Adolescent mental distress has increased in recent decades. It is unclear whether this is associated with changes in health-risk behaviors.
Methods
We analyzed five waves (2002–2018) of the repeated cross-sectional Swiss Health Behavior in School-aged Children study (ages 11–15; N = 30,122). Latent Profile Analyses identified health-risk behavior profiles in each wave using five indicators (physical inactivity, poor sleep, unhealthy diet, smoking, alcohol use). Associations with sociodemographic variables and mental distress (internalizing and somatic symptoms, life satisfaction) were tested using multinomial and linear regressions, including profile*sex interactions.
Results
A consistent four-profile solution (low-risk; high alcohol use/slightly elevated substance use; moderate substance use; highest risk) fit best across waves. The low-risk profile was most prevalent and increased in later cohorts (2014–2018), while elevated-risk profiles declined. Older adolescents were more likely to belong to elevated-risk groups, which were associated with greater mental distress, especially in earlier cohorts. No significant sex interactions were found.
Conclusion
Health-risk behavior profiles remained stable, but their associations with mental distress weakened over time. Prevention efforts should adapt to evolving adolescent contexts.
adolescencehealth behavior in school-aged children studyhealth-risk behaviorlatent profile analysismental distresssubstance usesecular changeThe author(s) declared that financial support was received for this work and/or its publication. We gratefully acknowledge funding provided through the PRC’s Seed Grants program, supported by the UZH Transform funding line (awarded to CJ and LB).In addition, CJ and LB are both currently supported by a Swiss National Science Foundation Postdoc Mobility Fellowship (Grant numbers: 225243 for Laura Bechtiger and 230406 for Clarissa Janousch). The funders had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.Introduction
Recent years have seen the rise of a youth mental health crisis. Although this issue has gained particular attention during and in the aftermath of the COVID-19 pandemic [1, 2], mental distress in young people has already been increasing prior to the pandemic [3]. This mental health crisis is characterized by higher levels of internalizing symptoms, somatic complaints, and reduced life satisfaction in more recent cohorts of adolescents, especially among girls [3–7]. Globally, approximately one in five children and adolescents experience depression or elevated depressive symptoms, underscoring the substantial burden of youth mental distress [8]. Research on potential drivers of the youth mental health crisis examined manifold factors, ranging from the rise of social media to sexual violence exposure to macro-economic factors, and typically yielded mixed findings for many indicators (see 8,9 for reviews). Some important and malleable risk factors have received less research attention as explanatory factors for secular changes in mental distress, including health-risk behaviors. Yet, understanding such malleable risk factors is of special interest for public health, considering the potential to foster youth mental health at the population level [9].
Why would health-risk behaviors be among the potential explanatory variables? First, health-risk behaviors, such as physical inactivity, an unhealthy diet, poor sleep, and legal or illegal substance use, have repeatedly been associated with mental distress [10–14]. Second, health-risk behaviors are also known to differ between males and females [15, 16]: During adolescence, females tend to report more physical inactivity than males [17], whereas males tend to report more unhealthy eating habits and more substance use than females [18–20]. Third, both gender differences in the prevalence of mental distress and gender differences in health-risk behavior have been found to emerge in or by early adolescence [21–24]. This makes the early to mid-adolescent period an important developmental time to understand this association.
Some evidence suggests that, next to secular trends in mental distress, health-risk behaviors show secular change as well. The direction of this change depends on the specific health behaviors in question. A recent overview notes that alcohol and tobacco use have declined over the past two decades, while other risks remain persistent concerns, reflecting changing patterns of adolescent health [25]. Consistent with this broader picture, the direction of change varies by behavior: physical activity and fitness have declined [26, 27], and adolescents now appear to sleep for shorter durations than earlier cohorts [28, 29], whereas smoking and drinking have decreased in many settings [30–32].
Not only has the prevalence of mental distress and some health-risk behavior changed in recent decades, but there is additional evidence that the associations between mental distress and health-risk have also changed across that timespan. For instance, the association between substance use and mental distress has grown stronger in adolescents born in the early 2000s compared to adolescents born in the early 1990s, despite the decrease in the overall prevalence of substance use [33]. This stronger association between substance use and mental distress was particularly observed in females [33]. In contrast, another study has found that mental distress and behavioral risk are less strongly associated in more recent than earlier cohorts of adolescents [34].
Existing studies on secular changes in health-risk behavior and the associations with mental distress have focused on specific behaviors rather than examining changes in patterns. Because risk and protective behaviors frequently cluster, analyzing single behaviors or simply adjusting one for another may obscure shared determinants [35]. Examining health-risk behavior profiles instead of isolated behaviors offers a more comprehensive view of risk [36, 37], as the risk for mental distress increases with the accumulation and co-occurrence of risk behaviors, potentially in a non-linear manner [11]. Two broad strategies can be used to define behavioral risk patterns: theory-driven classification based on predefined thresholds, and empirical classification using mixture models [36, 38]. While theory-based approaches offer conceptual clarity, they require validated and developmentally invariant cut-offs across behaviors that differ substantially in scale, distribution, and social meaning. In adolescent populations, such thresholds are often unavailable or vary across age and historical context. Empirical mixture models, therefore, provide a pragmatic approach for summarizing co-occurring behaviors without imposing a priori subgroup boundaries. Therefore, secular changes in health behavior profiles over the past two decades need to be systematically examined on their own and in relation to changes in mental distress, given preliminary evidence that the association between mental distress and health-related risk could have changed in recent decades [7].
The Current Study
This study leverages publicly available data from the Swiss Health Behavior in School-Aged Children study (HBSC), a repeated cross-sectional survey conducted from 2002 to 2018. Using five time points, we examined secular changes in health-risk behavior profiles (i.e., physical inactivity, poor diet, poor sleep, and substance use) and potential changes in the association between health-risk behaviors and mental distress across this period among adolescents aged 11, 13, and 15 years. In contrast to prior research that has focused solely on individual health-risk behaviors, we attempted to identify co-occurring behavioral risk patterns using Latent Profile Analysis (LPA). We also investigated whether the associations between the resulting health-risk profiles and three distinct, yet complementary indicators of adolescent mental health and wellbeing, namely, internalizing symptoms, somatic symptoms, and life satisfaction, changed across time. Given that recent increases in adolescent mental health problems have been particularly pronounced among girls, we further examined gender differences in both profile prevalence and associations with mental distress.
MethodsRecruitment and Participants
The HBSC study is an international comparative survey conducted in collaboration with the World Health Organization (WHO). This repeated cross-sectional study collects data every 4 years to monitor health behaviors, outcomes, and contextual factors among adolescents across >40 countries [39, 40]. This study uses publicly available data from the Swiss HBSC study (2002–2018), which is part of the international HBSC research network [41, 42]. The Swiss HBSC surveys adhered to a standardized international protocol to ensure consistency and comparability across countries and time periods.
Participants were recruited through a two-stage cluster sampling design. In the first stage, schools were randomly selected to represent various geographical regions of Switzerland. In the second stage, entire classes within these schools were randomly chosen to participate. This approach ensured a nationally representative sample of the target population, adolescents aged 11, 13, and 15 years (i.e., 5th, 7th, and 9th grades). For the current study, a total of 30,122 participants (50.3% girls) were included. Data were collected using standardized, self-administered paper-and-pencil questionnaires completed during school hours under teacher supervision. Responses were anonymized.
The study adhered to strict ethical guidelines as outlined in the HBSC international protocol. Parental consent was obtained where required, and participants were informed about the voluntary nature of their participation. Ethical approval was granted by the relevant Swiss authorities for each survey cycle.
MeasuresRisky Health Behavior
Physical inactivity was measured using the question: Over the past 7 days, on how many days were you physically active for a total of at least 60 min per day? Responses ranged from 0 days (0) to 7 days [7]. Responses were reverse-coded so that higher scores indicate lower physical activity.
