Abstract
Concerns have been raised about an increase in children’s mental health symptoms over the past 30 years, including after COVID-19 lockdowns. Yet, few studies have investigated variations over generations, while considering sex and socioeconomic status. We aimed to address this gap by comparing mental health symptoms (emotional distress, impulsivity/hyperactivity/inattention, disruptive behaviours) reported by classroom teachers of 11-year-olds in three population-based, prospective, representative cohorts in Quebec, Canada. Analyses included 1665 (83%) of the Quebec Longitudinal Study of Kindergarten Children, in 1993; 1305 (62%) of the Quebec Longitudinal Study of Child Development, 2009; and 3871 (100%) of the Quebec Survey of Child Development in Kindergarten, 2022; ~50% boys. Teacher-rated symptoms on the validated Social Behavior Questionnaire showed higher scores of emotional distress and impulsive/hyperactive/inattentive symptoms in 2022 than 2009, and higher in 2009 than 1993 (very small-to-small effect sizes: Cohen’s d 0.12 and 0.26 for emotional distress, 0.06 and 0.25 for impulsive/hyperactive/inattentive symptoms, respectively; P < 0.001). Disruptive behaviour symptoms scored lower in 2022 than 2009, though higher in 2009 than 1993, with very small effect sizes (Cohen’s d: -0.15 and 0.09, respectively). Boys presented more impulsive/hyperactive/inattentive and disruptive behaviour symptoms than girls; girls showed more emotional distress than boys. Children from economically disadvantaged households (lowest 20% of income distribution) presented higher symptoms rates than advantaged children. These findings provide novel and timely evidence about variations in children’s mental health symptom rates over three decades, underscoring the need for preventive interventions as early as elementary school, tailored differentially for boys and socioeconomically disadvantaged children.
Supplementary Information
The online version contains supplementary material available at 10.1007/s00787-025-02854-y.
Keywords: Child and adolescent mental health, COVID-19 pandemic, Epidemiology, Childhood depression, Inattention and hyperactivity, Temporal trends.
The children now love luxury; they have bad manners, contempt for authority; they show disrespect for elders and love chatter in place of exercise. Children are now tyrants, not the servants of their households. They contradict their parents, chatter before company, gobble up dainties at the table, cross their legs, and tyrannize their teachers.
― Kenneth John Freeman, with reference to Socrates’ era, c. 400 BC.
Introduction
Contemporary changes in behavioural and mental health problems among youth have long been cause for concern, with each generation expressing apprehension about young people’s ability to cope with the evolving demands of daily life. Nevertheless, with the development of standardized measures such as the Child Behavior Checklist [1], these concerns have found empirical support [2–11]. In particular, some cross-sectional studies have suggested increased mental health symptoms in children and teens, associated with and following the lockdowns and school closures related to the COVID-19 pandemic [6, 12–17]. In general, evidence suggests a possible increase in the prevalence of emotional and behavioural problems in Western developed countries [2–5], such as symptoms of anxiety and depression in adolescent girls [2, 4, 8, 9] and possibly symptoms of hyperactivity and inattention in boys [8, 9]. Data is unclear about disruptive behaviour problems, and questions remain over the specificity and scope of these changes [2, 4, 8, 9].
Importantly, recent estimates suggest that mental health disorders including anxiety, depression, and attention-deficit/hyperactivity disorder (ADHD) now affect approximately 11% of children and adolescents worldwide [5], representing the largest disease burden at ages 10–24 years [18]. Notably, ADHD assessment criteria and disorder definitions have broadened over time, confounding time trend analyses [19, 20]. And although boys are more likely to experience externalizing disorders such as ADHD and conduct disorders, and girls more commonly experience internalizing disorders such as anxiety and depression [2, 3], studies have shown mixed findings on whether sex differences for specific disorders have changed over time [2, 8]. Evidence from multiple timepoint comparisons over several decades using a harmonized set of cohorts and assessment methods is sorely needed but remains limited [4, 16].
Concerns have also been raised about associations between children’s mental health and socioeconomic inequalities [2, 4, 15]. Children from socioeconomically disadvantaged backgrounds tend to experience higher levels of internalizing and externalizing symptoms than their peers [4, 21], with a study reporting a temporal increase in this mental health gap [4]. Most of these studies were conducted prior to the COVID-19 pandemic [2, 8, 22] and it is unknown whether the gap is now greater or lesser [23], highlighting the need for updated research [24].
We aimed to compare mental health symptoms in children of similar age over the past 30 years, using similar cohorts with the results stratified by sex and household income. To do so, we selected 3 representative and prospective cohorts in Quebec, Canada, that have data on symptoms all assessed by an individual in the same capacity, i.e. classroom teachers, and using the same validated instrument, i.e. the Social Behavior Questionnaire [25]. We compared data at age 11 years, available in our cohorts in 1993, 2009, and 2022. Age 11 was chosen as data was relatively available in all 3 cohorts and represented a single timepoint. The goal was to better inform policymakers and enhance targeted preventive interventions. Coincidentally, using 2022 data also enabled us to compare pre- and post-COVID mental health symptom scores.
