Sage Journals: Discover world-class research

Abstract

Objective:

The dual-factor model of mental health postulates a role for positive mental health, alongside mental illness, in determining mental health care needs. Informed by this model, the present study delineated profiles of social-emotional competencies and difficulties during middle childhood in a population-based sample of girls and boys and determined their association with adolescent mental disorder diagnoses.

Method:

Latent profile analyses were conducted across five indices of social-emotional competency and four indices of psychopathology that were measured by questionnaire self-report among 13,349 girls and 13,488 boys at age ~11 years. The association of the profiles with adolescent presentations to hospital or ambulatory services (ages ~12–17 years) were determined using logistic regression.

Results:

Analyses yielded five profiles in each sex: complete mental health (44% girls; 42% boys), average mental health (30%; 33%), internalising symptoms with poor relationship skills (9%; 7%), externalising symptoms with poor self-management (12%; 9%) and low mental health (5%; 10%). Profiles associated differentially with adolescent presentations with any mental disorder, externalising disorders, internalising disorders and self-harm/suicidal ideation, identified in linked health records. Greater odds of any and specific mental disorder diagnoses were characteristic of all profiles relative to average mental health (adjusted odds ratios [aOR]: 1.7–3.3) except complete mental health (aOR: 0.7–0.9), with different strengths of association according to profile.

Conclusion:

Combining information on social-emotional competencies and psychopathology in middle childhood may help refine the provision of mental health promotion and early intervention to alleviate adolescent mental disorder.

Keywords

Latent profile analysis mental disorders dual-factor model of mental health wellbeing psychological distress

With rates of youth mental disorders and psychological distress increasing globally (Brazel et al., 2023; Sacco et al., 2024; Slade et al., 2024; Tkacz and Brady, 2021), and adolescence marking the peak period of onset for many mental disorders (Solmi et al., 2022), school-based tiered mental health and wellbeing interventions have become a popular, evidence-based avenue of mitigating burden on mental health services (Grové and Laletas, 2020; Humphrey, 2023; Lee et al., 2023). These tiered interventions incorporate universal prevention for all students, alongside targeted interventions for students with emerging psychopathology, and individualised intensive interventions for students with established mental health difficulties. By universally fostering social-emotional wellbeing for all children, these approaches seek to prevent an increasing accumulation of need at the targeted and intensive intervention levels (de Bruin et al., 2023; Greenberg et al., 2017; Guest et al., 2024). They align well with growing evidence in support of a ‘dual-factor model of mental health’, wherein positive mental health and mental illness represent distinct (albeit correlated) factors rather than the extreme ends of a single dimension (Greenspoon and Saklofske, 2001; Suldo and Shaffer, 2008). Positive mental health is defined by the World Health Organisation (2025) as ‘a state of mental well-being that enables people to cope with the stresses of life, realize their abilities, learn and work well, and contribute to their community’. The model postulates that consideration of indicators of social-emotional wellbeing alongside symptoms of ill health may better differentiate children’s need for interventions (Iasiello and Van Agteren, 2020) and help inform the level of intervention required. Informed by the model, the present study sought to examine population patterns in children’s co-presentation of social-emotional competencies and symptoms of psychopathology during middle childhood and determine how these patterns related to children’s subsequent presentations to health services with mental disorders during adolescence.

The dual-factor model of mental health proposes four different expressions (profiles) of social-emotional wellbeing and psychopathology in the population, namely: (1) ‘complete mental health’, represented by high wellbeing and low psychopathology; (2) ‘troubled’, reflected in low wellbeing and high psychopathology; (3) ‘vulnerable’, represented by low wellbeing but also low psychopathology; and (4) ‘symptomatic but content’, reflected in high psychopathology but also high wellbeing (Greenspoon and Saklofske, 2001; Lyons et al., 2013). Commonly, prior studies have tested the model by imposing threshold scores to distinguish the four profiles of high versus low levels of wellbeing and psychopathology and then examined their associations with predictors or outcomes (Iasiello and Van Agteren, 2020). More recently, several studies have adopted person-centred techniques as a more flexible and data-driven approach to identifying underlying subgroups within the population, and examined their alignment with the four theorised expressions (Clark et al., 2024). These methods, including latent class/profile analyses (LCA/LPA), classify individuals into homogeneous subgroups based on response patterns across a set of wellbeing and psychopathology indices (Collins and Lanza, 2009).

Varying support for the four expressions in the dual-factor model has emerged from these analyses, conducted in child and adolescent community samples across a diversity of social-emotional wellbeing and psychopathology indicators, as summarised in Table 1. Across these studies, differing numbers of profiles emerged, including 3- (Clark and Malecki, 2022), 4- (Kim et al., 2018; Moore et al., 2019; Petersen et al., 2020) and 5-profile (Gregory et al., 2024) solutions. Multiple investigations were unable to detect a ‘vulnerable’ profile (Clark and Malecki, 2022; Gregory et al., 2024; Kim et al., 2018; Moore et al., 2019) with average mental health profiles (moderate wellbeing and moderate psychopathology) typically emerging instead (Gregory et al., 2024; Kim et al., 2018; Moore et al., 2019). These divergent findings in the number of profiles and their characteristics may reflect differences in sample size, cultural settings, developmental stages and/or measures used, with some assessments encompassing only internalising or externalising symptoms or only social (interpersonal) or emotional (intrapersonal) wellbeing. Notably, however, LCA/LPA studies stratified by gender have identified consistent profile solutions for girls and boys (Gregory et al., 2024; Kim et al., 2018), but with boys being more likely to present the ‘complete mental health’ rather than ‘troubled’ or ‘symptomatic but content’ profiles (Clark and Malecki, 2022).

Table 1.

Summary of latent profile/class analyses conducted among children and adolescents and informed by the dual-factor model of mental health.