Poor sleep: Participants were asked about their sleep difficulties over the last 6 months with the question: “How often have you experienced difficulties falling asleep?” Response options ranged from about every day [1] to rarely or never [5]. Responses were reverse-coded so higher scores indicate poorer sleep.
Poor diet: Participants were asked four items about their dietary habits: “How many times a week do you usually eat fruits/eat vegetables/eat sweets/drink Coke or other soft drinks?” Items were rated on a seven-point scale, ranging from never [1] to more than once daily [7]. Responses for fruit and vegetable consumption were reverse-coded so that higher scores indicated lower consumption. A mean value was calculated across these four variables.
Smoking: From 2002 to 2014, smoking was assessed with the item “How often do you smoke tobacco at present?” and rated on a four-point scale from every day [1] to never [4]. In 2018, smoking was assessed with the item “In the past 30 days, on how many days (if any) have you smoked cigarettes?”, which was rated on a seven-point scale from never [1] to 30 days (or more) [7]. In 2014, when both items were included to assess smoking behavior, the 2018 item was used in those instances where the earlier smoking variable was not available (97/6634 cases). In order to harmonize the variables across the waves, we reversed-coded the item values from 2002 to 2014 so that higher values indicated more frequent smoking. We rescaled the 2018 variable to align with the 1-4 range while retaining the 7-point coding.
Alcohol use: From 2002 to 2014, alcohol use was assessed with three items “At present, how often do you drink anything alcoholic, such as beer/such as wine/such as spirits?” and rated on a five-point scale from every day [1] to never [5]. In 2018, alcohol use was assessed with the item “In the past 30 days, on how many days (if any) have you drunk alcohol?”, which was rated on a seven-point scale from never [1] to 30 days (or more) [7]. Again, to harmonize the variables across the waves, we used the maximum frequency reported for beer/wine/spirits for 2002 to 2014, respectively, before reverse-coding the mean scores so that higher values indicated more frequent alcohol use. Also, we rescaled the 2018 variable to align with the 1-5 range while retaining the 7-point coding.
Health Outcome Variables
Internalizing symptoms: Participants rated the frequency of feeling nervous, dizzy, low, and irritable on a five-point Likert scale, ranging from about every day [1] to rarely or never [5]. A reverse mean score was generated, with higher scores indicating more internalizing symptoms. Cronbach’s alphas are shown in the Supplementary Table S1.
Somatic symptoms: Participants rated the frequency of headaches, stomachaches, and backaches within the last 6 months on a five-point Likert scale, ranging from about every day [1] to rarely or never [5]. A reverse mean score was generated, with higher scores indicating more frequent somatic symptoms. Cronbach’s alphas are shown in the Supplementary Table S1.
Life satisfaction: Respondents indicated their perceived life satisfaction on a 10-point Likert scale, ranging from the worst possible life (0) to the best possible life (10) [43].
All of these measures are intended to capture population-level variation in mental distress and wellbeing rather than to assess clinical disorders, and have shown to have adequate validity and reliability in previous research [44–46].
Sociodemographic Variables
Sociodemographic characteristics included in the analysis were age (i.e., ages 11, 13, or 15), self-reported participant gender (1 = male; 2 = female), and family socioeconomic status (SES). As an indicator of family SES, a version of the Family Affluence Scale [47] was coded as the sum of four items that were included at each assessment (i.e., having a family car, target participant having their own bedroom, the number of computers in the house, and number of family holidays abroad in the past year).
Statistical Analysis
LPA was conducted using Mplus (Version 8.10; [48]), R version 4.4.2 [49] and the MplusAutomation package [50] to identify distinct subgroups of health behaviors separately for each time point but across all age groups. The analysis included five health-risk indicators: physical inactivity, sleep problems, unhealthy diet, smoking, and alcohol consumption. Variables were z-standardized to account for measurement variability. Models were estimated with the maximum likelihood with robust standard errors (MLR) estimator. Variance within profiles was set to zero to facilitate model estimation. Full information maximum likelihood (FIML [51])was used to account for missing data.
To determine the optimal number of latent classes, models ranging from one to six classes were evaluated, and their fit was compared using several statistical criteria, including the Akaike Information Criterion (AIC; [52]), the Bayesian Information Criterion (BIC; [53], sample size-adjusted BIC (aBIC; [53, 54]), entropy values, the Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT; [55]), the Vuong-Lo-Mendell-Rubin Likelihood Ratio Test (VLMR-LRT; [55]), and the Bootstrapped Likelihood Ratio Test (BLRT; [56–58]. Final model selection balanced parsimony with interpretability, guided by substantive relevance and statistical fit indices (see Supplementary Material for an overview of model indices; Supplementary Table S2).
To explore the association between distinct risky behavior profiles and sociodemographic characteristics and health outcomes, we saved the most likely profile membership from the LPA models and used it as an outcome/predictor in subsequent regression models. We estimated four sets of regression models: First, we examined the association between sociodemographic variables and health-risk behavior profiles using multinomial logistic regression. Second, we examined health risk behavior profiles as bivariate predictors (dummy-coded) of mental distress, using linear regression models. Third, we examined health-risk behavior profiles as predictors of wellbeing, adjusting for sociodemographic characteristics. Fourth, we also included the profile*sex interactions in the adjusted models. As a sensitivity analysis, we additionally estimated health-risk profiles across all waves combined and examined their associations with sociodemographic characteristics and mental distress.
Results
Table 1 presents descriptive statistics of the main study variables per survey wave (see also Supplementary Figures S1, S2 for age-specific and sex-specific trend graphs).
Descriptive statistics of main study variables. Health Behavior in School Aged Children study, Switzerland, 2002–2018.
Variables
2002 (N = 4,679)
2006 (N = 4,621)
2010 (N = 6,678)
2014 (N = 6,634)
2018 (N = 7,510)
n
M
SD
n
M
SD
n
M
SD
n
M
SD
n
M
SD
Health risk behaviors
Alcohol use
4,586
1.75
1.04
4,579
1.59
0.93
6,595
1.70
0.99
6,579
1.47
0.83
7,412
1.19
0.49
Smoking
4,661
1.34
0.86
4,603
1.21
0.67
6,636
1.25
0.73
6,579
1.15
0.56
7,463
1.09
0.43
Physical inactivity
4,633
3.11
2.01
4,517
3.25
1.96
6,615
3.03
1.81
6,525
2.84
1.86
7,438
2.84
1.87
Sleep problems
4,655
1.17
1.23
4,501
1.33
1.31
6,549
1.33
1.33
6,511
1.44
1.34
7,439
1.40
1.37
Unhealthy diet
4,673
3.33
1.01
4,595
3.13
1.00
6,667
3.18
1.00
6,614
3.11
0.98
7,506
3.03
0.98
Mental distress
Internalizing symptoms
4,663
2.01
0.78
4,512
2.10
0.83
6,563
2.04
0.83
6,526
2.12
0.84
7,551
2.09
0.86
Life satisfaction
4,646
7.82
1.71
4,464
7.78
1.80
6,537
7.70
1.84
6,448
7.73
1.76
7,405
7.69
1.85
Somatic symptoms
4,667
1.85
0.79
4,519
1.90
0.83
6,576
1.92
0.86
6,529
1.97
0.84
7,552
1.94
0.85
Latent Profile Analysis
The LPA analysis suggested that a four-profile solution best fit the data at all time points. The resulting health-risk behavior profiles are illustrated in Figure 1. When estimating profiles separately within each wave, the most frequent health-risk behavior profile with a prevalence of 74.3%–93.6% was always a low-risk profile, characterized by low values on each risk behavior. Second, we identified high alcohol use (2002–2010) and slightly elevated substance use (2014–2018) profiles, which showed elevated rates of substance use compared to the low-risk profile, though at varying degrees depending on the time point (prevalence: 3.4%–10.1%). Third, we identified a moderate substance use profile that was characterized by high levels of substance use (prevalence: 1.3%–8.5%). Fourth, we identified a highest risk profile, with elevated rates of substance use but also slightly elevated risk for other health-risk behaviors (prevalence: 1.7%–7.3%). Thus, profiles mainly differed in their substance use behavior (Figure 1; Supplementary Figures S1, S2).