Methods
Study design and participants
We obtained and compared data from three population-based, longitudinal cohorts in Quebec, Canada. All original study protocols were approved by their respective institutional Research Ethics Boards, including the Sainte Justine Hospital Research Centre, the University of Montreal, and the Quebec Statistics Institute (Institut de la Statistique du Québec), as applicable. Written informed consent was obtained from participants, parents and teachers at each data collection point, as applicable. The protocol for the present study was approved by the Sainte Justine Hospital Research Center ethic committee.
1) Quebec Longitudinal Study of Kindergarten Children (QLSKC).
The QLSKC was initiated in the 1986–87 and 1987–88 school years, designed to be representative of all children attending kindergarten in French-language public schools of urban and rural Quebec [26]. In Quebec, all children age 5 years by September 30 of the school year were eligible to attend kindergarten, which was widely attended. The cohort included 2000 children as well as a similarly selected sample of 1017 “disruptive” children who scored ≥ 80th percentile on the Social Behavior Questionnaire. The subset of 1017 was not used in the present study (n = 2000). Assessments in the QLSKC were made yearly from ages 6–12 years and again at ages 15, 22, and 29 years.
2) Quebec Longitudinal Study of Child Development (QLSCD).
The QLSCD is a representative birth cohort of infants born at 24–42 weeks’ gestation to French or English-speaking mothers from October 1997 to July 1998, in all of Quebec except for Indigenous territories (n = 2120). Longitudinal assessments were done annually from age 5 months to 8 years, and biennially thereafter to age 25 years [27]. As no data collection was conducted at age 11 years in this cohort, we averaged the values obtained at ages 10 and 12 years.
3)Quebec Survey of Child Development in Kindergarten (QSCDK).
The QSCDK was initiated in 2016–2017 in all English- and French-language kindergarten classes in public and private schools, except for those in Indigenous territories and special needs schools (n = 83,335, representing 94.8% of the eligible population) [28]. Of these, 8,760 children were randomly selected by the Quebec Statistics Institute between July to September 2021 to participate in a longitudinal study, and 4,524 provided consent. Later, from March to June 2022, the Quebec Statistic Institute invited 8,800 children (Grade 5, age 11) who participated to the QSCDK to join a new data collection wave, of whom 8,217 consented to participate. Finally, the Quebec Statistics Institute linked the data of the 3,871 children who participated in the three phases.
Measures
Mental health symptoms
In all cohorts, classroom teachers assessed their students’ mental health symptoms during the spring of the school year of data collection. We used data from the QLSKC and QLSCD at age 11 years; in the QLSCD, as only ages 10 and 12 years were available, we used the arithmetic mean of the two. The validated Social Behavior Questionnaire (SBQ) [25] included 18 items in 3 dimensions: (1) emotional distress; (2) symptoms of impulsivity/hyperactivity/inattention; and (3) disruptive behaviours (Supplementary Table 1). Questions addressed issues in the past 6 months on a 3-point Likert scale (“never or not true”, “sometimes or somewhat true”, and “often or very true”). Each dimension score was calculated as the average of the items, with higher scores indicating greater mental health symptoms. We examined internal consistency within each cohort, using Cronbach’s alpha (Supplementary Table 2). We also examined the internal structure of the SBQ using first-order confirmatory factor analysis (Supplementary Table 3).
Covariates
Within each cohort, data was collected on sociodemographic covariates both at baseline and thereafter (annually/biennially or at age 11, depending on the cohort). We considered 12 covariates in our analyses: child sex (boy or girl, as indicated by parent at baseline); maternal and paternal ages at childbirth; and maternal and paternal birthplace (Canada, elsewhere); as well as updated (at age 11 years) maternal and paternal education (high school or higher, no high school diploma), maternal and paternal working status (working, unemployed), family status (intact, non-intact), number of children in the household, and household income in the past 12 months (QLSKC and QLSCD : < CA$20,000 vs. $20,000–79,999 vs. ≥ $80,000; QSCDK: < CA$60,000 vs. $60,000-159,999 vs. ≥ $160,000). We categorized household income as disadvantaged (lowest 20% of the distribution), median (median 60%), or advantaged (highest 20%).
Statistical analysis
We used Student’s t-tests for continuous variables and χ2 tests for categorical variables. Children with more than 30% missing data on mental health symptoms were excluded from analyses. To address both the issue of attrition and selective non-response on specific items in each cohort, we first compared baseline sociodemographic characteristics of children included in the analyses to those excluded. Weights included covariates for which the distribution was different between included participants in the final sample and those lost to attrition/missing. We then used inverse probability weighting, with multiple imputation by chained equations (m = 50), to maintain representativeness of the initial cohort. For the QSCDK (2022 cohort), the Quebec Statistics Institute provided estimated weights to ensure representativeness.