Study Reference	Location	Sample size	Age at assessment	Indices	Class/Profile Solutions	Equivalent Class/Profile within the Dual-Factor Model
Clark and Malecki (2022)	United States	404	11–14 years(Grades 6–8)	- Life satisfaction (SLSS)- Affect (PANAS-C)- Internalising problems (YSR)- Externalising problems (YSR)	1. Complete mental health (55%)2. Symptomatic but content (34%)3. Troubled (11%)	1. Complete mental health2. Symptomatic but content3. Troubled
Gregory et al. (2024)	Australia	75,757	8–18 years (Grades 4–12)	- Life satisfaction (SLS)- Optimism (MDI)- Happiness (EPOCH)- Sadness (SPQ)- Worries (WEC)	1. Complete mental health (23%)2. Good mental health (33%)3. Moderate mental health (27%)4. Symptomatic but content (9%)5. Troubled (8%)	1. Complete mental health — —2. Symptomatic but content3. Troubled
Kim et al. (2018)	South Korea	1,757	9–12 years(Grades 4–6)	- Gratitude (SEHS-P)- Zest (SEHS-P)- Optimism (SEHS-P)- Persistence (SEHS-P)- Emotional difficulties (M&Ms)- Behavioural difficulties (M&Ms)	1. Flourishing (14–17%) ^a 2. Moderate flourishing (29–38%)3. Moderate languishing (38–40%)4. Languishing (8–17%)	1. Complete mental health — —2. Troubled
Moore et al. (2019)	United States	332	~14–17 years(Grades 9–11)	- Belief in self (SEHS-S)- Belief in others (SEHS-S)- Emotional competence (SEHS-S)- Engaged living (SEHS-S)- Internalising problems (SDQ)- Externalising problems (SDQ)	1. Complete mental health (21–41%)^b 2. Moderately mentally healthy (32–44%)3. Symptomatic but content (20–31%)4. Troubled (4–6%)	1. Complete mental health —2. Symptomatic but content3. Troubled
Petersen et al. (2020)	United Kingdom	3,340	8–9 years(Grades 4–5)	- Subjective Wellbeing (KIDSCREEN-27)- Conduct problems (SDQ)- Emotional symptoms (SDQ)	1. Complete mental health (57%)2. Vulnerable (13%)3. Conduct problems but content (12%)4. Emotional symptoms but content (18%)	1. Complete mental health2. Vulnerable3. Symptomatic but content4. Symptomatic but content

SLSS = Students’ Life Satisfaction Scale (Huebner, 1991); PANAS-C = Positive and negative Affect Scale for Children (Laurent et al., 1999); YSR = Youth Self-Report (Achenbach and Rescorla, 2001); SLS = Satisfaction with Life Scale (Gadermann et al., 2010); MDI = Middle Years Development Index (Schonert-Reichl et al., 2013); EPOCH = EPOCH Measure of Adolescent Wellbeing (Kern et al., 2015); SPQ = Seattle Personality Questionnaire (Kusche et al., 1988); WEC = Wellbeing and Engagement Collection (Gregory et al., 2021); SEHS-p = Social Emotional Health Survey – Primary (Furlong et al., 2013); M&Ms = Me and My School Questionnaire (Deighton et al., 2013); SEHS-S = Social-Emotional Health Survey – Secondary (You et al., 2014); SDQ = Strengths and Difficulties Questionnaire (Goodman, 2001); KIDSCREEN-27 (Ravens-Sieberer et al., 2007).

differentiated among girls and boys.

differentiated by grade levels.

Limited research has explored longitudinal associations between these childhood/adolescent profiles and later mental illness. In their study of US students (Grades 9–11), Moore et al. (2019) indicated significantly higher later mean depression and anxiety symptoms in Grade 12 for the ‘troubled’ and ‘symptomatic but content’ profiles relative to the ‘complete mental health’ and ‘moderate mental health’ profiles. Other evidence may be inferred from a study that explored associations between six early childhood profiles (derived on eight teacher-reported indices of social-emotional competencies and difficulties among ~34,000 Canadian Kindergarten students aged ~5 years) and mental health conditions recorded in health records up to the age of 14 years (Thomson et al., 2019). Although not explicitly informed by the dual-factor model, four of the six profiles, were broadly consistent with the ‘complete mental health’, ‘troubled’ and ‘vulnerable’ profiles and a ‘symptomatic but content’ profile featuring prominent aggressive and hyperactive/inattentive behaviours. In addition, there were two profiles akin to the ‘moderate’ and ‘good mental health’ profiles identified by Gregory et al. (2024) and others, though with varied patterns of particular social-emotional competencies and difficulties. After adjusting for various sociodemographic covariates, relative to ‘complete mental health’, both the ‘troubled’ and the ‘symptomatic but content’ profiles had increased odds of conduct disorders, attention-deficit/hyperactivity disorders (ADHD) and of multiple disorders, but no association with depressive or anxiety disorders. Increased odds of multiple disorders were also apparent for the ‘vulnerable’ relative to ‘complete mental health’ profile. Further research is needed to determine whether associations between the profiles and depressive/anxiety disorders might be identified during adolescence, as children mature towards the peak age for onset of these disorders (McGrath et al., 2023; Solmi et al., 2022), and/or when using self-reports gathered during middle childhood relative to kindergarten-teacher reports of children’s social-emotional functioning. Given recognised sex-based disparities in mental health disorder prevalence and patterns (Eaton et al., 2012), research is also needed to determine whether the association of the profiles with later mental disorders varies by sex.

The present study examined these questions using data from a large population-based sample of ~27,000 children drawn from the New South Wales Child Development Study (NSW-CDS; Green et al., 2024). LPA analyses determined middle childhood mental health profiles, separately for girls and boys, according to nine indices (described in Table 2) of social-emotional competencies (self-awareness, self-management, social awareness, relationship skills and responsible decision-making) and symptoms of psychopathology (emotional symptoms, peer relationship problems, conduct problems and hyperactivity-inattention). This study subsequently investigated associations between these profiles and later presentations to health services with mental disorder diagnoses during adolescence (~12–17 years). Building on prior Australian research showing girls (cf. boys) at a higher risk of mental health service use, internalising disorders and suicidality (Islam et al., 2020; Watkeys et al., 2024), the study examined, separately by sex, the association of the middle childhood profiles with adolescent health services presentations for any mental disorder, and externalising disorders, internalising disorders and self-harm/suicidal ideation. It was hypothesised that the LPA solutions would demonstrate partial support for the dual-factor model among girls and boys, with the emergence of a ‘moderate mental health’ profile likely in place of a ‘vulnerable’ profile. It was also hypothesised that children classified into profiles characterised by lower social-emotional competencies and higher psychopathology scores would demonstrate significantly greater odds of later mental disorders than children in a profile characterised by higher competencies and lower psychopathology scores.

Table 2.

Social-emotional competencies (from the Middle Childhood Survey – Social-Emotional Learning) and social-emotional difficulties (from the Strengths and Difficulties Questionnaire).