Health-risk behavior profiles for each assessment. Health Behavior in School Aged Children study, Switzerland, 2002–2018.
Line graph panels compare health risk profiles in Switzerland overall and by survey year (2002, 2006, 2010, 2014, 2018) for five behaviors: alcohol consumption, smoking, physical inactivity, sleep problems, and unhealthy diet. Risk groups were low risk, slightly elevated substance use, moderate substance use, high alcohol use, and highest risk, and these were color-coded and plotted as means. Smoking and alcohol use distinguish highest risk groups, with their proportions labeled above charts.
A secular trend in the prevalence of health risk behaviors was observed. More adolescents were assigned to a low-risk group in 2014 and 2018 compared to the earlier years. This is reflected in a decrease in the prevalence of all three other groups in 2014 and 2018 compared to the previous years.
Sociodemographic Predictors of Latent Profile Membership
Table 2 shows sociodemographic predictors of profile membership across the years. Most consistently, older adolescents were more likely to be in the high alcohol use/slightly elevated substance use, moderate substance use, or highest risk profile than younger adolescents. Age also discriminated between these three health-risk profiles. Relative to 11-year-olds, for 13- and 15-year-olds the odds rose for highest risk vs. high alcohol use/slightly elevated substance use and for moderate substance use vs. high alcohol use/slightly elevated substance use. Sex differences are primarily seen in the years from 2002 to 2010. Compared with boys, girls had lower odds of belonging to any elevated-risk profile vs. low-risk. However, conditional on elevated risk, girls were more likely to be in moderate substance use vs. high alcohol use/slightly elevated substance use and more likely to be in highest risk vs. high alcohol use/slightly elevated substance use in 2002.
Associations of sociodemographic characteristics with health-risk behavior profiles. Health Behavior in School Aged Children study, Switzerland, 2002–2018.
Sociodemographic variables
High alcohol use/slightly elevated substance use vs. low-risk
Moderate substance use vs. low-risk
Highest risk vs. low-risk
Moderate substance use vs. high alcohol use/slightly elevated substance use
Highest risk vs. high alcohol use/slightly elevated substance use
Moderate substance use vs. highest risk
OR
95% CI
OR
95% CI
OR
95% CI
OR
95% CI
OR
95% CI
OR
95% CI
2002
Sex
0.40
0.32–0.49
0.85
0.68–1.06
0.65
0.51–0.83
2.14
1.62–2.82
1.63
1.22–2.18
1.31
0.97–1.77
SES
1.13
1.02–1.25
0.99
0.89–1.11
0.92
0.81–1.04
0.88
0.77–1.01
0.81
0.70–0.94
1.08
0.93–1.26
Age
2.93
2.60–3.31
2.52
2.22–2.86
4.99
4.16–5.98
0.86
0.73–1.01
1.70
1.39–2.09
0.50
0.41–0.62
2006
Sex
0.42
0.33–0.53
0.61
0.45–0.82
0.82
0.64–1.05
1.46
1.01–2.09
1.96
1.43–2.69
1.35
0.93–1.96
SES
1.15
1.03–1.29
0.89
0.77–1.04
0.90
0.79–1.02
0.78
0.65–0.93
0.78
0.67–0.91
1.00
0.83–1.20
Age
2.82
2.46–3.23
2.24
1.89–2.66
4.33
3.60–5.22
0.79
0.64–0.98
1.54
1.23–1.91
1.93
1.51–2.47
2010
Sex
0.56
0.47–0.66
0.60
0.50–0.73
0.64
0.51–0.82
1.09
0.86–1.38
1.16
0.88–1.52
0.94
0.70–1.25
SES
1.19
1.09–1.30
1.09
0.99–1.20
0.93
0.83–1.05
0.91
0.81–1.03
0.78
0.68–0.89
1.17
1.01–1.34
Age
2.68
2.42–2.97
2.82
2.51–3.18
5.38
4.42–6.55
1.05
0.91–1.23
2.01
1.62–2.48
0.52
0.42–0.65
2014
Sex
0.73
0.56–0.95
0.81
0.57–1.15
0.82
0.59–1.14
1.11
0.72–1.70
1.12
0.74–1.68
0.99
0.62–1.58
SES
1.22
1.06–1.41
0.99
0.83–1.19
0.93
0.78–1.10
0.81
0.65–1.01
0.76
0.61–0.94
1.07
0.84–1.36
Age
1.93
1.67–2.23
2.64
2.11–3.31
5.45
4.04–7.35
1.37
1.05–1.79
2.82
2.03–3.92
0.49
0.33–0.70
2018
Sex
0.98
0.75–1.27
0.86
0.57–1.29
0.63
0.43–0.92
0.88
0.54–1.41
0.64
0.41–1.01
1.36
0.79–2.35
SES
1.09
0.95–1.25
1.07
0.87–1.31
0.88
0.74–1.04
0.98
0.77–1.25
0.81
0.65–0.99
1.21
0.93–1.58
Age
2.93
2.46–3.48
6.27
4.22–9.33
5.65
4.02–7.93
2.14
1.40–3.29
1.93
1.32–2.81
1.11
0.66–1.86
OR, odds ratio; CI=confidence interval. Coefficients are bolded when the confidence interval does not cross zero.
Mental Health Outcomes of Health-Risk Behavior Profile Membership
Table 3 shows the results from bivariate and multivariate regression analyses of health-risk behavior profiles predicting outcomes. Bivariately, assignment to the high alcohol use/slightly elevated substance use profile was associated with more somatic and internalizing symptoms in 2010 compared to those assigned to the low-risk profile. Assignment to the moderate substance use profile was associated with increased somatic symptoms in 2002 and 2010, with more internalizing symptoms in 2002 and 2010, and with lower life satisfaction in 2010 compared to those assigned to the low-risk profile. Assignment to the highest risk profile was associated with increased somatic symptoms in 2002 and 2010, more internalizing symptoms in 2002 and 2006, and lower life satisfaction in 2002, 2006, and 2010 compared to those assigned to the low-risk profile. Moreover, highest risk profiles were significantly associated with more somatic symptoms, more internalizing symptoms, and lower life satisfaction. Except for the difference between the highest risk profile compared to the low-risk profile in somatic symptoms and life satisfaction in 2010, these findings held up when accounting for socioeconomic covariates. Between the three elevated risk profiles, differences in somatic symptoms, internalizing symptoms, and life satisfaction were small and inconsistent across waves.
Regression results of the health-risk behavior profiles-mental distress association. Health Behavior in School Aged Children study, Switzerland, 2002–2018.
