkers and binge drinkers were significantly higher among those whose best friend smoked or drank alcohol, as compared to adolescents with non-smoking or non-drinking best friends. Peer group’s smoking is in a gradient-like relationship with all of the examined health risk behaviours. Higher rate of smoking friends is significantly related with higher odds of smoking, drinking and binge drinking. Gradient-like relationship was also found between friends’ drinking and adolescents’ alcohol use. If the majority or all of friends drink alcohol, the odds of smoking (p < 0.05; OR = 1.84; 95%CI = 1.14–2.97) and binge drinking (p < 0.001; OR = 2.87; 95%CI = 1.75–4.70) was also higher compared to other adolescents. According to these results there is a strong and positive association between smoking and drinking of friends (including the best friend and peer group) and the students’ health risk behaviours.

Among the social psychological variables, shyness is significantly related to all examined health risk behaviours. Higher level of shyness is correlated with lower odds ratio of smoking (p < 0.05; OR = 0.97; 95%CI = 0.95-0.99), drinking (p < 0.05; OR = 0.97; 95%CI = 0.95–0.99) and binge drinking (p < 0.05; OR = 0.97; 95%CI = 0.95-0.99). There is also a significant relationship between loneliness and both smoking (p < 0.01; OR = 0.95; 95%CI = 0.92–0.98) and drinking (p < 0.01; OR = 0.97; 95%CI = 0.93-0.98). Both smoking (p < 0.05; OR = 1.24; 95%CI = 1.04-1.48) and drinking (p < 0.001; OR = 1.45; 95%CI = 1.21-1.74) correlate positively with age in our sample. Gender is related only to drinking; boys have a higher odds of drinking (p < 0.01; OR = 1.99; 95%CI = 1.29-3.07) than girls.

The sample was divided into two groups by gender, and binary logistic regression analyses were also conducted separately. Table 3 presents the regression estimates among boys. Comparing with the results of the whole sample, several differences were found. Social behavioural factors were still better predictors than social psychological variables, but the correlations between social behavioural variables and health risk behaviours are weaker. We have not found significant relationship between the best friend’s smoking and binge drinking, and between the best friend’s alcohol use and smoking. In addition, friends’ alcohol use does not correlate with smoking and drinking in this subgroup.

Among the social psychological variables, competitiveness has been found to be a risk factor of drinking. In addition, shyness does not correlate with smoking and there is a significant negative relationship between loneliness and binge drinking (p < 0.01; OR = 0.91; 95%CI = 0.86-0.96). Age also correlates with drinking (p < 0.01; OR = 1.49; 95%CI = 1.02-2.16) and binge drinking (p < 0.01; OR = 1.67; 95%CI = 1.19-2.36).

Odds ratios of logistic regression among girls are presented in Table 4. All of the examined social behavioural variables correlate significantly with all variables of health risk behaviours (that is, smoking, drinking and binge drinking) of the adolescent girls. On the other hand, less social psychological variables are related to health risk behaviours compared to boys. Significant relationship has been found only between loneliness and girls’ drinking (p < 0.05; OR = 0.96; 95%CI = 0.92-0.99), as well as shyness and prevalence of smoking (p < 0.05; OR = 0.97; 95%CI = 0.94-0.99). Age correlates only with drinking in this subgroup (p < 0.01; OR = 1.37; 95%CI = 1.11-1.68).

Our study focused on the correlations between social psychological and social behavioural variables and health risk behaviours in a sample of Hungarian adolescents. Among social behavioural variables, both smoking and drinking of the best friend and peer group increased the odds of health risk behaviours in both genders (the only exclusion is the peer group’s alcohol use among boys). Cross-correlations (alcohol use of adolescent – smoking of the best friend/peer group and vice versa) were also significant in both health risk behaviours among girls and alcohol use among boys. Binge drinking also correlated with social behavioural variables. Smoking and alcohol use of friends and the peer group increased the risk of binge drinking among both girls and boys. The same results were established in previous international and Hungarian studies [19‐23, 32]. The role of these social factors proved to be contextual. Thus, in school classes where the prevalence of risk behaviours is higher, the society’s norms support or accept smoking and/or alcohol drinking. In these school classes social variables had a greater effect [33]. In comparison with other European countries, ratios of both alcohol use and smoking of Hungarian high school students are very high [31]. According to these results the norm of several Hungarian high school classes accepts or supports these behaviours.