Domain		Competency	Definition
Social-Emotional competencies	Intrapersonal	Self-Awareness	The ability to understand one’s own emotions, thoughts and behaviours, and have a well-grounded sense of confidence and purpose.Example item: I easily learn my school work
	Intrapersonal	Self-Management	The ability to manage one’s emotions, thoughts and behaviours effectively in different situations and to achieve one’s goals.Example item: When I am angry, I throw or break things [reversed]
	Interpersonal	Social Awareness	The ability to understand and empathise with the perspectives of others, including those from diverse backgrounds, cultures and contexts.Example item: I try to be nice to other people. I care about their feelings
	Interpersonal	Relationship Skills	The ability to establish and maintain healthy and supportive relationships with diverse individuals and groups through effective communication and collaboration.Example item: My school is a place where I get on well with other students in my class
	Decision-Making	Responsible Decision-Making	The ability to make caring and constructive decisions about one’s personal behaviour and to consider the consequences of one’s actions.Example item: I keep my things in order
Social-Emotional Difficulties	Internalising	Emotional Symptoms	Assesses the experiences of emotional distress, including nervousness, worry and depression.Example item: I am often unhappy, depressed or tearful
	Internalising	Peer Relationship Problems	Assesses difficulties forming positive friendships with peers of the same age and/or experiences of negative peer relationships.Example item: Other children or young people pick on me or bully me
	Externalising	Conduct Problems	Assesses behavioural problems, including difficulty controlling temper, disobedience, fighting, lying or stealing.Example item: I get very angry and often lose my temper
	Externalising	Hyperactivity-Inattention	Assesses difficulties with staying still, concentrating, thinking before acting and completing tasks.Example item: I am easily distracted, I find it difficult to concentrate

Method

Participants and procedure

Participants in this study (n = 26,837; 49.7% girls) were a population-based subsample of children in the NSW-CDS cohort who completed the self-report Middle Childhood Survey in 2015 (when aged 11–12 years; Laurens et al., 2017) and had linked health records available in the study’s third record linkage (Green et al., 2024). Supplementary Table S1 provides sample demographic details, relative to the full cohort. As at March 2021 (the last date of linked mental health records), the study sample was, on average, 17.52 years old (SD = 0.38; Median = 17.50; IQR = 17.25–17.75 years). Children who had mental disorder presentations prior to January 2016 (n = 1,903) were retained in the derivation of mental health profiles but excluded from analyses of the association of profiles with adolescent mental disorder diagnoses.

Ethical approval (HREC/18/CIPHS/49) for the linkage was provided by the NSW Population and Health Services (2019/ETH01571) and ACT Health Human Research Ethics Committees (2019.STE.00184), with relevant data custodian approvals. Probabilistic linkage of records was conducted by the Centre for Health and Record Linkage, managed by a third party under a waiver of consent process consistent with the Australian National Health and Medical Research Council’s (2007) National Statement of Ethical Conduct in Human Research. False linkage rate across all record sets was estimated at 0.5%.

Measures

Social-emotional competencies and difficulties

Children’s self-reports of five social-emotional competencies (as defined by the Collaborative for Academic, Social and Emotional Learning [CASEL], 2005, 2025) were assessed using the Middle Childhood Survey–Social-Emotional Learning (MCS-SEL; Carpendale et al., 2023a), namely: self-awareness (four items), self-management (three items), social awareness (five items), relationship skills (five items) and responsible decision-making (three items). Internalising and externalising psychopathology were assessed via the four self-report scales (five items per scale) of the Strengths and Difficulties Questionnaire (SDQ; Goodman, 2001), namely: emotional symptoms, peer relationship problems, conduct problems and hyperactivity/inattention. All items were rated on a three-level response scale: not true (scored 0), somewhat true (1) and certainly true (2). Internal consistency of the five competencies in this sample is satisfactory (McDonald’s ω: 0.65–0.76; Carpendale et al., 2023a; 2023b) and acceptable to good for the four psychopathologies (ordinal α: 0.69–0.80; Laurens et al., 2017). LPA was conducted on averaged, centred scores (according to sex-based means) for these nine constructs.

Mental disorders

Health service contacts for mental disorder diagnoses were recorded in any of four record sets provided by the NSW Ministry of Health: Admitted Patient Data Collection (APDC; records from January 2002–March 2021; including presentations in public hospitals, public psychiatric hospitals, multi-purpose services, private hospitals and private day procedure centres in NSW); Emergency Department Data Collection (EDDC; January 2005–March 2021; including presentations at emergency departments of public hospitals and in-scope contracted private hospitals in NSW); NSW Mental Health Ambulatory data collection (MH-AMB; November 2002–December 2020, which includes care provided to non-admitted patients in mental health day programmes, psychiatric outpatient and community outreach services [e.g. home visits]); and Controlled Drugs Data Collection (CoDDaC: June 2004–June 2020, which includes the prescription of Schedule 8 drugs such as psychostimulants).

Four binary outcomes were explored: (1) any mental disorder diagnoses; (2) externalising disorders, including conduct disorder, mixed disorders of conduct and emotion, cluster B personality disorders and substance use disorders; (3) internalising disorders, including affective and anxiety disorders and (4) self-harm or suicidal ideation presentations (Supplementary Table S2 details the ICD-10-AM codes related to these outcomes). Presentation of these disorders from January 2016 (the year after Middle Childhood Survey completion in 2015) were coded 1. The reference group (coded 0) for all analyses was the group of children who had no recorded mental health diagnoses at any time.

Sociodemographic variables

Each child’s sex (as assigned at birth) and Aboriginal and Torres Strait Islander background (0 = non-Indigenous; 1 = Indigenous) was determined by consensus across NSW-CDS linked records. Area socioeconomic disadvantage was coded from the child’s residential postcode in 2015 (0 = bottom four quintiles, reflecting less disadvantage; 1 = highest quintile, reflecting greatest disadvantage), according to the Socio-Economic Indexes for Areas (SEIFA) Index of Relative Socio-Economic Disadvantage (IRSD) ratings (Australian Bureau of Statistics, 2011). Children’s geographical remoteness was coded according to children’s residential postcodes using ratings from the Accessibility and Remoteness Index of Australia (0 = major cities; 1 = inner regional, outer regional, remote and very remote areas; Department of Health and Aged Care, 2001).