Regression models
Bivariate
Multivariate
Somatic symptoms
Internalizing
Life satisfaction
Somatic symptoms
Internalizing
Life satisfaction
β
95% CI
β
95% CI
β
95% CI
β
95% CI
β
95% CI
β
95% CI
Reference category: low-risk
2002
High alcohol use/slightly elevated substance use
0.05***
−0.04, 0.14
0.07***
−0.02, 0.16
−0.04**
−0.14, 0.05
0.07***
−0.03, 0.16
0.09***
−0.01, 0.18
−0.06***
−0.15, 0.04
Moderate substance use
0.11***
0.01, 0.21
0.12***
0.02, 0.23
−0.10***
−0.20, 0.01
0.11***
0.00, 0.21
0.12***
0.02, 0.23
−0.09***
−0.20, 0.02
Highest risk
0.13***
0.02, 0.24
0.17***
0.06, 0.28
−0.15***
−0.26, −0.04
0.13***
0.02, 0.25
0.18***
0.06, 0.29
−0.14***
−0.25, −0.02
2006
High alcohol use/slightly elevated substance use
0.07***
−0.03, 0.18
0.06***
−0.05, 0.16
−0.08***
−0.18, 0.03
0.08***
−0.03, 0.19
0.06***
−0.05, 0.17
−0.07***
−0.18, 0.04
Moderate substance use
0.06***
−0.08, 0.21
0.08***
−0.06, 0.23
−0.12***
−0.27, 0.03
0.06***
−0.08, 0.21
0.08***
−0.01, 0.23
−0.11***
−0.25, 0.04
Highest risk
0.11***
−0.00, 0.23
0.13***
0.02, 0.25
−0.17***
−0.29, −0.05
0.10***
−0.02, 0.22
0.13***
0.00, 0.25
−0.14***
−0.27, −0.02
2010
High alcohol use/slightly elevated substance use
0.10***
0.02, 0.20
0.08***
0.00, 0.16
−0.05***
−0.13, 0.03
0.10***
0.02, 0.18
0.09***
0.01, 0.17
−0.04**
−0.12, 0.04
Moderate substance use
0.11***
0.01, 0.23
0.11***
0.02, 0.21
−0.12***
−0.22, −0.03
0.11***
0.01, 0.20
0.12***
0.03, 0.21
−0.11***
−0.20, −0.02
Highest risk
0.12***
0.02, 0.18
0.10***
−0.01, 0.22
−0.14***
−0.25, −0.02
0.11***
−0.01, 0.22
0.11***
−0.01, 0.22
−0.11***
−0.22, 0.01
2014
High alcohol use/slightly elevated substance use
0.07***
−0.06, 0.19
0.09***
−0.03, 0.22
−0.06***
−0.19, 0.07
0.06***
−0.06, 0.19
0.10***
−0.03, 0.22
−0.06***
−0.18, 0.07
Moderate substance use
0.07***
−0.11, 0.24
0.07***
−0.10, 0.24
−0.10***
−0.27, 0.07
0.06***
−0.11, 0.23
0.07***
−0.10, 0.24
−0.09***
−0.26, 0.08
Highest risk
0.11***
−0.05, 0.26
0.08***
−0.08, 0.24
−0.08***
−0.24, 0.08
0.09***
−0.07, 0.25
0.08***
−0.08, 0.24
−0.06***
−0.23, 0.10
2018
High alcohol use/slightly elevated substance use
0.08***
−0.05, 0.20
0.11***
−0.02, 0.23
−0.09***
−0.21, 0.04
0.06***
−0.06, 0.19
0.10***
−0.03, 0.22
−0.07***
−0.19, 0.06
Moderate substance use
0.04**
−0.16, 0.24
0.06***
−0.14, 0.26
−0.05***
−0.25, 0.15
0.02*
−0.17, 0.22
0.05***
−0.14, 0.25
−0.04**
−0.24, 0.16
Highest risk
0.10***
−0.08, 0.28
0.09***
−0.09, 0.26
−0.09***
−0.27, 0.09
0.09***
−0.09, 0.27
0.08***
−0.09, 0.26
−0.08***
−0.26, 0.10
Reference category: high alcohol use/slightly elevated substance use
2002
Moderate substance use
0.06***
−0.07, 0.19
0.06***
−0.06, 0.19
−0.06**
−0.19, 0.07
0.04*
−0.08, 0.17
0.05*
−0.08, 0.17
−0.04*
−0.17, 0.09
Highest risk
0.09***
−0.04, 0.23
0.12***
−0.02, 0.25
−0.11***
−0.25, 0.02
0.08***
−0.06, 0.21
0.11***
−0.03, 0.24
−0.09***
−0.23, 0.04
2006
Moderate substance use
0.01
−0.16, 0.19
0.04*
−0.13, 0.22
−0.06***
−0.24, 0.11
0.01
−0.16, 0.18
0.04*
−0.13, 0.21
−0.06***
−0.23, 0.11
Highest risk
0.05**
−0.10, 0.20
0.09***
−0.07, 0.24
−0.10***
−0.26, 0.05
0.03
−0.12, 0.18
0.07***
−0.08, 0.22
−0.08***
−0.23, 0.07
2010
Moderate substance use
0.03
−0.09, 0.14
0.04**
−0.07, 0.16
−0.08***
−0.20, 0.03
0.02
−0.09, 0.13
0.04**
−0.07, 0.16
−0.07***
−0.19, 0.04
Highest risk
0.05***
−0.08, 0.18
0.05**
−0.08, 0.18
−0.10***
−0.23, 0.03
0.04**
−0.09, 0.17
0.04**
−0.09, 0.17
−0.08***
−0.21, 0.05
2014
Moderate substance use
0.02
−0.19, 0.22
0.00
−0.21, 0.21
−0.05***
−0.27, 0.16
0.04**
−0.15, 0.24
0.00
−0.21, 0.21
−0.05**
−0.26, 0.16
Highest risk
0.05***
−0.15, 0.25
0.01
−0.19, 0.20
−0.03*
−0.23, 0.17
0.01
−0.20, 0.22
0.00
−0.20, 0.20
−0.02
−0.22, 0.18
2018
Moderate substance use
−0.01
−0.24, 0.22
−0.01
−0.24, 0.22
0.00
−0.23, 0.24
−0.02
−0.24, 0.21
−0.01
−0.24, 0.22
0.01
−0.23, 0.24
Highest risk
0.04**
−0.17, 0.26
0.01
−0.21, 0.22
−0.03**
−0.25, 0.18
0.05***
−0.17, 0.26
0.01
−0.20, 0.23
−0.03*
−0.24, 0.18
Reference category: highest risk
2002
Moderate substance use
−0.03
−0.18, 0.11
−0.06***
−0.20, 0.08
0.06***
−0.08, 0.21
−0.04
−0.18, 0.10
−0.07**
−0.21, 0.08
0.06**
−0.09, 0.20
2006
Moderate substance use
−0.03
−0.21, 0.15
−0.02
−0.21, 0.16
0.02
−0.16., 0.20
−0.02
−0.20, 0.16
−0.02
−0.20, 0.16
0.01
−0.17, 0.19
2010
Moderate substance use
−0.04*
−0.18, 0.10
−0.01
−0.15, 0.13
0.04*
−0.10, 0.18
−0.03
−0.16, 0.11
−0.01
−0.15, 0.13
0.02
−0.11, 0.16
2014
Moderate substance use
−0.03*
−0.26, 0.20
−0.00
−0.23, 0.23
−0.03
−0.26, 0.21
−0.03
−0.26, 0.20
−0.00
−0.23, 0.23
−0.03
−0.26, 0.20
2018
Moderate substance use
−0.05***
−0.31, 0.21
−0.02
−0.28, 0.25
0.03*
−0.23, 0.30
−0.06***
−0.32, 0.20
−0.02
−0.28, 0.24
0.03*
−0.23, 0.30
Multivariate models adjusted for sex, age, and family SES. Coefficients are bolded when the confidence interval does not cross zero.
*p < 0.05; **p < 0.01; ***p < 0.001.