We compared SBQ mean scores across cohorts using Students’ t-tests as well as Cohen’s d effect sizes (very small, < 0.20; small, 0.20–0.50; medium, 0.50–0.80; large, 0.80–1.20; very large, 1.20–2.0; huge, >2.0 [29]). Finally, we used weighted linear models to quantify associations between mental health symptoms with sex or household income. We calculated coefficients, 95% confidence intervals, and P values for each SBQ dimension. Statistical significance was set at P < 0.05; all statistical analyses were performed using IBM SPSS Statistics (version 27).
Results
Participants
The number of children with complete data was as follows, for QLSKC (in 1993), QLSCD (in 2009), and QSCDK (in 2022), respectively: 1665 (83% of original cohort), 1305 (62%), and 3871 (100%). Each had about 50% boys (Table 1). Sociodemographic data reflected temporal trends in that both maternal and paternal ages at the birth of the child were somewhat higher in the most recent cohort, parental birthplaces outside of Canada increased with greater immigration, and high school graduation and employment rates increased for both parents. However, relative family socioeconomic status remained similar between the 3 cohorts.
Table 1.
Sociodemographic characteristics, collected at child’s age 11 years, for each of the three cohorts, after weightinga
| QLSKC 1993 (n = 1665) |
QLSCDb 2009 (n = 1305) |
QSCDK 2022 (n = 3871) |
P valuec | |
|---|---|---|---|---|
| Child age, years, med [min, max] | 11.0 [10.5, 11.5] | 11.3 [9.7, 12.6] | 11.1 [10.4, 12.5] | < 0.001 |
| Child sex, boys, n (%) | 834 (50.1) | 667 (51.1) | 1947 (50.3) | 0.84 |
| Maternal age at child’s birth, years, mean (SD) | 26.9 (4.5) | 28.9 (5.3) | 31.1 (4.9) | < 0.001 |
| Paternal age at child’s birth, years, mean (SD) | 29.5 (4.9) | 31.8 (5.6) | 33.9 (5.9) | < 0.001 |
| Maternal birthplace, Canada, n (%) | 1507 (96.9) | 1138 (87.3) | 2803 (76.3) | < 0.001 |
| Paternal birthplace, Canada, n (%) | 1406 (95.3) | 1051 (87.4) | 2377 (74.6) | < 0.001 |
| Maternal high school graduation, yes, n (%) | 1207 (74.8) | 1094 (87.8) | 3523 (96.2) | < 0.001 |
| Paternal high school graduation, yes, n (%) | 1034 (70.1) | 896 (84.7) | 2969 (93.9) | < 0.001 |
| Maternal employment, yes, n (%) | 984 (72.6) | 1048 (84.1) | 3228 (88.5) | < 0.001 |
| Paternal employment, yes, n (%) | 1149 (89.8) | 1018 (96.2) | 3050 (96.6) | < 0.001 |
| Non intact family, yes, n (%) | 324 (23.6) | 479 (38.1) | 902 (23.3) | < 0.001 |
| Family socioeconomic status, n (%) | < 0.001 | |||
| Disadvantaged | 296 (22.1) | 279 (22.1) | 724 (18.7) | |
| Medium | 716 (53.4) | 631 (50.0) | 2261 (58.4) | |
| Advantaged | 328 (24.5) | 351 (27.8) | 890 (23.0) | |
| Number of children in the household | - | 2.3 (0.87) | 2.4 (0.9) | < 0.001 |
| Language spoken at home | < 0.001 | |||
| French and other | 859 (98.0) | 1101 (87.6) | 3263 (84.3) | |
| English and other (except French) | 8 (0.9) | 112 (8.9) | 337 (8.7) | |
| Neither French nor English | 10 (1.1) | 44 (3.5) | 271 (7.0) |
Abbreviations: QLSKC, Quebec Longitudinal Study of Kindergarten Children, 1993; QLSCD, Quebec Longitudinal Study of Child Development, 2009; QSCDK, Quebec Survey of Child Development in Kindergarten, 2022
Data compiled from final master files of each cohort, © Gouvernement du Québec, Institut de la statistique du Québec
a Inverse probability weighting to preserve representativeness of each cohort at baseline, considering loss of participants due to attrition
b In the QLSCD study, as age 11 years was not directly available, we used the mean of values at ages 10 and 12
c Comparisons between cohorts using t-test, Chi-square test, or ANOVA tests, as appropriate. Significance set at P < 0.05
In the 1993 cohort, participants differed from non-participants (excluded due to attrition or to incomplete data on mental health symptoms) in parental age at childbirth, parental birthplace, and family status (Supplementary Table 4). In addition, in the 2009 cohort, non-participants were more likely to be boys and from more disadvantaged backgrounds (Supplementary Table 5). After inverse probability weighting, these sociodemographic differences were significantly reduced (Supplementary Tables 6, 7). No child was excluded from the QSCDK (2022) cohort.