Statistical analysis

Latent profile analyses

Latent profiles were specified in Mplus version 8 (Muthén and Muthén, 1998) for girls and boys, separately. Models were repeated, specifying two to k profiles, until the model’s best loglikelihood value was not replicated, indicating a local solution rather than a general solution. Profiles were estimated using Maximum Likelihood with Robust standard errors, accounting for non-normality (Vermunt and Magidson, 2002). Models were specified with 500 initial-stage random starts, 50 final-stage optimizations (Nylund-Gibson and Choi, 2018) and, for the time-intensive likelihood bootstrap tests, 100 draws were requested (Tein et al., 2013). The optimal model (i.e. number of profiles) was determined by evaluating class separation (success of classification of individuals) via entropy values; model parsimony via log-likelihood values, Akaike Information Criteria (AIC) and Bayesian Information Criterion (BIC); and distributional properties via the adjusted Lo-Mendell-Rubin adjusted likelihood ratio test (LMR-LRT) and bootstrapped likelihood ratio test (BLRT) (Weller et al., 2020). Entropy values of 0.40, 0.60 and 0.80 were interpreted as low, moderate and high class separation, respectively (Clark and Muthén, 2009). Smaller loglikelihood, AIC and BIC values were optimal, and significant (p < 0.05) LMR-LRT and BLRT values indicated improved model fit relative to the previous solution. Solutions with profile sizes of ⩾5% were preferred, on account of the practical applications of small profiles (Weller et al., 2020). Profile enumeration also considered practical and theoretical relevance.

Profiles and later mental health service use

Once profile solutions for girls and boys were selected and the most likely profile memberships extracted for each child, a series of bivariate (unadjusted) logistic regressions were conducted to explore associations between the mental health profiles (relative to an average/moderate mental health profile) and mental health presentations, with children who had no presentations as the reference group. Analyses were repeated controlling for sociodemographic covariates (Aboriginal/Torres Strait Islander background, area-based socioeconomic disadvantage and geographical remoteness). The logistic regression analyses yielded odds ratios and their 95% confidence intervals (CIs), with significance denoted by CIs that did not cross 1.00. Odds ratios of 1.00–1.49 (or 1.00–0.67) were interpreted as small effects, 1.50–2.49 (or 0.66–0.40) as medium, 2.50–4.00 (or 0.39–0.25) as large and >4.00 (or <0.25) as very large effects (Rosenthal, 1996).

Results

Latent profile analysis

For girls and boys, solutions are reported up to nine profiles (Table 3), beyond which the LMR-LRT was consistently non-significant, indicating no improvement in fit. With each increase in the number of profiles specified, small improvements were observed in AIC and BIC. For both samples, LMR-LRT was non-significant at the six-profile solution, indicating six profiles did not significantly improve fit above the five-profile model. In both samples, a five-profile solution was the last model before the smallest profiles consistently comprised less than 5% of children and before posterior probabilities fell below 80%. Accordingly, the most likely profile memberships from the five-profile solutions for girls and boys were extracted for each child.

Table 3.

Model fit comparison of latent profile analysis of social-emotional competencies and psychopathology among girls and boys.

Number of Profiles	Log Likelihood	AIC	BIC	Entropy	LMR-LRT	BLRT	Lowest Class Probability	Lowest Class Size	Lowest Class Proportion
Girls
1	-65,287.92	130,611.84	130,746.83	-	-	-	-	-	-
2	-52,519.67	105,095.33	105,305.31	0.868	25,270.48***	25,536.51***	0.94	3,674	27.5
3	-48,632.03	97,340.07	97,625.04	0.828	7,694.27***	7,775.27***	0.89	1,470	11.0
4	-47,429.13	94,954.27	95,314.23	0.808	2,380.74***	2,405.80***	0.85	603	4.5
5	-46,197.51	92,511.03	92,945.98	0.806	2,437.59***	2,463.24***	0.82	666	5.0
6	-45,344.50	90,825.00	91,334.95	0.820	1,688.25	1,706.03***	0.80	506	3.8
7	-44,689.24	89,534.49	90,119.42	0.795	1,296.86*	1,310.51***	0.77	486	3.6
8	-44,105.34	88,386.67	89,046.60	0.811	1,155.65	1,167.81***	0.77	389	2.9
9	-43,624.64	87,445.29	88,180.21	0.803	951.37	961.38***	0.74	264	2.0
Boys
1	-71,714.42	143,464.84	143,600.01	-	-	-	-	-	-
2	-60,098.68	120,253.36	120,463.63	0.829	22,989.72***	23,231.48***	0.93	4,583	34.0
3	-56,847.44	113,770.89	114,056.25	0.795	6,434.81***	6,502.48***	0.88	2,024	15.0
4	-55,765.63	111,627.27	111,987.73	0.798	2,141.10*	2,163.62***	0.81	1,363	10.1
5	-54,733.85	109,583.71	110,019.26	0.795	2,042.09**	2,063.56***	0.80	936	6.9
6	-53,920.18	107,976.37	108,487.02	0.787	1,610.41	1,627.34***	0.80	600	4.4
7	-53,240.04	106,636.07	107,221.82	0.780	1,346.14	1,360.30***	0.76	452	3.4
8	-52,619.87	105,415.73	106,076.58	0.796	1,227.43	1,240.34***	0.76	327	2.4
9	-52,198.94	104,593.88	105,329.81	0.759	833.10*	841.86***	0.71	259	1.9

AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion; LMR-LRT = Lo-Mendell-Rubin Adjusted Likelihood Ratio Test; BLRT = Bootstrapped Likelihood Ratio Test.

***

p < .001; **p < .010; *p < .050

Mean values of the nine constructs for children in each of the five profiles are displayed in Figure 1 (and detailed in Supplementary Table S3). Profile patterns were sufficiently comparable for girls and boys to support consistent naming, albeit cross-tabulation analyses revealed small but significantly different prevalence distributions by sex for each of the five profiles. Prevalences for girls and boys in each profile were: complete mental health (complete-MH: 44.3% girls; 41.6% boys); average mental health (average-MH: 30.2%; 32.9%); externalising symptoms with poor self-management (externalising-SM: 11.7%; 8.7%); internalising symptoms with poor relationship skills (internalising-RS: 8.8%; 6.9%); and low mental health (low-MH: 5.0%; 9.9%). Despite the overall similarities in profile patterns between girls and boys, some distinctions were noteworthy. Among children in the externalising-SM profile, boys had above-average relationship skills whereas girls reported average relationship skills. Girls in the low-MH profile demonstrated generalised psychopathology, with the highest scores on internalising and externalising symptoms of any profile, whereas boys in this profile demonstrated lower levels of internalising symptoms than boys in the internalising-RS profile, and were instead characterised by pronounced externalising symptoms.

Figure 1.

Latent profile analysis solutions evidencing five profiles for girls and boys.