Sex Differences in the Health-Risk Behavior and Mental Distress Association
Sex-stratified associations were small and imprecise, likely due to attenuated profile–distress links in later cohorts, reduced power in sex-specific subsamples, and wave-specific class composition (see Supplementary Tables S3, S4 for multivariate regression models run separately for males and females and the corresponding profile × sex interaction tests). In males, assignment to the highest risk profile compared to the low-risk profile was associated with more internalizing symptoms in 2002. Assignment to the high alcohol use/slightly elevated substance use profile compared to the low-risk profile was associated with more somatic symptoms in 2010. In females, assignment to the highest risk profile compared to the low-risk profile was associated with more internalizing symptoms in 2002. Assignment to the moderate substance use risk profile compared to the low-risk profile was associated with more internalizing symptoms and lower life satisfaction in 2010. However, confidence intervals for the profile-by-sex interactions all included zero, indicating that associations between health-risk behavior profiles and mental distress indicators did not differ significantly by sex.
Sensitivity Analysis
Results from the sensitivity analysis with the overall profiles are presented in the Supplementary Tables S5, S6, Supplementary Figure S3. Briefly, the overall profiles reflected the profiles identified for 2014 and 2018 better than the profiles identified for 2002, 2006, and 2010 (i.e., finding a slightly elevated substance use profile instead of a high alcohol use profile). Associations with sociodemographic characteristics changed in the earlier waves in that most associations with sex and SES are not significant, whereas the associations with age remained. Also, most of the associations with mental distress were non-significant, with the exception of the highest risk/low risk contrast in 2002.
Discussion
The present study examined secular changes in health-risk behavior profiles among Swiss adolescents from 2002 to 2018, using five waves of HBSC data, and how these profiles relate to indicators of mental distress and wellbeing (i.e., somatic symptoms, internalizing symptoms, and life satisfaction). Using latent profile analysis, we identified four distinct health-risk profiles for each time point: a large low-risk group (i.e., low-risk), and three smaller groups (i.e., high alcohol use/slightly elevated substance use; moderate substance use; highest risk), characterized by varying levels of substance use and, to a lesser extent, other risk behaviors. Similar patterns have been reported in other HBSC participant countries, with international HBSC reports documenting declining adolescent substance use alongside stable or increasing mental distress in recent cohorts [40]. Comparable trends have also been observed in Norway, where the population-based Young-HUNT study reported concurrent declines in health-risk behaviors and increases in mental health problems over time [59].
Secular Trends in Health-Risk Profiles
Across the study period, the prevalence of the low-risk profile increased, as the prevalence of all other health-risk behaviors decreased. This decrease was most pronounced for the profiles characterized by different levels of substance use only (i.e., high alcohol use/slightly elevated substance use; moderate substance use). This pattern is consistent with broader secular trends for lower age-specific prevalence of alcohol consumption and smoking among high-income countries like Switzerland [31, 60, 61], as well as with a previous study that documented an increase of low health-risk behaviors in youth [34]. It should be noted that, due to a change in measurement for alcohol and nicotine use in 2018 compared to the earlier assessments, the exact means of substance use in the profile estimation cannot be directly compared, as the variables required harmonization. Nevertheless, substance use behavior, and smoking in particular, differentiated the high-risk profiles, with minor variability in other behaviors such as sleep, diet, and exercise.
Associations With Mental Distress and Wellbeing
Associations between health-risk behavior profiles and the three indicators of mental distress were small and inconsistent. Although adverse health-risk behavior profiles were associated with more somatic symptoms, more internalizing symptoms, and a lower life satisfaction than those assigned to the low-risk group, this trend was primarily observed in the first waves (2002–2010). After 2014, these correlations decreased or disappeared, suggesting a potential cohort effect. One potential reason for this is that substance use, historically a marker of psychological vulnerability, may have lost its strong link with mental health as its prevalence declined and social meaning shifted [33]. As substance has become less common among adolescents and young adults, fitness-oriented lifestyles have gained importance [62]. Alternatively, the decoupling of risky behaviors and wellbeing in later cohorts might reflect changes in other risk and protective contexts, such as academic pressure, economic stressors, or improved prevention and support structures [43, 63]. Considering that the associations of health-risk profiles and mental distress did not remain significant in a sensitivity analysis where profiles were estimated based across all survey years and that there are overlapping confidence intervals across the waves, secular changes should be interpreted cautiously.
Variation in mental health outcomes did not vary significantly across the different elevated-risk profiles, suggesting that merely the presence of more than one risk behavior (potentially substance use), rather than their pattern, seems to account for the association with distress [11]. Controlling for sex, age, and SES made a minimal difference in the health-risk behavior and mental distress associations, reaffirming that the observed pattern is not the result of demographic confounding. Although the associations between behavior profiles and mental distress were not significant, health-risk behaviors remain public-health relevant, as previous studies show they predict long-term health outcomes [63]. Even small differences in adolescent health may widen over time, contributing to health inequalities in adulthood [9, 64]. Promoting healthy routines during this formative period can therefore yield lasting benefits for adult wellbeing.
Gender Interactions
Girls in higher-risk profiles reported slightly more mental health problems than boys, aligning with broader evidence of increasing gender disparities in adolescent internalizing symptoms [4, 23]. However, these differences were small and not statistically significant, as confidence intervals included zero, indicating no reliable sex-based variation in the link between health-risk behaviors and mental distress. Gendered social roles, less adaptive coping mechanisms, and increased susceptibility to psychosocial stressors are some possible explanations for the overall higher burden among girls [65]. Despite the fact that boys were marginally more likely to belong to high-risk groups, our results indicate that intervention and prevention should be attentive to the mental health needs of both genders.
Further Sociodemographic Differences
Although sociodemographic characteristics did not account for the (small) associations between health-risk behavior profiles and mental distress, associations with health-risk behaviors were observed. Across all waves, older adolescents were more likely to be assigned to one of the substance use profiles. This pattern was to be expected given the age-of-onset of substance use in Switzerland, and is consistent with developmental theories highlighting increased exposure and opportunity for risk behavior with age [64, 66, 67]. Socioeconomic differences were less consistent but suggested that lower family wealth increased the likelihood of being part of the highest risk rather than the low-risk profile. This underscores the importance of targeting prevention resources to vulnerable groups. These associations were overall quite stable, indicating that even as overall risk behaviors decline, age, sex, and SES continue to shape adolescents’ health-behavior clustering.
Contextual Interpretation
The limited variation in activity, diet, and sleep in the health-risk behavior profiles that emerged from the data analysis could reflect structural protection, such as mandatory physical education at school [68, 69] and the wide availability of healthy foods [70]. In contrast, differences in alcohol and nicotine use differentiate the emerging risk profiles, potentially reflecting their normalization and easy access in Switzerland (i.e., the legal purchasing age for beer and wine is 16 years, and only since October 2024, a nationwide minimum purchase age of 18 applies for tobacco products and e-cigarettes, [71]). Although informal enforcement has tightened in recent years, social norms still determine prevalence and perceived risk [72]. Future research would benefit from examining how shifting cultural and regulatory conditions influence the interrelation between adolescent behavior patterns and mental health. In particular, cross-national comparisons are needed to determine whether the Swiss patterns observed in this study reflect broader international trends or country-specific dynamics shaped by policy and cultural context.
Strengths and Limitations
This study has several strengths, including the use of large, nationally representative data spanning 16 years, harmonized measures of mental distress and health-risk behaviors across five repeated cross-sectional assessments, and a latent profile approach to capture health-risk behavior patterns across five common health-risk behaviors.