Overall mental health symptoms scores
SBQ scores in all cohorts showed good internal consistency for each dimension, at 0.72–0.88, with most ≥ 0.85 (Supplementary Table 2). Emotional distress and impulsive/hyperactive/inattentive symptom scores, as rated by classroom teachers, were generally higher in 2022 than in 2009, and higher in 2009 than in 1993 (Cohen’s d indicating very small-to-small effect sizes of 0.12 and 0.26 for emotional distress, respectively; and of 0.06 and 0.25 for impulsive/hyperactive/inattentive symptoms, respectively; P < 0.001) (Table 2 Fig. 1). Teacher ratings of disruptive behaviour symptoms showed lower scores in 2022 than in 2009, but higher in 2009 than 1993, with very small effect sizes (Cohen’s d of −0.15 and 0.09, respectively) (Table 2; Fig. 1). Note that observed differences between 2022 and 2009 in this paragraph would include effects of the COVID-19 pandemic, if any.
Table 2.
Mental health symptom scores at age 11 years, as per teacher-rated social behavior questionnaire
| Mean scoresa | QLSKC 1993 (n = 1665) |
QLSCDb 2009 (n = 1305) |
QSCDK 2022 (n = 3871) |
P valuec |
|---|---|---|---|---|
| Emotional distress | 1.86 (2.03) | 2.39 (2.08) | 2.67 (2.42) | < 0.001 |
| Impulsive/Hyperactive/Inattentive | 2.15 (2.40) | 2.76 (2.46) | 2.90 (2.57) | < 0.001 |
| Disruptive behaviors | 1.17 (1.82) | 1.34 (1.86) | 1.07 (1.78) | < 0.001 |
Abbreviations: QLSKC, Quebec Longitudinal Study of Kindergarten Children; QLSCD, Quebec Longitudinal Study of Child Development; QSCDK, Quebec Survey of Child Development in Kindergarten
Data compiled from final master files of each cohort, © Gouvernement du Québec, Institut de la statistique du Québec
a All values shown as mean (SD)
b In the QLSCD study, as age 11 years was not directly available, we used the mean of values at ages 10 and 12
c Comparisons between cohorts using ANOVA tests, significance set at P < 0.05
Fig. 1.
Mental health symptom scores at age 11 years (as per teacher-rated Social Behavior Questionnaire) in each of the 3 cohorts, according to household income. Abbreviations: QLSKC, Quebec Longitudinal Study of Kindergarten Children, 1993; QLSCD, Quebec Longitudinal Study of Child Development, 2009; QSCDK, Quebec Survey of Child Development in Kindergarten, 2022. Data compiled from the final master files of each cohort, © Gouvernement du Québec, Institut de la statistique du Québec. We used Student’s t-tests to compare mean symptom scores. Effect sizes were determined using Cohen’s d tests: very small, < 0.20; small, 0.20–0.50; medium, 0.50–0.80; large, 0.80–1.20; very large, 1.20–2.0; huge, >2.0. * P< 0.05; ** P < 0.01; *** P < 0.001
Socioeconomic and gender differences
In all cohorts, children from disadvantaged households had higher symptom scores in all dimensions than children in advantaged households (Fig. 1). Household income group differences remained relatively stable over time.
Boys showed significantly higher scores than girls in impulsive/hyperactive/inattentive symptoms and disruptive behaviours (P < 0.001) but not in emotional distress (Fig. 2; Table 3). Among both boys and girls, emotional distress and impulsive/hyperactive/inattentive symptoms were generally higher in 2022 than in 2009, and higher in 2009 than in 1993. However, in both boys and girls, disruptive behaviour symptom scores were highest in 2009.
Fig. 2.
Mental health symptom scores at age 11 years (as per teacher-rated Social Behavior Questionnaire) in each of the 3 cohorts, according to child sex. Abbreviations: QLSKC, Quebec Longitudinal Study of Kindergarten Children, 1993; QLSCD, Quebec Longitudinal Study of Child Development, 2009; QSCDK, Quebec Survey of Child Development in Kindergarten, 2022. Data compiled from the final master files of each cohort, © Gouvernement du Québec, Institut de la statistique du Québec. We used Student’s t-tests to compare mean symptom scores. Effect sizes were determined using Cohen’s d tests: very small, < 0.20; small, 0.20–0.50; medium, 0.50–0.80; large, 0.80–1.20; very large, 1.20–2.0; huge,>2.0. * P < 0.05; ** P < 0.01; *** P < 0.001
Table 3.