Profiles’ associations with adolescent mental health service presentations

Table 4 presents the distribution of children within each profile according to adolescent health services presentations with mental disorder diagnoses (aged ~12–17 years). Table 5 displays the odds of children in each profile presenting with mental disorders, each relative to average-MH. For the most part, children from the low-MH, externalising-SM and internalising-RS profiles had significantly greater odds of mental disorders across all four outcomes. After adjusting for sociodemographic covariates, children in these three risk profiles were 1.7 to 3.3 times as likely to have any mental disorder in adolescence relative to average-MH, representing moderate to large effects. For the most part, children with low-MH demonstrated the largest odds of specific mental disorders in adolescence (adjusted odds ratios [aOR]: girls = 3.4–4.9, boys = 1.9–3.3). Children in the externalising-SM and internalising-RS profiles demonstrated elevated odds of all outcomes, with externalising-SM having greatest odds of externalising disorders (aOR: girls = 2.5, boys = 2.2) and internalising-RS having greatest odds of internalising disorders (aOR: girls = 2.4, boys = 2.0) and self-harm/suicidal ideation (aOR: girls = 2.6, boys = 1.9). No significant difference was found in the odds of self-harm/suicidal ideation between boys in the externalising-SM and average-MH profiles. Children with complete-MH had significantly reduced odds of all four outcomes relative to average-MH (aOR = 0.4–0.9), with two exceptions: for girls, no significant differences between complete-MH and average-MH were observed in odds of externalising disorders nor self-harm/suicidal ideation.

Table 4.

Mental health presentations post-2015 (ages ~12–17 years) according to mental health profiles in middle childhood.

	Sample proportion		No presentations		Any mental health presentation		Externalising disorders		Internalising disorders		Self-Harm/Suicidal Ideation
	%	(n)	%	(n)	%	(n)	%	(n)	%	(n)	%	(n)
Girls
Total Sample (n=13,349)			84.9	(11,330)	10.3	(1,371)	1.7	(228)	5.8	(770)	3.4	(456)
Low Mental Health	5.0	(666)	65.7	(431)	23.9	(159)	4.1	(27)	14.9	(99)	10.7	(71)
Externalising symptoms with poor Self-Management	11.7	(1,565)	78.3	(1,225)	14.5	(227)	3.5	(54)	8.6	(135)	5.4	(85)
Internalising symptoms with poor Relationship Skills	8.8	(1,171)	76.0	(890)	16.7	(196)	2.1	(25)	10.3	(121)	6.1	(71)
Average Mental Health	30.2	(4,037)	86.5	(3,492)	8.7	(352)	1.4	(58)	4.8	(193)	2.6	(105)
Complete Mental Health	44.3	(5,910)	89.5	(5,292)	7.4	(437)	1.1	(64)	3.8	(222)	2.1	(124)
Boys
Total Sample (n=13,488)			84.8	(11,433)	5.9	(800)	1.3	(148)	2.6	(344)	1.2	(163)
Low Mental Health	9.9	(1,336)	70.5	(942)	10.3	(138)	2.6	(35)	4.0	(54)	2.6	(35)
Externalising symptoms with poor Self-Management	8.7	(1,172)	77.6	(910)	10.0	(116)	2.0	(23)	4.2	(49)	1.5	(18)
Internalising symptoms with poor Relationship Skills	6.9	(936)	78.2	(732)	8.4	(79)	1.8	(17)	4.5	(42)	2.1	(20)
Average Mental Health	32.9	(4,436)	85.4	(3,790)	5.6	(250)	0.9	(41)	2.5	(113)	1.3	(57)
Complete Mental Health	41.6	( 5,608)	90.2	(5,059)	3.9	(217)	0.6	(32)	1.5	(86)	0.6	(33)

Excludes children with mental health service presentations prior to 01-Jan-2016 (n = 1,903).

Table 5.

Logistic regression results for mental health presentations according to mental health profile.

		Any mental presentation	Externalising disorders	Internalising disorders	Self-harm/Suicidal ideation
Girls
Low Mental Health	uOR [95% CIs]	3.66 [2.96, 4.53]	3.77 [2.36, 6.02]	4.16 [3.20, 5.40]	5.48 [3.99, 7.53]
Low Mental Health	aOR [95% CIs]	3.26 [2.63, 4.05]	3.36 [2.09, 5.39]	3.60 [2.75, 4.70]	4.91 [3.56, 6.78]
Externalising symptoms with poor Self-Management	uOR [95% CIs]	1.84 [1.54, 2.20]	2.65 [1.82, 3.87]	1.99 [1.59, 2.51]	2.31 [1.72, 3.09]
Externalising symptoms with poor Self-Management	aOR [95% CIs]	1.72 [1.44, 2.07]	2.46 [1.69, 3.60]	1.83 [1.45, 2.31]	2.15 [1.60, 2.88]
Internalising symptoms with poor Relationship Skills	uOR [95% CIs]	2.19 [1.81, 2.64]	1.69 [1.05, 2.72]	2.46 [1.94, 3.12]	2.65 [1.95, 3.62]
Internalising symptoms with poor Relationship Skills	aOR [95% CIs]	2.11 [1.74, 2.55]	1.63 [1.01, 2.63]	2.36 [1.86, 3.01]	2.56 [1.87, 3.50]
Average Mental Health*
Complete Mental Health	uOR [95% CIs]	0.82 [0.71, 0.95]	0.73 [0.51, 1.04]	0.76 [0.62, 0.93]	0.78 [0.60, 1.01]
Complete Mental Health	aOR [95% CIs]	0.85 [0.73, 0.98]	0.75 [0.52, 1.07]	0.79 [0.64, 0.96]	0.80 [0.62, 1.04]
Boys
Low Mental Health	uOR [95% CIs]	2.22 [1.78, 2.77]	3.44 [2.18, 5.42]	1.92 [1.38, 2.68]	2.47 [1.61, 3.79]
Low Mental Health	aOR [95% CIs]	2.13 [1.70, 2.66]	3.26 [2.06, 5.16]	1.89 [1.35, 2.64]	2.46 [1.60, 3.78]
Externalising symptoms with poor Self-Management	uOR [95% CIs]	1.93 [1.53, 2.44]	2.34 [1.40, 3.91]	1.81 [1.28, 2.55]	1.32 [0.77, 2.25]
Externalising symptoms with poor Self-Management	aOR [95% CIs]	1.87 [1.48, 2.37]	2.20 [1.31, 3.70]	1.78 [1.26, 2.51]	1.31 [0.77, 2.24]
Internalising symptoms with poor Relationship Skills	uOR [95% CIs]	1.64 [1.26, 2.13]	2.15 [1.21, 3.80]	1.92 [1.34, 2.77]	1.82 [1.09, 3.04]
Internalising symptoms with poor Relationship Skills	aOR [95% CIs]	1.66 [1.27, 2.17]	2.18 [1.23, 3.86]	1.96 [1.36, 2.82]	1.85 [1.10, 3.09]
Average Mental Health*
Complete Mental Health	uOR [95% CIs]	0.65 [0.54, 0.78]	0.59 [0.37, 0.93]	0.57 [0.43, 0.76]	0.43 [0.28, 0.67]
Complete Mental Health	aOR [95% CIs]	0.68 [0.56, 0.82]	0.61 [0.38, 0.97]	0.59 [0.44, 0.78]	0.44 [0.29, 0.68]