Some limitations must be noted. First, the cross-sectional nature of each assessment makes it impossible to disentangle the direction of the health-risk behavior-mental distress associations. These are likely bidirectional [73]. Second, several measurement comparability limitations need to be addressed. The measurement of alcohol use and smoking behavior changed in 2018 compared to the earlier assessments, which may have introduced bias due to reduced comparability. Since the aim of this study was not to compare absolute means but to understand relative differences in alcohol/smoking frequency between groups at the same time point, we consider this harmonization appropriate for our analytic goals. In addition, binge or heavy episodic drinking could not be examined, as this measure was not assessed consistently across survey waves, which may limit the identification of more problematic patterns of alcohol use. Future HBSC waves that include binge drinking indicators could help distinguish frequency of alcohol use from more harmful drinking patterns and clarify their links with adolescent mental distress. Furthermore, e-cigarette use was not assessed prior to 2022, and sleep was limited to difficulties falling asleep, without information on sleep maintenance or sleep duration, which may underestimate certain contemporary risk behaviors and sleep-related problems. Third, mental distress outcomes were assessed using brief self-report scales designed for large-scale surveillance rather than clinical diagnosis. In particular, the somatic symptom scale should be interpreted as an index of symptom burden rather than a unidimensional latent construct, which may partly explain modest internal consistency. Fourth, all data were self-reported, which can be especially problematic in studies of non-conforming behaviors, such as risky behaviors. Indeed, such underreporting has been documented and is driven by factors like social desirability, stigma, and fear of negative consequences [74, 75]. Lastly, latent profile approaches are exploratory by nature and depend on model selection choices [76]. LPA solutions should not be interpreted as discrete or naturally occurring subgroups but may reflect continuous variation in health-risk behavior. Profile definitions are also dependent on model specification and indicator selection and are therefore best understood as descriptive summaries. We decided to estimate wave-specific profiles to align with the repeated cross-sectional design of the HBSC study that allows to compare different snapshots across time. At the same time, theory-driven classifications face comparable limitations, including reliance on arbitrary cut-offs and limited flexibility in the context of secular change.
Conclusion
In summary, our study shows that adolescent health-risk behavior profiles remain a useful way to capture co-occurring behaviors, but their associations with mental distress are inconsistent and have weakened in more recent cohorts. This shift may reflect changing social meanings of substance use and structural protections in Switzerland. While prevention efforts should continue to target adolescent substance use because of its established health risks, these findings suggest that health-risk behaviors are less likely to be contributing significantly to the ongoing mental health crisis, highlighting the need to reassess which behavioral patterns are most relevant for adolescent mental health today.
Ethics Statement
The studies involving humans were approved by in each HBSC country, the HBSC study protocols were approved by the relevant institutional review boards. In Switzerland, this was the cantonal ethics committee of the canton of Vaud. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin.
Author Contributions
Conceptualization: LB and CJ. Methodology: LB, CJ, andHT. Formal analysis: LB, CJ, and HT. Data curation: LB, CJ, and HT. Writing original draft: LB, CJ, and HT. Writing – review and editing: LB, CJ, HT, DB, and LE. Visualization: LB, CJ, and HT. Supervision: LB and CJ. Project administration: LB and CJ. Funding acquisition: LB and CJ. All authors contributed to the article and approved the submitted version.
Acknowledgements
We gratefully acknowledge the UZH Population Research Center for supporting this research through a Seed Grant. We also thank Prof. Lilly Shanahan and Prof. Boris B. Quednow for their valuable feedback on earlier versions of this work. HBSC is an international study carried out in collaboration with the World Health Organization, Regional Office for Europe (WHO/EURO). The International Coordinator of the HBSC study is Dr. Joanna Inchley, University of Glasgow, Scotland. The Data Bank Manager is Professor Oddrun Samdal, at the University of Bergen in Norway. For details, see http://www.hbsc.org/. In Switzerland, HBSC is coordinated by Dr. Marina Delgrande Jordan at Addiction Switzerland and financed by the Federal Department of Health and the Swiss cantons.
Conflict of Interest
The authors declare that they do not have any conflicts of interest.
Generative AI Statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Supplementary Material
The Supplementary Material for this article can be found online at: https://www.ssph-journal.org/articles/10.3389/ijph.2026.1609345/full#supplementary-material
ReferencesOverhageLHailuRBuschABMehrotraAMichelsonKAHuskampHA. Trends in Acute Care Use for Mental Health Conditions Among Youth During the COVID-19 Pandemic. JAMA Psychiatry (2023) 80(9):924–32. 10.1001/jamapsychiatry.2023.219537436733SamjiHWuJLadakAVossenCStewartEDoveNReview: Mental Health Impacts of the COVID-19 Pandemic on Children and Youth – A Systematic Review. Child Adolesc Ment Health (2022) 27(2):173–89. 10.1111/camh.1250134455683CollishawS. Annual Research Review: Secular Trends in Child and Adolescent Mental Health. J Child Psychol Psychiatry (2015) 56(3):370–93. 10.1111/jcpp.1237225496340CampbellOLKBannDPatalayP. The Gender Gap in Adolescent Mental Health: A Cross-National Investigation of 566,829 Adolescents Across 73 Countries. SSM - Popul Health (2021) 13:100742. 10.1016/j.ssmph.2021.10074233748389GondekDVandecasteeleLSánchez-MiraNSteinmetzSMehmetiTVoorpostelM. The COVID-19 Pandemic and Wellbeing in Switzerland-Worse for Young People?Child Adolesc Psychiatry Ment Health (2024) 18(1):67. 10.1186/s13034-024-00760-w38844962McGorryPDMeiCDalalNAlvarez-JimenezMBlakemoreSJBrowneVThe Lancet Psychiatry Commission on Youth Mental Health. Lancet Psychiatry (2024) 11(9):731–74. 10.1016/S2215-0366(24)00163-939147461PatalayPGageSH. Changes in Millennial Adolescent Mental Health and Health-Related Behaviours over 10 Years: A Population Cohort Comparison Study. Int J Epidemiol (2019) 48(5):1650–64. 10.1093/ije/dyz00630815691LuBLinLSuX. Global Burden of Depression or Depressive Symptoms in Children and Adolescents: A Systematic Review and Meta-Analysis. J Affect Disord (2024) 354:553–62. 