Mental health mean scores in the three cohorts according to child sex
| QLSKC 1993 n = 1,665 |
QLSCD 2009 n = 1,305 |
QSCDK 2022 n = 8,217 |
P valuea | |
|---|---|---|---|---|
| Emotional distress | ||||
| Boys | 1.90 (2.10) | 2.61 (2.23) | 2.67 (2.46) | < 0.001 |
| Girls | 1.83 (1.95) | 2.16 (1.89) | 2.68 (2.37) | < 0.001 |
| Impulsive/Hyperactive/Inattentive | ||||
| Boys | 2.85 (2.62) | 3.58 (2.59) | 3.60 (2.72) | < 0.001 |
| Girls | 1.44 (1.93) | 1.91 (1.98) | 2.19 (2.20) | < 0.001 |
| Disruptive behaviors | ||||
| Boys | 1.74 (2.13) | 1.89 (2.14) | 1.52 (2.09) | < 0.001 |
| Girls | 0.60 (1.19) | 0.76 (1.29) | 0.62 (1.25) | 0.03 |
Abbreviations: QLSKC, Quebec Longitudinal Study of Kindergarten Children; QLSCD, Quebec Longitudinal Study of Child Development; QSCDK, Quebec Survey of Child Development in Kindergarten
Data compiled from final master files of each cohort, © Gouvernement du Québec, Institut de la statistique du Québec
a All values shown as mean (SD)
b In the QLSCD study, as age 11 years was not directly available, we used the mean of values at ages 10 and 12
c Comparisons between cohorts using ANOVA tests, significance set at P < 0.05
Discussion
Despite growing concerns of worsening mental health in children, few studies have directly compared the rates of mental health symptoms in representative, population-based cohorts assessed cross-sectionally at the same age but separated by decades in time. In this study, we compared mental health symptoms in 11-year-olds in 3 population-based cohorts (i.e. in 1993, 2009, 2022, respectively), using comparable measures (validated questionnaire) completed by comparable raters (classroom teachers). We found that classroom teachers in 2022 reported higher levels of emotional distress and impulsivity/hyperactivity/inattention symptoms in their students than teachers in 2009; and teachers in 2009 reported higher levels than those in 1993. Effect sizes between 1993, 2009, and 2022 were small to very small. However, for disruptive behaviours, teachers in 2022 reported lower levels than did teachers of 11-year-olds in 2009, though 2009 was higher than 1993. Viewed differently, our results suggest that 2 out of 3 types of symptoms were scored higher in post-COVID times than in a cohort assessed 13 years earlier, while the third actually decreased; supporting concerns that COVID may represent an additional aggravating factor for certain mental health symptoms in 11-year-old children.
Several studies have attempted to address the temporal trends of increasing mental health issues in youth [3, 4, 11, 16]. However, whereas we focused on children on the cusp of puberty, others focused on older ages, namely adolescents and young adults [3, 16]. For instance, Botha et al. in Australia concluded that “poorer mental health of Millennials is driving the apparent deterioration in population-level mental health” [3]. One British study in 11-year-olds of 3 different generations stopped at 2012 [4]. Similarly to ours, they found ongoing differences between advantaged and disadvantaged participants. A 2023 comparison of 2 cohorts of children ages 4–17 in the UK showed that trajectories of emotional problems were different in the more recent cohort, and emerging earlier [2]. There are other studies and literature reviews, but some ended in 2014 or earlier, i.e. more than 10 years ago [22, 30] and others relied on different questionnaires and subscales, making further comparisons difficult [2, 3].
The difference in mean symptom scores between financially disadvantaged and advantaged children in our study remained stable over time, suggesting no intensification (or reduction) of inequalities. This finding is in line with a study on teenagers and young adults by Kiviruusu in Finnish youth between 2015 and 2023 [16], but not by an earlier 2015 paper by Collishaw and colleagues for 1999–2012 [22]. It is possible that economic factors specific to each country contributed to these discrepancies, notably during the COVID-19 pandemic when financial support, public health policies, and lockdowns differed considerably [31]. Previous research indicates that poverty likely has direct negative effects on child mental health [32]. For instance, reductions in public spending in several European countries over the past decade have disproportionately impacted social support services and mental health resources for economically disadvantaged families [33]. While our findings did not show a widening of inequalities, they underscored the ongoing importance of policy aimed at protecting the mental health of children from vulnerable socioeconomic backgrounds.
With respect to sex differences, boys in our study were consistently rated as exhibiting more symptoms than girls, except for emotional distress. These sex differences remained relatively stable over cohorts of different generations, similarly to previous studies [4, 16]. Boys are consistently more likely to exhibit externalizing disorders, such as ADHD and conduct disorders [3, 30]. Conversely, girls are more likely to present internalizing disorders, such as anxiety and depression [2, 5, 8]. However, we did not measure disorders or diagnoses but strictly observed symptoms. Notably, studies in older children/teenagers have reported increasing disparities between sexes over decades, with changes less pronounced in boys [2, 3, 8, 22]. Future research should explore targeted interventions to effectively address sex-specific vulnerabilities to mental health problems during critical developmental periods [2].
Several reasons may explain why teachers reported higher levels of mental health symptoms and same/lower levels of disruptive behaviours for same-aged children in recent generations. First, teachers may be more aware and better trained to recognize symptoms in children [8, 22]. Second, the widening range of symptoms included in diagnostic criteria over time (for example from the DSM 5 [19] compared to the DSM 4 [20]) might have led to increased identification of certain symptoms [34]. Third, lower stigma today may make children more comfortable expressing or displaying symptoms openly [8, 22]. Nonetheless, the documented increase in emergency visits for suicide attempts – which are less likely to be influenced by awareness or stigma – suggests that at least part of the reported increase reflects a factual worsening in youth mental health [35]. Finally, it is possible that children’s mental health has worsened over time [8, 22]. For example, Botha and colleagues (2023) found that Millennials (i.e. those born in the 1990 s) experienced higher emotional distress symptoms than those born both earlier and later, pointing to a possible cohort-specific effect [3]. However, a recent review concluded that increases in adolescent (ages 12–18 years) depression across 1991–2020 were more consistently attributed to period effects – such as the rise of social media, educational pressures, economic instability, increase of single-parent households, the COVID-19 pandemic – affecting all youths broadly, regardless of birth-cohort [36]. Interpreting such trends remains complex, as societal changes may differentially impact successive cohorts, particularly when experienced during sensitive developmental windows. Unlike studies attempting to disentangle age, period, and cohort effects simultaneously [3], the present study compared mental health symptoms at a fixed age (11 years) across three time points. By holding age constant, our design avoids the identification problem inherent in age–period–cohort models and allows for a more straightforward comparison of generational changes in mental health in late childhood.