Some of the included social psychological variables were also related to both adolescents’ alcohol use and smoking. Shyness among girls and loneliness among boys decreased the odds of smoking. Thus, our results are not consistent with some previous international studies [8, 9]. High level of feeling lonely decreased the risk of alcohol use among both genders. In addition, shyness also correlated negatively with this health risk behaviour among boys. Accordingly, the results of our study only partially support previous findings which described loneliness as a risk factor of alcohol use among adolescents [34‐40]. But this result is in concordance with other studies that did not support this finding [15]. We should take into account that adolescence is a life period where peer norms and shared activities play an important role. While loneliness may play a negative role in adolescents’ psychosocial well being, it may prevent youth from common alcohol and cigarette use. Shyness may operate in a similar way [14]. On the contrary, adolescents’ competitive attitudes were linked with higher odds of alcohol use as well similar to previous studies on the link between competitiveness/hostility/Type A behaviour and health risk behaviours [4].

Overall, the results of our study suggest the following: (1) social behavioural variables, including the best friend’s and peer group’s behaviours may strongly determine alcohol and cigarette use of Hungarian adolescents; (2) among social psychological variables high level of competitiveness may raise, while feeling of loneliness and shyness decrease the odds of risk behaviours.

The negative correlation between loneliness and shyness, and health risk behaviours may sound unexpectedly; however, this may be closely connected to the peer group effect and the role of the social network in an adolescent’s life. Firstly, our results, in parallel with several previous findings, established that the peer group may strongly determine adolescents’ health behaviours [19‐23, 32]. In addition, based on the international statistics on high school students’ smoking and alcohol use, acceptance or support of these risk behaviours is significantly higher in Hungarian high school classes, compared to West European countries [31]. Consequently, there is a strong social pressure on Hungarian youth to smoke and drink. The strength of this social pressure on an individual student depends on the student’s role in the social network and the social hierarchy of the high school class [41‐44]. Due to an adolescent’s shyness he/she may become marginalised in the class [42‐45]. On the other hand, social pressure on this student is lighter, which may decrease the odds of health risk behaviours in the class where alcohol use and smoking is accepted or supported by peers [41]. This can be an explanation for the phenomenon that according to our results, health risk behaviours were negatively related to loneliness and shyness.

Several limitations of our study should be noted. Self-reported data was used on adolescents’ mental health and health behaviours, without any clinical diagnoses or objective source. In addition, our study is cross-sectional; therefore, we cannot provide cause-and-effect relationships. We think, for a better understanding of cause-and-effect relationships, more extensive studies should be conducted, using a longitudinal design. Among social psychological variables, the need to belong scale has a Cronbach’s alpha coefficient of 0.60 in our data. Thus, we think that this scale, which have been used with higher reliability in several international studies [16], needs further validation in Hungarian adolescent populations. The dichotomous nature of many of our variables makes it difficult to compare our results with studies using non-dichotomous variables. Thus, other variables should also be considered, such as frequency of smoking and drinking, socioeconomic status, lifestyle, social support or introversion-extroversion. Finally, our findings may have limited generalisability because of the study’s specific cultural context, sample size, and imbalance of the sample (e.g., grammar school students were overrepresented comparative of other types of high school student population of Hungary). On the other hand, we emphasize the relevance of population based studies from different countries, regions and cultures that may show different results in the pattern and strength of these correlations. All in all, these findings may increase our understanding of the role of social psychological and social behavioural factors in adolescents’ health risk behaviours. Finally, our results also suggest that there is a need for more research focusing on the social behavioural and social psychological context of adolescent health risk behaviours in population based studies. Particularly, using different non-western societies to map features of these associations.

With our study we tried to give a useful addition to the practice of high school prevention. According to our results, we suggest that teachers and actors of prevention policy should pay more attention to adolescents’ social psychological attributes and social (peer) norms of high school classes. These indicators correlate with the odds of health risk behaviours, especially alcohol use and smoking, which can have a negative effect on youth future physical health status [46]. In addition, our results also suggest that prevention cannot function effectively without actively involving the whole high school class, particularly where an adolescent’s peer group accepts or supports smoking and drinking. Thus, individual prevention cannot be effective without changing peer group norms into health protective ones.