Reference group; Bolded text = p < .05; non-bolded text = p > .05; CI = confidence intervals; uOR = unadjusted odds ratios; aOR = adjusted odds ratios accounting for Aboriginal and Torres Strait Islander background (No [coded 0] and Yes [1]), Socio-Economic Indexes for Areas Index of Relative Socio-Economic Disadvantage quintile (bottom quintiles of disadvantage [0], top quintile of disadvantage [1]), and Accessibility and Remoteness Index of Australia geographical region (major cities [0], inner and outer regional, remote, or very remote [1]).

Discussion

Latent profile analyses of middle childhood (age ~11 years) social-emotional competencies and difficulties yielded a five-profile solution for girls and boys, and these profiles differentially related to health service presentations for mental disorders during adolescence (ages ~12–17 years). Between 42–44% of girls and boys were classified as demonstrating complete mental health (high competencies, low psychopathology). Another 30–33% of children demonstrated average mental health, while the remaining 25–26% of children demonstrated profiles of low mental health (girls = 5%, boys = 10%), externalising symptoms with poor Self-Management (girls = 12%, boys = 9%) or internalising symptoms with poor Relationship Skills (girls = 9%, boys = 7%). Children in these symptomatic and low mental health profiles demonstrated between a 1.7- to 3.3-times greater odds of mental disorders across the adolescent period, relative to average-MH. That is, these differentiated profiles of competencies and difficulties distinguished children at risk of presentations with any, and with particular, adolescent mental disorder diagnoses, providing information that may help the tailoring of tiered interventions for children according to their co-presentation of mental health and illness.

This comprehensive exploration of patterns of middle childhood interpersonal and intrapersonal competencies and internalising and externalising psychopathology identified an average/moderate mental health profile rather than a ‘vulnerable’ profile (low psychopathology, low wellbeing), as reported by others (Gregory et al., 2024; Kim et al., 2018; Moore et al., 2019). Similarly to the prior study that assessed both internalising and externalising psychopathology (Petersen et al., 2020), two distinct profiles emerged separating children according to their experiences of these symptoms. These externalising-SM and internalising-RS profiles align somewhat with the theoretically proposed ‘symptomatic but content’ class, defined by high psychopathology symptoms and high wellbeing (Greenspoon and Saklofske, 2001; Lyons et al., 2013). Excepting self-management, children in the externalising-SM profile demonstrated average (girls) to above average (boys) levels of all other social-emotional competencies. Similarly, children in the internalising-RS profile reported average levels of all competencies, excepting relationship skills. These profiles indicate that social-emotional competence can exist despite psychopathology symptoms, excepting perhaps selected competencies closely aligned with their respective indicators of difficulties (i.e. self-management with conduct problems, and relationships skills with peer relationship problems). These competencies denote additional protective factors that might be bolstered to mitigate against psychopathology symptoms. School-based promotion of these skills requires the provision of universal social-emotional learning programmes targeting all five competencies (Carpendale et al., 2025), but where indicated, this whole-school provision should be supplemented by targeted and intensive interventions delivered to students requiring additional support, such as via small group-based activities or individualised support plans delivered to children at risk of or experiencing mental health difficulties (Close et al., 2023). Tailoring of these higher-tier interventions should be informed by comprehensive assessment of a student’s social-emotional competencies alongside their difficulties, so that they respond to specific needs (e.g. relationships skills promotion for internalising-RS children).

This study demonstrated longitudinal associations between middle childhood mental health profiles and adolescent mental disorders. The complete-MH profile was, for the most part, protective against later disorders relative to average-MH. In a prior study (Moore et al., 2019), ‘symptomatic but content’ and ‘troubled’ profiles were later characterised by greater depression and anxiety symptoms than ‘moderate’ and ‘complete mental health’ profiles. Children classified here into the symptomatic (externalising-SM; internalising-RS) or low-MH profiles similarly had significantly greater odds of any, internalising, and externalising disorder presentations, relative to the average-MH profile, even when adjusting for sociodemographic covariates. Findings are also consistent with research indicating homotypic and heterotypic continuity between internalising and externalising disorders (Copeland et al., 2013; Shevlin et al., 2017). That is, children with pronounced externalising symptoms (low-MH; externalising-SM) and internalising symptoms (low-MH; internalising-RS) demonstrated significantly greater odds (cf. average-MH) of both externalising and internalising disorders. Whereas odds of self-harm/suicidal ideation among girls were significantly higher for all three risk profiles relative to average-MH, this was not replicated among boys in the externalising-SM profile. This may reflect the paradox by which males are less likely to engage in self-harm than females (Diggins et al., 2024), despite being more likely to die by suicide (Barrigon and Cegla-Schvartzman, 2020). Alternatively, the greater relationship skills of externalising-SM boys, relative to girls in the equivalent profile, may provide the foundation for stronger relationships with family and peers, and act as a protective factor.

A strength of this study was the use of a large, representative sample of Australian young people followed over time, but several limitations are notable. Non-significant differences in odds of mental disorders could be a result of the small number of individuals with mental disorders among some of the profiles (e.g. 1.5% of externalising-SM boys with self-harm/suicidal ideation presentations and 1.1% of complete-MH girls with externalising disorders), resulting in wide confidence intervals that broach 1.0 (Reeve, 2018). Data was limited to hospital and ambulatory service presentations for mental disorders and does not account for presentations to other types of health services, such as general practitioners, private psychologists or psychiatrists or counsellors. Children who had diagnoses recorded prior to/in the year of the Middle Childhood Survey (2015; Laurens et al., 2017) were removed from logistic regression analyses (n = 1,903), as we sought to explore the association of middle childhood profiles with later mental health service use among children who had not previously accessed such services. Thus, this study was not equipped to explore associations between these profiles and neurodevelopmental disorders, for which peak onset is in early childhood (Solmi et al., 2022). Finally, this study employed brief measures of social-emotional competencies and difficulties (3–5 items per scale), limiting coverage of these domains. Replication using more comprehensive measures would ascertain the consistency of profile solutions.