Strengths and limitations
A key strength of our study was the comparison of cross-sectional data from a series of three population-based representative cohorts assessed in different decades in which participants turned age 11 years, allowing the generalizability of results to non-clinical samples. Second, assessments were conducted prospectively in real-world settings, more accurately portraying mental health in the community [37]. Third, we used ratings by classroom teachers, as these were available in all three cohorts, contrary to parent ratings. Teacher ratings are often used in population-based studies to inform children’s mental health and provide a sense of what is normative versus atypical (especially regarding externalizing symptoms) [38]; however, these ratings inherently reflect expectations specific to the school setting. Fourth, regardless of cohort, all children were assessed using the same validated instrument, namely the Social Behavior Questionnaire, which presents good psychometric properties [25], thereby limiting bias and interference from changes in diagnostic criteria.
Limitations included not adding parent-reported assessments, as these were available only partially in two cohorts and entirely unavailable in the third. Second, cohort databases were stored separately, limiting direct comparison to investigate whether the internal structure of the questionnaire or its use by teachers varied according to contemporary practice. We were also unable to investigate potential interactions between cohort, sex, and socioeconomic status, limiting our ability to formally assess associations between these factors. Third, two of our cohorts presented differential attrition; we offset this by using a robust method (inverse probability weighting) to maintain initial representativeness of each cohort [39]. Fourth, generalizability of results might be limited due to cultural differences within and outside Quebec. However, and importantly, all cohorts were representative of the Quebec population at their inception (other than residents of Indigenous areas). Further, the city of Montreal was included, which is multicultural and multiethnic [40], thereby reflecting various cultural backgrounds in the analyses. Lastly, we focused exclusively on children at age 11, an important developmental period just prior to the “turbulent teens”, but nonetheless only one age point on the trajectory from childhood to young adulthood. Further research could focus on including more timepoints in childhood to clarify developmental changes in mental health across childhood [2].
In conclusion, our research points to temporal trends of increasing mental health symptoms in children as young as age 11 years. There were differences in mental health symptoms between boys and girls, and in each cohort, those from low-income households exhibited the highest symptom levels. Considering that childhood alone accounts for approximately one-quarter of the total lifetime mental health burden [5, 41] and represents an economic cost of approximately 798€ billion per year in Europe [42], strategies for identifying individuals at risk and for designing preventive interventions hold the potential for meaningful individual, familial, and societal benefits [10]. Given these findings, policymakers should consider supporting early preventive interventions in elementary schools, tailored differentially for boys and for children of lower-income households. Further, clinicians may consider adding teachers’ perceptions when assessing children’s mental health, with respect to other children in the class.
Supplementary Information
Below is the link to the electronic supplementary material.
Author contributions
OC: conceptualization, investigation, methodology, writing, reviewing, editing. MO: conceptualization, methodology, writing, reviewing, editing, supervision. CG: conceptualization, writing, reviewing, editing. BDT: conceptualization, writing, reviewing. MCG: conceptualisation, writing, reviewing. OTF: conceptualization, methodology, data analysis. DB: writing, editing, reviewing. FV: conceptualization, methodology, writing, reviewing, editing. MB: conceptualization, methodology, writing, editing, reviewing. RET: conceptualization, methodology, writing, editing, reviewing. SMC: conceptualization, project administration, writing, editing, supervision.
Data availability
Participants of the study did not consent for publish sharing of their data. Thus, the data that support the findings of this study are not openly available but are accessible to researchers through the Centre d’accès aux données de recherche de l’Institut statistique du Québec. To access the data, researchers need to submit for evaluation a project (including for example the aim, analysis plan, researchers on the project) to the Research Data Access Point team. Further information and support can be found on the Research Data Access Point website (https ://www.stat.gouv.qc.ca/resea rch/#/accueil).
Declarations
Competing interests
The authors declare no competing interests.
