This is the first study to employ PBT to comprehensively examine persistent problem behaviors among adolescents. Psychosocial factors had different effects on the initiation and persistence of problem drinking. Male sex, higher levels of depression at T2, more peers with whom they can talk to, and smoking at T2 were the common predictors for the initiation and persistence of problem drinking at T2. Top- school ranking and low impulsivity at T2 could prevent them from the initiation of problem drinking, whereas teachers’ and parental support at T2 may reduce the risk of persistent problem drinking, but such effects attenuated after adjusting for current smoking.
Our results indicated a significant overall effect of sex, with male sex significantly predicting the initiation and persistence of problem drinking one year later. Sex differences in neurobiological-related development, psychiatric comorbidities, or social-related factors may impact the risk of problem drinking distinctively among adolescent boys and girls [20]. For instance, literature suggested that men may be more sensitive to the effects of mesolimbic dopamine signals and may have stronger responses to alcohol’s rewarding effects than women do beck [21, 22]. Males may be more susceptible to the rewarding effects of alcohol, while female may sense its negative effects even at lower dosages [23]. Puberty begins earlier in women, the earlier maturation of women’s prefrontal cortex may help develop inhibitory control abilities to sensation seeking [24]. When responding to stress, female may demonstrate more internalizing or depressive symptoms, while male may have more externalizing or disinhibitions [25] and may not be able to control their impulses for alcohol use. Learning and memory-related neurological processes have also been shown to have sex variations. For example, the benefit of exercise in reducing adolescent alcohol use were greater in women than in men [26]. With regard to social-related aspects, research showed that adolescent boys were more prone to the pro-drinking norms than girls [27], and women may be judged more negatively than men when drinking in some cultural context [28]. Accordingly, the issue of gender difference should be taken into account when planning for prevention programs targeting on the initiation or persistence of alcohol use among adolescents.
Consistent with previous studies [29, 30], depression at T2 was associated with both initiating and persistent problem drinking at T2. Depression and accompanying rumination might influence adolescents to use alcohol as a ‘self-medication’ to mitigate these symptoms, and thus increased the risk for developing and persisting the problem drinking [30]. Moreover, problem drinking could disrupt the functioning of brain systems, leading to simultaneous increases in symptoms of depression and alcohol use [31]. Interventions aiming at the prevention and management of depression may have important preventive impacts on alcohol misuse. Additionally, early detections of depression might help prevent the development of the initiation or persistence of problem drinking.
The initiation of problem drinking was positively associated with the number of peers with whom they can talk to at T1; whereas the persistence of problem drinking was positively associated with the number of peers with whom they can talk to at T2. A previous study in Taiwan also found that effects of peers operated differentially by prior alcohol experience. For the alcohol naïve youngsters, having peers against alcohol drinking could reduce subsequent drinking occasions. For the alcohol experienced youngsters, drinking peers and attending classes of higher drinking norms could increase future drinking occasions [32]. Future research should further clarify the role of peer support and influence in adolescent problem drinking initiation and continuation.
Our findings that smoking status was associated with the initiation and persistence of problem drinking provide further support to previous literature reporting the association between smoking and persistent problem drinking [33]. Adolescents who smoked in the consecutive year had the highest risk of problem drinking as compared to other risk factors. This could imply smoking at T2 predicts the high probability of starting and persisting alcohol use. In this study, the possible protective effects of parental supports to hinder persistent problem drinking in Model 3 became non-significant when controlling for current smoking at T2 might imply that when considering managements and preventions of alcohol use problems early in adolescents, teenage smokers are a high risk group that requires greater attention. Public health policy should emphasize early screening and interventions in tobacco and alcohol use in schools to prevent problematic use and reduce morbidity and mortality associated with alcohol concurrent use in this high- risk group.
The result that top- school ranking could protect adolescents from initiating problem drinking in this study could be related to various factors. Low academic achievers are more likely to consume alcohol, smoke cigarettes, and use illegal substances than higher achievers [34]. In Taiwan, low-ranking school students had higher rates of problem drinking and were more likely to be exposed to substance-using peers than their counterparts [32, 35]. The psychogenic theory proposes that students initiate substance use to cope with their anxiety over academic failure [36]. Further, low school-ranking could lead to low self-esteem, which may trigger problem drinking in students. In addition, we found that adolescents who have more parents with whom they can talk to or with perceived support from teachers at T2 could have greater opportunities to be protected against persistent problem drinking in the consecutive year. In a systemic review, four protective factors (parental monitoring, the quality of parent– child relationship parental support, and parental involvement) were identified as longitudinal predictors of both alcohol initiation and levels of later alcohol use/misuse [37]. Adolescents who have one parent or ideally two parents to talk to could involve several protective factors such as a close parent–child relationship. In a cohort study conducted by [38], the support from teacher was associated with a significantly lower risk of alcohol abuse. In our study, the support from teachers could reduce the risk of persistent problem drinking but did not prevent teens from the onset of problem drinking. This could be related to the Taiwanese campus culture that values academic achievement; thus, teachers usually focus on academic teaching and there is a lack of systematic prevention of drinking problems. Future research is needed to examine students’ perspectives on the mechanisms of initiating problem drinking. School-based interventions providing a less stressful school atmosphere, emphasized less on academic achievements but more on physical or mental health, as well as attractive educational activities or multi-component collaborations between parents and teachers to motivate students to learn may help promote a positive school environment for low academic achievers [39].
The univariate analysis in Table 1 shows that the group with persistent use of alcohol had higher impulsivity scores at T2 than the cessation group, and there was no difference at T1. This might suggest the effect of alcohol on impeding adolescents’ developmental improvement of impulsivity. Impulsivity might play a determinant and a consequence role of problem drinking. On the other hand, the alcohol initiation group had higher impulsivity scores at both T1 and T2 than the non-alcohol group. Research has consistently demonstrated an association between impulsivity and problem drinking [40]. The direction of the association, however, is debatable owing to the derogatory effects of alcohol on brain areas associated with behavioral control. Our longitudinal study could shed light upon a two-way association between impulsivity and drinking.
This study used a large community sample of adolescents from different areas of Taipei. A further strength is that a broad array of well-established self-report psychosocial measures to explore this less studied topic of risk factors for persistent problem drinking among adolescents in a school setting. This study had several limitations. First, the attrition at follow-up was high. Second, information was based solely on self-reports. We did not design a relevant questionnaire specifically for peer influence, eg. whether their friends drink alcohol or not, therefore, we cannot explore influences from friends or social media in the initiation or persistence of problem drinking. Future research may include more confirmations of peer relationships or psychiatric or alcohol use diagnoses to investigate more on possible mechanisms of psychopathology.
In conclusion, a better understanding of the difference between risk and protective factors on alcohol involvement may help inform prevention and intervention efforts. Male or depressed teenagers that smoke or attached to the peer environment of drinking may require more attention for timely interventions. Future studies may continue to focus on possible preventive strategies on the identification and management of the initiation or persistence of problem drinking among adolescents with higher risks according to the PBT.