10.1016/j.jad.2024.03.07438490591FirthJSolmiMWoottonREVancampfortDSchuchFBHoareEA Meta‐Review of “Lifestyle Psychiatry”: The Role of Exercise, Smoking, Diet and Sleep in the Prevention and Treatment of Mental Disorders. World Psychiatry (2020) 19(3):360–80. 10.1002/wps.2077332931092JackaFNReavleyNJJormAFToumbourouJWLewisAJBerkM. Prevention of Common Mental Disorders: What Can We Learn from Those Who Have Gone Before and Where Do We Go Next?Aust N Z J Psychiatry (2013) 47(10):920–9. 10.1177/000486741349352323798717LoewenOKMaximovaKEkwaruJPFaughtELAsbridgeMOhinmaaALifestyle Behavior and Mental Health in Early Adolescence. Pediatrics (2019) 143(5):e20183307. 10.1542/peds.2018-330731004047O’NeilCENicklasTAZanovecMFulgoniVL. Diet Quality Is Positively Associated with 100% Fruit Juice Consumption in Children and Adults in the United States: NHANES 2003-2006. Nutr J (2011) 10(1):1–10. 10.1186/1475-2891-10-1721208446TaylorGMJMunafòMR. Does Smoking Cause Poor Mental Health?Lancet Psychiatry (2019) 6(1):2–3. 10.1016/S2215-0366(18)30459-030527762WijbengaLde WinterAFAlmansaJVolleberghWAMKorevaarELHofstraJMultiple Health Risk Behaviors and Mental Health from a Life Course Perspective: The Dutch TRAILS Study. Prev Med (2022) 154:106870. 10.1016/j.ypmed.2021.10687034780855MacArthurGJSmithMCMelottiRHeronJMacleodJHickmanMPatterns of Alcohol Use and Multiple Risk Behaviour by Gender During Early and Late Adolescence: The ALSPAC Cohort. J Public Health (2012) 34(Suppl. l_1):i20–30. 10.1093/pubmed/fds00622363027MahalikJRLevine ColeyRMcPherran LombardiCDoyle LynchAMarkowitzAJJaffeeSR. Changes in Health Risk Behaviors for Males and Females from Early Adolescence Through Early Adulthood: Health Psychology. Health Psychol (2013) 32(6):685–94. 10.1037/a003165823477574SterdtELierschSWalterU. Correlates of Physical Activity of Children and Adolescents: A Systematic Review of Reviews. Health Educ J (2014) 73(1):72–89. 10.1177/0017896912469578ChenPJacobsonKC. Developmental Trajectories of Substance Use from Early Adolescence to Young Adulthood: Gender and Racial/Ethnic Differences. J Adolesc Health (2012) 50(2):154–63. 10.1016/j.jadohealth.2011.05.01322265111LeeJAllenJ. Gender Differences in Healthy and Unhealthy Food Consumption and Its Relationship with Depression in Young Adulthood. Community Ment Health J (2021) 57(5):898–909. 10.1007/s10597-020-00672-x32602082WangYCaoPLiuFChenYXieJBaiBGender Differences in Unhealthy Lifestyle Behaviors Among Adults with Diabetes in the United States Between 1999 and 2018. Int J Environ Res Public Health (2022) 19(24):24. 10.3390/ijerph19241641236554290HankinBLAbramsonLYMoffittTESilvaPAMcGeeRAngellKE. Development of Depression from Preadolescence to Young Adulthood: Emerging Gender Differences in a 10-Year Longitudinal Study. J Abnorm Psychol (1998) 107(1):128–40. 10.1037/0021-843X.107.1.1289505045RollsBJFedoroffICGuthrieJF. Gender Differences in Eating Behavior and Body Weight Regulation. Health Psychol (1991) 10(2):133–42. 10.1037/0278-6133.10.2.1332055211SalkRHHydeJSAbramsonLY. Gender Differences in Depression in Representative National Samples: Meta-Analyses of Diagnoses and Symptoms. Psychol Bull (2017) 143(8):783–822. 10.1037/bul000010228447828WadeTJCairneyJPevalinDJ. Emergence of Gender Differences in Depression During Adolescence: National Panel Results from Three Countries. J Am Acad Child Adolesc Psychiatry (2002) 41(2):190–8. 10.1097/00004583-200202000-0001311837409MyttonOTDonaldsonLGoddingsALMathewsGWardJLGreavesFChanging Patterns of Health Risk in Adolescence: Implications for Health Policy. Lancet Public Health (2024) 9(8):e629–34. 10.1016/S2468-2667(24)00125-738996502DaleneKEAnderssenSAAndersenLBSteene-JohannessenJEkelundUHansenBHSecular and Longitudinal Physical Activity Changes in Population-Based Samples of Children and Adolescents. Scand J Med Sci Sports (2018) 28(1):161–71. 10.1111/sms.1287628299832FühnerTKlieglRArntzFKriemlerSGranacherU. An Update on Secular Trends in Physical Fitness of Children and Adolescents from 1972 to 2015: A Systematic Review. Sports Med (2021) 51(2):303–20. 10.1007/s40279-020-01373-x33159655KeyesKMMaslowskyJHamiltonASchulenbergJ. The Great Sleep Recession: Changes in Sleep Duration Among US Adolescents, 1991–2012. Pediatrics (2015) 135(3):460–8. 10.1542/peds.2014-270725687142MatriccianiLOldsTPetkovJ. In Search of Lost Sleep: Secular Trends in the Sleep Time of School-Aged Children and Adolescents. Sleep Med Rev (2012) 16(3):203–11. 10.1016/j.smrv.2011.03.00521612957BallJGruczaRLivingstonMter BogtTCurrieCde LoozeM. The Great Decline in Adolescent Risk Behaviours: Unitary Trend, Separate Trends, or Cascade?Soc Sci Med (2023) 317:115616. 10.1016/j.socscimed.2022.11561636563586MiechRKeyesKMO’MalleyPMJohnstonLD. The Great Decline in Adolescent Cigarette Smoking Since 2000: Consequences for Drug Use Among US Adolescents. Tob Control (2020) 29(6):638–43. 10.1136/tobaccocontrol-2019-05505231941823PapeHRossowIBrunborgGS. Adolescents Drink Less: How, Who and Why? A Review of the Recent Research Literature. Drug Alcohol Rev (2018) 37(S1):S98–114. 10.1111/dar.1269529573020GageSHPatalayP. Associations Between Adolescent Mental Health and Health-Related Behaviors in 2005 and 2015: A Population Cross-Cohort Study. J Adolesc Health (2021) 69(4):588–96. 10.1016/j.jadohealth.2021.03.00233867232ColeyRLLeerJLanteriL. Trends in Mental and Behavioral Health Risks in Adolescents: 1999-2021. Pediatrics (2025) 155:e2024068774. 10.1542/peds.2024-06877440096857PoortingaW. The Prevalence and Clustering of Four Major Lifestyle Risk Factors in an English Adult Population. Prev Med (2007) 44(2):124–8. 10.1016/j.ypmed.2006.10.00617157369McAloneyKGrahamHLawCPlattL. A Scoping Review of Statistical Approaches to the Analysis of Multiple Health-Related Behaviours. Prev Med (2013) 56(6):365–71. 10.1016/j.ypmed.2013.03.00223518213ProchaskaJJSpringBNiggCR. Multiple Health Behavior Change Research: An Introduction and Overview. Prev Med (2008) 46(3):181–8. 10.1016/j.ypmed.2008.02.00118319098GibbonsFXKingsburyJHGerrardM. Social-Psychological Theories and Adolescent Health Risk Behavior. Soc Personal Psychol Compass (2012) 6(2):170–83. 10.1111/j.1751-9004.2011.00412.xSamdalOKellyCCraigWHancockJWoldBAarøLEFour Decades of Advancing Research on Adolescent Health and Informing Health Policies: The Health Behaviour in School-Aged Children Study. Int J Public Health (2025) 70:1608136. 10.3389/ijph.2025.160813640201422InchleyJCStevensGWJMSamdalOCurrieDB. Enhancing Understanding of Adolescent Health and Well-Being: The Health Behaviour in School-Aged Children Study. J Adolesc Health (2020) 66(6):S3–5. 10.1016/j.jadohealth.2020.03.01432446607Delgrande JordanMSchmidhauserVBalsigerN. Étude Health Behaviour in School-Aged Children (HBSC) 2022 En Suisse. Rapport de Méthodes. Addict Suisse (2023). 10.58758/rech151MoorIWinterKBilzLBuckschJFinneEJohnNThe 2017/18 Health Behaviour in School-Aged Children (HBSC) Study – Methodology of the World Health Organization’s Child and Adolescent Health Study. Robert Koch-Institut (2020) 5:88–102. 