References
- 1.Achenbach TM (1991) Manual for child behavior checklist 4–18, 1991 profile. Univ Vermont/Dept Psychiatry, Burlington, VT [Google Scholar]
- 2.Armitage JM, Kwong ASF, Tseliou F et al (2023) Cross-cohort change in parent-reported emotional problem trajectories across childhood and adolescence in the UK. Lancet Psychiatry 10:509–517. 10.1016/S2215-0366(23)00175-X [DOI] [PubMed] [Google Scholar]
- 3.Botha F, Morris RW, Butterworth P, Glozier N (2023) Generational differences in mental health trends in the twenty-first century. Proc Natl Acad Sci 120(e2303781120). 10.1073/pnas.2303781120 [DOI] [PMC free article] [PubMed]
- 4.Collishaw S, Furzer E, Thapar AK, Sellers R (2019) Brief report: a comparison of child mental health inequalities in three UK population cohorts. Eur Child Adolesc Psychiatry 28:1547–1549. 10.1007/s00787-019-01305-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kieling C, Buchweitz C, Caye A et al (2024) Worldwide prevalence and disability from mental disorders across childhood and adolescence: evidence from the global burden of disease study. JAMA Psychiatry 81:347–356. 10.1001/jamapsychiatry.2023.5051 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lebrun-Harris LA, Ghandour RM, Kogan MD, Warren MD (2022) Five-Year trends in US children’s health and Well-being, 2016–2020. JAMA Pediatr 176:e220056. 10.1001/jamapediatrics.2022.0056 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.McGorry PD, Mei C, Dalal N et al (2024) The lancet psychiatry commission on youth mental health. Lancet Psychiatry 11:731–774. 10.1016/S2215-0366(24)00163-9 [DOI] [PubMed] [Google Scholar]
- 8.Patalay P, Gage SH (2019) Changes in millennial adolescent mental health and health-related behaviours over 10 years: a population cohort comparison study. Int J Epidemiol 48:1650–1664. 10.1093/ije/dyz006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Twenge JM, Joiner TE, Rogers ML, Martin GN (2018) Increases in depressive symptoms, suicide-Related outcomes, and suicide rates among U.S. Adolescents after 2010 and links to increased new media screen time. Clin Psychol Sci 6:3–17. 10.1177/2167702617723376 [Google Scholar]
- 10.Ravens-Sieberer U, Otto C, Kriston L et al (2015) The longitudinal BELLA study: design, methods and first results on the course of mental health problems. Eur Child Adolesc Psychiatry 24:651–663. 10.1007/s00787-014-0638-4 [DOI] [PubMed] [Google Scholar]
- 11.Kretschmer T, Hickman M, Doerner R et al (2014) Outcomes of childhood conduct problem trajectories in early adulthood: findings from the ALSPAC study. Eur Child Adolesc Psychiatry 23:539–549. 10.1007/s00787-013-0488-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Dreyer BP (2023) Stepping up to the Plate—The role of pediatricians in addressing the child and adolescent mental health crisis. JAMA Netw Open 6:e2249578. 10.1001/jamanetworkopen.2022.49578 [DOI] [PubMed] [Google Scholar]
- 13.de Oliveira JMD, Butini L, Pauletto P et al (2022) Mental health effects prevalence in children and adolescents during the COVID-19 pandemic: A systematic review. Worldviews Evid Based Nurs 19:130–137. 10.1111/wvn.12566 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Racine N, McArthur BA, Cooke JE et al (2021) Global prevalence of depressive and anxiety symptoms in children and adolescents during COVID-19: a meta-analysis. JAMA Pediatr 175:1142–1150 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Santomauro DF, Herrera AMM, Shadid J et al (2021) Global prevalence and burden of depressive and anxiety disorders in 204 countries and territories in 2020 due to the COVID-19 pandemic. Lancet 398:1700–1712. 10.1016/S0140-6736(21)02143-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kiviruusu O, Ranta K, Lindgren M et al (2024) Mental health after the COVID-19 pandemic among Finnish youth: a repeated, cross-sectional, population-based study. Lancet Psychiatry 11:451–460. 10.1016/S2215-0366(24)00108-1 [DOI] [PubMed] [Google Scholar]
- 17.Wolf K, Schmitz J (2023) Scoping review: longitudinal effects of the COVID-19 pandemic on child and adolescent mental health. Eur Child Adolesc Psychiatry 1–56. 10.1007/s00787-023-02206-8 [DOI] [PMC free article] [PubMed]
- 18. Institute for Health Metrics and Evaluation (2022) Global Burden of Disease. Institute for Health Metrics and Evaluation, University of Washington, Seattle, WA. https://www.healthdata.org
- 19.American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders (DSM-5®). American Psychiatric Publishing, Washington, UNITED STATES [Google Scholar]
- 20. American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders: DSM-IV. American Psychiatric Association
- 21.Newlove-Delgado T, McManus S, Sadler K et al (2021) Child mental health in England before and during the COVID-19 lockdown. Lancet Psychiatry 8:353–354. 10.1016/S2215-0366(20)30570-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Collishaw S (2015) Annual research review: secular trends in child and adolescent mental health. J Child Psychol Psychiatry 56:370–393. 10.1111/jcpp.12372 [DOI] [PubMed] [Google Scholar]