Future longitudinal research could explore potential predictors of membership in each of the profiles, such as early childhood mental disorders, parental mental illness or adverse childhood experiences. Continued follow-up of the sample into early adulthood would capture more mental health presentations, particularly for disorders that peak in onset during adulthood (e.g. substance use, mood and personality disorders and schizophrenia; Solmi et al., 2022). Where this study was limited to self-report assessment of children’s perceived social-emotional competencies and difficulties at age ~11 years, future research could determine whether these profiles are replicated among different age groups and using teacher- and/or parent-report data.

This study demonstrates the utility of middle childhood mental health profiles that account for both social-emotional competencies and difficulties in indexing risk for mental disorders in adolescence (ages ~12–17 years). Middle childhood represents an opportune time for prevention and early intervention, prior to the difficult transition to adolescence – the peak period of onset for many mental disorders (McGrath et al., 2023; Solmi et al., 2022). Social-emotional competencies are teachable and can be promoted in schools via effective, evidence-based programming (Carpendale et al., 2025; Durlak et al., 2022). Further research is required to determine the effectiveness of these programmes for averting mental health difficulties and to inform where the provision of targeted and intensive supports – tailored to students’ social-emotional needs and emerging/established psychopathology – may be required in addition to the whole-school teaching of all competencies.

Supplemental Material

sj-docx-1-anp-10.1177_00048674261426776 – Supplemental material for Middle childhood profiles of social-emotional competencies and difficulties differentiate risk of health service presentations with adolescent mental disorders

Supplemental material, sj-docx-1-anp-10.1177_00048674261426776 for Middle childhood profiles of social-emotional competencies and difficulties differentiate risk of health service presentations with adolescent mental disorders by Emma J Carpendale, Melissa J Green, Oliver J Watkeys, Stacy Tzoumakis, Vaughan J Carr and Kristin R Laurens in Australian & New Zealand Journal of Psychiatry

Footnotes

Acknowledgements

This research used population data owned and managed by the NSW Ministry of Health, ACT Health, the Australian Coordinating Registry (on behalf of Australian Registries of Births, Deaths and Marriages, Australian Coroners and the National Coronial Information System) and the University of New South Wales. The findings and views reported are those of the authors and should not be attributed to these Departments and Agencies, or the NSW or Australian Governments. Record linkages were conducted by the NSW Centre for Health Record Linkage (CHeReL).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was conducted by the University of New South Wales with financial support from National Health and Medical Research Council Project Grants (APP1058652 and APP1148055), an Australian Research Council Future Fellowship (FT170100294 awarded to KRL) and Discovery Early Career Researcher Award (DE210100113 awarded to ST), a Department of Health and Aged Care Medical Research Future Fund Million Minds Mental Health Grant (APP2006436) and Suicide Prevention Australia (Postdoctoral Fellowship awarded to OJW) with funding provided by the Australian Department of Health, Disability and Ageing.

Ethical approval and informed consent

Ethics approval was provided by the NSW Population and Health Services and ACT Health Human Research Ethics Committees (HREC/18/CIPHS/49).

Consent for a school to participate in the Middle Childhood Survey (MCS) was provided in online (electronic) form by the school Principal (school leader). Child recruitment for the MCS within participating schools followed a child and parent opt-out consent procedure using online forms, written instruction and verbal instruction. The use of opt-out consent procedures (for parents and children) was guided by the Australian National Health and Medical Research Council National Statement of Ethical Conduct in Human Research (chapter 2.3), which specifies conditions under which these procedures are appropriate.

ORCID iDs

Emma J Carpendale

Melissa J Green

Stacy Tzoumakis

Vaughan J Carr

Kristin R Laurens

Data availability statement

Data used in this project has been provided by government agencies for the research purposes of the NSW Child Development Study and are unable to be shared with third parties or deposited into data repositories. Collaborative research activities may be possible depending on scope and resources; alternatively, data can be accessed by applications to relevant data custodians.

Supplemental material

Supplemental material for this article is available online.

References

Achenbach

Rescorla

(2001) Manual for the ASEBA School-Age Forms and Profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth, and Families.

Australian Bureau of Statistics (2011) Census of population and housing: Socio-Economic Indexes for Areas (SEIFA), 2011. Available at: https://www.ausstats.abs.gov.au/ausstats/subscriber.nsf/LookupAttach/2033.0.55.001Publication28.03.131/$File/2033.0.55.001%20SEIFA%202011%20Technical%20Paper.pdf (accessed 25 November, 2025).

Barrigon

Cegla-Schvartzman

(2020) Sex, gender, and suicidal behavior. In: Baca-Garcia

(ed.) Behavioral Neurobiology of Suicide and Self Harm (Current Topics in Behavioral Neurosciences, vol. 46). Cham: Springer, pp. 89–115.

Brazel

Allison

Bastiampillai

, et al. (2023) Child and adolescent mental health services in Australia: A descriptive analysis between 2015–16 and 2019–20. Australasian Psychiatry 31: 445–451.

Carpendale

Green

Dix

, et al. (2025) An exploratory evaluation of universal social-emotional learning programs delivered during elementary school to Australian students. Journal of School Psychology 110: 101447.

Carpendale

Green

Williams

, et al. (2023a) Item response theory analysis of self-reported social-emotional learning competencies in an Australian population cohort aged 11 years. School Psychology 38: 247–263.

Carpendale

Green

Williams

, et al. (2023b) Measurement invariance across sex and language background of a self-report social-emotional learning questionnaire for middle childhood. Child Indicators Research 32: e1962.

Clark

Malecki

(2022) Adolescent mental health profiles through a latent dual-factor approach. Journal of School Psychology 91: 112–128.

Clark

Blyth

Plant

, et al. (2024) Dual-factor mental health in adolescence: Comparing classification methods. Journal of School Psychology 107: 101391.

10.

Clark

Muthén

(2009) Relating Latent Class Analysis Results to Variables Not Included in the Analysis. Los Angeles, CA: University of California.

11.

Killingly

Gaumer-Erickson

, et al. (2023) What is Social and Emotional Learning (SEL) and why is it important? In: Graham

(ed.) Inclusive Education for the 21st Century. London: Routledge, pp. 275–301.

12.

Collaborative for Academic, Social and Emotional Learning [CASEL] (2025) Fundamentals of SEL. Chicago, IL: Collaborative for Academic Social and Emotional Learning.

13.