10.25646/690435146275ArmitageJMCollishawSSellersR. Explaining Long-Term Trends in Adolescent Emotional Problems: What We Know from Population-Based Studies. Discov Soc Sci Health (2024) 4(1):14. 10.1007/s44155-024-00076-2LevinKACurrieC. Reliability and Validity of an Adapted Version of the Cantril Ladder for Use with Adolescent Samples. Soc Indic Res (2014) 119(2):1047–63. 10.1007/s11205-013-0507-4HauglandSWoldBStevensonJAaroeLEWoynarowskaB. Subjective Health Complaints in Adolescence: A Cross-National Comparison of Prevalence and Dimensionality. Eur J Public Health (2001) 11(1):4–10. 10.1093/eurpub/11.1.411276570GariepyGMcKinnonBSentenacMElgarFJ. Validity and Reliability of a Brief Symptom Checklist to Measure Psychological Health in School-Aged Children. Child Indic Res (2016) 9(2):471–84. 10.1007/s12187-015-9326-2CurrieCMolchoMBoyceWHolsteinBTorsheimTRichterM. Researching Health Inequalities in Adolescents: The Development of the Health Behaviour in School-Aged Children (HBSC) Family Affluence Scale. Soc Sci Med (2008) 66(6):1429–36. 10.1016/j.socscimed.2007.11.02418179852MuthénLKMuthénBO. Mplus User’s Guide. Eigth Edition. Los Angeles, CA: Muthén & Muthén (2008).R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing (2023). Available online at: https://www.R-project.org/ (Accessed May 2, 2023).HallquistMNWileyJF. Mplusautomation: An R Package for Facilitating Large-Scale Latent Variable Analyses in M plus. Struct Equ Model Multidiscip J (2018) 25(4):621–38. 10.1080/10705511.2017.140233430083048ChamHReshetnyakERosenfeldBBreitbartW. Full Information Maximum Likelihood Estimation for Latent Variable Interactions with Incomplete Indicators. Multivar Behav Res (2017) 52(1):12–30. 10.1080/00273171.2016.124560027834491AkaikeH. Information Theory and an Extension of the Maximum Likelihood Principle. In: ParzenETanabeKKitagawaG, editors. Selected Papers of Hirotugu Akaike. New York, NY: Springer (1998). p. 199–213. 10.1007/978-1-4612-1694-0_15SchwarzG. Estimating the Dimension of a Model. Ann Stat (1978) 6(2):461–4. 10.1214/aos/1176344136ScloveSL. Application of Model-Selection Criteria to Some Problems in Multivariate Analysis. Psychometrika (1987) 52(3):333–43. 10.1007/BF02294360LoY. Testing the Number of Components in a Normal Mixture. Biometrika (2001) 88(3):767–78. 10.1093/biomet/88.3.767McLachlanGJLeeSXRathnayakeSI. Finite Mixture Models. Annu Rev Stat Its Appl (2019) 6:355–78. 10.1146/annurev-statistics-031017-100325McLachlanGJPeelD. Finite Mixture Models. John Wiley & Sons (2000). p. 468.McCutcheonAL. Latent Class Analysis. Beverly Hills, CA: Sage (1987).64KrokstadMASundERangulVBaumanAOlssonCBjerkesetO. Secular Trends in Risk Factors for Adolescent Anxiety and Depression Symptoms: The Young-HUNT Studies 1995–2019, Norway. Eur Child Adolesc Psychiatry (2024) 33(11):3819–27. 10.1007/s00787-024-02373-238578474LoozeMdeRaaijmakersQBogtTterBendtsenPFarhatTFerreiraMDecreases in Adolescent Weekly Alcohol Use in Europe and North America: Evidence from 28 Countries from 2002 to 2010. Eur J Public Health (2015) 25(Suppl. l_2):69–72. 10.1093/eurpub/ckv03125805792DumontSMarques-VidalPFavrod-CouneTThelerJMGaspozJMBroersBAlcohol Policy Changes and 22-Year Trends in Individual Alcohol Consumption in a Swiss Adult Population: A 1993–2014 Cross-Sectional Population-Based Study. BMJ Open (2017) 7(3):e014828. 10.1136/bmjopen-2016-01482828298370European School Survey Project on Alcohol and Other Drugs (ESPAD). New ESPAD Survey Results: Teen Substance Use Down, but New Risks Emerging. Lisbon: European Union Drugs Agency (2025). Available online at: https://www.euda.europa.eu/publications/data-factsheets/espad-2024-key-findings_en (Accessed September 22, 2025).KeyesKMPlattJM. Annual Research Review: Sex, Gender, and Internalizing Conditions Among Adolescents in the 21st Century – Trends, Causes, Consequences. J Child Psychol Psychiatry (2023) 17:jcpp.13864–407. 10.1111/jcpp.1386437458091PattonGCSawyerSMSantelliJSRossDAAfifiRAllenNBOur Future: A Lancet Commission on Adolescent Health and Wellbeing. The Lancet (2016) 387(10036):2423–78. 10.1016/S0140-6736(16)00579-127174304De LoozeMECosmaAElgarFJSchrijversKInchleyJWalshSDExploring Mechanisms Behind the Increasing Gender Gap in Adolescent Psychological Symptoms, 2002–2022: The Role of National‐Level Gender Equality. J Child Psychol Psychiatry (2025) 66(5):737–51. 10.1111/jcpp.1408139644150QuednowBBSteinhoffABechtigerLRibeaudDEisnerMShanahanL. High Prevalence and Early Onsets: Legal and Illegal Substance Use in an Urban Cohort of Young Adults in Switzerland. Eur Addict Res (2021) 28(3):186–98. 10.1159/00052017834864731DegenhardtLStockingsEPattonGHallWDLynskeyM. The Increasing Global Health Priority of Substance Use in Young People. Lancet Psychiatry (2016) 3(3):251–64. 10.1016/S2215-0366(15)00508-826905480LamprechtMBürgiRGebertAStammHP. Sport Schweiz 2020: Kinder-Und Jugendbericht. Magglingen: Bundesamt für Sport BASPO (2021). Available online at: https://www.sportobs.ch/inhalte/Downloads/Kinder_und_Jugendbericht_2020_d.pdf (Accessed September 22, 2025).MeyerURothRZahnerLGerberMPuderJJHebestreitHContribution of Physical Education to Overall Physical Activity. Scand J Med Sci Sports (2013) 23(5):600–6. 10.1111/j.1600-0838.2011.01425.x22151355de MestralCStringhiniSMarques-VidalP. Barriers to Healthy Eating in Switzerland: A Nationwide Study. Clin Nutr (2016) 35(6):1490–8. 10.1016/j.clnu.2016.04.00427091772Bundesamt für Gesundheit (BAG). Bundesamt Für Gesundheit BAG. Tabakproduktegesetz (2024). Available online at: https://www.bag.admin.ch/de/tabakproduktegesetz (Accessed September 22, 2025).KeyesKMSchulenbergJEO’MalleyPMJohnstonLDBachmanJGLiGBirth Cohort Effects on Adolescent Alcohol Use: The Influence of Social Norms from 1976 to 2007. Arch Gen Psychiatry (2012) 69(12):1304–13. 10.1001/archgenpsychiatry.2012.78722868751HooshmandSWilloughbyTGoodM. Does the Direction of Effects in the Association Between Depressive Symptoms and Health-Risk Behaviors Differ by Behavior? A Longitudinal Study Across the High School Years. J Adolesc Health (2012) 50(2):140–7. 10.1016/j.jadohealth.2011.05.01622265109JanouschCEggenbergerLSteinhoffAJohnson-FergusonLBechtigerLLoherMWords Versus Strands: Reliability and Stability of Concordance Rates of Self-Reported and Hair-Analyzed Substance Use of Young Adults over Time. Eur Addict Res (2024) 31(1):60–74. 10.1159/00054171339561726PalamarJJLeAGuarinoHMateu-GelabertP. A Comparison of the Utility of Urine- and Hair Testing in Detecting Self-Reported Drug Use Among Young Adult Opioid Users. Drug Alcohol Depend (2019) 200:161–7. 10.1016/j.drugalcdep.2019.04.00831146203NylundKLAsparouhovTMuthénBO. Deciding on the Number of Classes in Latent Class Analysis and Growth Mixture Modeling: A Monte Carlo Simulation Study. Struct Equ Model Multidiscip J (2007) 14(4):535–69. 10.1080/10705510701575396
Edited by:Robert Wellman, UMass Chan Medical School, United States
Reviewed by:Michael Chaiton, University of Toronto, Canada
Reiner Hanewinkel, Institute for Therapy and Health Research (IFT-Nord), Germany