- 23.Deng J, Zhou F, Hou W et al (2023) Prevalence of mental health symptoms in children and adolescents during the COVID-19 pandemic: A meta-analysis. Ann N Y Acad Sci 1520:53–73 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Harrison L, Carducci B, Klein JD, Bhutta ZA (2022) Indirect effects of COVID-19 on child and adolescent mental health: an overview of systematic reviews. BMJ Global Health 7:e010713 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Collet OA, Orri M, Tremblay RE et al (2022) Psychometric properties of the social behavior questionnaire (SBQ) in a longitudinal population-based sample. Int J Behav Dev 01650254221113472. 10.1177/01650254221113472
- 26.Rouquette A, Côté SM, Pryor LE et al (2014) Cohort profile: the Quebec longitudinal study of kindergarten children (QLSKC). Int J Epidemiol 43:23–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Orri M, Boivin M, Chen C et al (2020) Cohort profile: Quebec longitudinal study of child development (QLSCD). Soc Psychiatry Psychiatr Epidemiol 10.1007/s00127-020-01972-z [DOI] [PubMed]
- 28.Simard M, Lavoie A, Audet N (2018) Enquête québécoise sur le développement des enfants à la maternelle 2017. Institut de la statistique du Québec, Québec. https://statistique.quebec.ca/fr/document/enquete-quebecoise-sur-le-developement-des-enfants-a-la-maternelle-2017-portrait-statistique-pour-le-quebec-et-ses-regions-administratives
- 29.Cohen J (1988) Statistical power analysis for the behavioral sciences, 2nd edn. Routledge, Hillsdale, N.J [Google Scholar]
- 30.Polanczyk GV, Willcutt EG, Salum GA et al (2014) ADHD prevalence estimates across three decades: an updated systematic review and meta-regression analysis. Int J Epidemiol 43:434–442. 10.1093/ije/dyt261 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.The World Bank, UNESCO, UNICEF (2021) The State of the Global Education Crisis: A Path to Recovery. Washington, DC; Paris; New York: The World Bank, UNESCO, and UNICEF. https://documents1.worldbank.org/curated/en/416991638768297704/pdf/The-State-of-the-Global-Education-Crisis-A-Path-to-Recovery.pdf
- 32.Costello EJ, Compton SN, Keeler G, Angold A (2003) Relationships between poverty and psychopathology: a natural experiment. JAMA 290:2023–2029. 10.1001/jama.290.15.2023 [DOI] [PubMed] [Google Scholar]
- 33.Stuckler D, Reeves A, Loopstra R et al (2017) Austerity and health: the impact in the UK and Europe. Eur J Public Health 27:18–21. 10.1093/eurpub/ckx167 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Cummings JR, Hu X, Marchak J et al (2025) Trends in mental health diagnoses among publicly insured children. JAMA. 10.1001/jama.2025.4605 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kim Y, Krause TM, Lane SD (2023) Trends and seasonality of emergency department visits and hospitalizations for suicidality among children and adolescents in the US from 2016 to 2021. JAMA Netw Open 6:e2324183. 10.1001/jamanetworkopen.2023.24183 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Askari MS, Belsky DW, Olfson M et al (2024) An integrative literature review of birth cohort and time period trends in adolescent depression in the united States. Soc Psychiatry Psychiatr Epidemiol 59:899–915. 10.1007/s00127-023-02527-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Moffitt TE, Caspi A, Taylor A et al (2010) How common are common mental disorders? Evidence that lifetime prevalence rates are doubled by prospective versus retrospective ascertainment. Psychol Med 40:899–909. 10.1017/S0033291709991036 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Green J, Staff L, Bromley P et al (2021) The implications of face masks for babies and families during the COVID-19 pandemic: A discussion paper. J Neonatal Nurs 27:21–25. 10.1016/j.jnn.2020.10.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Ridgeway G, Kovalchik SA, Griffin BA, Kabeto MU (2015) Propensity score analysis with survey weighted data. J Causal Inference 3:237–249. 10.1515/jci-2014-0039 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Government of Canada SC (2024) Ethnocultural diversity in Canadian cities. https://www.statcan.gc.ca/o1/en/plus/7238-ethnocultural-diversity-canadian-cities. Accessed 22 May 2025
- 41.McGrath JJ, Al-Hamzawi A, Alonso J et al (2023) Age-of-onset and cumulative risk of mental disorders: A cross-national analysis of population surveys data based on 156,331 respondents from 29 countries. Lancet Psychiatry 10:668–681. 10.1016/S2215-0366(23)00193-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Wittchen HU, Jacobi F, Rehm J et al (2011) The size and burden of mental disorders and other disorders of the brain in Europe 2010. Eur Neuropsychopharmacol 21:655–679. 10.1016/j.euroneuro.2011.07.018 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Participants of the study did not consent for publish sharing of their data. Thus, the data that support the findings of this study are not openly available but are accessible to researchers through the Centre d’accès aux données de recherche de l’Institut statistique du Québec. To access the data, researchers need to submit for evaluation a project (including for example the aim, analysis plan, researchers on the project) to the Research Data Access Point team. Further information and support can be found on the Research Data Access Point website (https ://www.stat.gouv.qc.ca/resea rch/#/accueil).