Collaborative for Academic, Social and Emotional Learning [CASEL] (2005) Safe and Sound: An Educational Leader’s Guide to Evidence-Based Social and Emotional Learning (SEL) Programs. Chicago, IL: Collaborative for Academic Social and Emotional Learning.

14.

Collins

Lanza

(2009) Latent Class and Latent Transition Analysis: With Applications in the Social, Behavioral, and Health Sciences. Hoboken, NJ: John Wiley & Sons.

15.

Copeland

Adair

Smetanin

, et al. (2013) Diagnostic transitions from childhood to adolescence to early adulthood. Journal of Child Psychology and Psychiatry 54: 791–799.

16.

de Bruin

Killingly

Graham

(2023) Multi-tiered systems of support: What are they, where did they come from, and what are the lessons for Australian schools? In: Graham

(ed.) Inclusive Education for the 21st Century. London: Routledge, pp. 157–175.

17.

Deighton

Tymms

Vostanis

, et al. (2013) The development of a school-based measure of child mental health. Journal of Psychoeducational Assessment 31: 247–257.

18.

Department of Health and Aged Care [Australian Government] (2001) Measuring Remoteness: Accessibility/Remoteness Index of Australia (ARIA) (Occasional papers: New series no.14). Canberra, ACT, Australia: Department of Health and Aged Care.

19.

Diggins

Heuvelman

Pujades-Rodriguez

, et al. (2024) Exploring gender differences in risk factors for self-harm in adolescents using data from the millennium cohort study. Journal of Affective Disorders 345: 131–140.

20.

Durlak

Mahoney

Boyle

(2022) What we know, and what we need to find out about universal, school-based social and emotional learning programs for children and adolescents: A review of meta-analyses and directions for future research. Psychological Bulletin 148: 765.

21.

Eaton

Keyes

Krueger

, et al. (2012) An invariant dimensional liability model of gender differences in mental disorder prevalence: Evidence from a national sample. Journal of Abnormal Psychology 121: 282–288.

22.

Furlong

You

Renshaw

, et al. (2013) Preliminary development of the positive experiences at school scale for elementary school children. Child Indicators Research 6: 753–775.

23.

Gadermann

Schonert-Reichl

Zumbo

(2010) Investigating validity evidence of the satisfaction with life scale adapted for children. Social Indicators Research 96: 229–247.

24.

Goodman

(2001) Psychometric properties of the strengths and difficulties questionnaire. Journal of the American Academy of Child & Adolescent Psychiatry 40: 1337–1345.

25.

Green

Watkeys

Harris

, et al. (2024) Cohort profile: The New South Wales Child Development Study (NSW-CDS) – wave 3 (Child age ~18 years). International Journal of Epidemiology 53: dyae069.

26.

Greenberg

Domitrovich

Weissberg

, et al. (2017) Social and emotional learning as a public health approach to education. The Future of Children 27: 13–32.

27.

Greenspoon

Saklofske

(2001) Toward an integration of subjective well-being and psychopathology. Social Indicators Research 54: 81–108.

28.

Gregory

Lewkowicz

Engelhardt

, et al. (2021) Data resource profile: The South Australian wellbeing and engagement collection (WEC). International Journal of Epidemiology 51: 16–16.

29.

Gregory

Monroy

Grace

, et al. (2024) Mental health profiles and academic achievement in Australian school students. Journal of School Psychology 103: 101291.

30.

Grové

Laletas

(2020) Promoting student wellbeing and mental health through social and emotional learning. In: Graham

(ed.) Inclusive Education for the 21st Century. London: Routledge, pp. 317–335.

31.

Guest

Ross

Childs

, et al. (2024) Embedding social emotional learning from the bottom up in multi-tiered services and supports frameworks. Psychology in the Schools 61: 2745–2761.

32.

Huebner

(1991) Initial development of the student’s life satisfaction scale. School Psychology International 12: 231–240.

33.

Humphrey

(2023) Universal, school-based social and emotional learning interventions and their potential to improve students’ mental health. In: Evans

Owens

Bradshaw

, et al (eds) Handbook of School Mental Health (Issues in Clinical Child Psychology). Cham: Springer, pp. 9–19.

34.

Iasiello

Van Agteren

(2020) Mental health and/or mental illness: A scoping review of the evidence and implications of the dual-factor model of mental health. Evidence Base: A Journal of Evidence Reviews in Key Policy Areas 2020: 1–45.

35.

Islam

Khanam

Kabir

(2020) The use of mental health services by Australian adolescents with mental disorders and suicidality: Findings from a nationwide cross-sectional survey. PLoS One 15: e0231180.

36.

Kern

Benson

Steinberg

, et al. (2015) The EPOCH measure of adolescent well-being. Psychological Assessment 28: 586–597.

37.

Kim

Dowdy

Furlong

, et al. (2018) Complete mental health screening: Psychological strengths and life satisfaction in Korean students. Child Indicators Research 12: 901–915.

38.

Kusche

Greenberg

Beilke

(1988) Seattle Personality Questionnaire for Young School-Aged Children. Unpublished Personality Questionnaire. Seattle, WA: University of Washington, Department of Psychology.

39.

Laurens

Tzoumakis

Dean

, et al. (2017) The 2015 Middle Childhood Survey (MCS) of mental health and well-being at age 11 years in an Australian population cohort. BMJ Open 7: e016244.

40.

Laurent

Catanzaro

Joiner

Jr , et al. (1999) A measure of positive and negative affect for children: Scale development and preliminary validation. Psychological Assessment 11: 326–338.

41.

Lee

Skeen

Melendez-Torres

, et al. (2023) School-based socio-emotional learning programs to prevent depression, anxiety and suicide among adolescents: A global cost-effectiveness analysis. Epidemiology and Psychiatric Sciences 32: e461–11.

42.

Lyons

Huebner

Hills

(2013) The dual-factor model of mental health: A short-term longitudinal study of school-related outcomes. Social Indicators Research 114: 549–565.

43.

McGrath

Al-Hamzawi

Alonso

, et al. (2023) Age of onset and cumulative risk of mental disorders: A cross-national analysis of population surveys from 29 countries. The Lancet Psychiatry 10: 668–681.

44.

Moore

Dowdy

Nylund-Gibson

, et al. (2019) An empirical approach to complete mental health classification in adolescents. School Mental Health 11: 438–453.

45.

Muthén

(1998) Mplus User’s Guide, 7th edition. Los Angeles, CA: Muthén & Muthén.

46.

National Health and Medical Research Council (2007) The national statement on ethical conduct in human research. Available at: www.nhmrc.gov.au/about-us/publications/national-statement-ethical-conduct-human-research-2007-updated-2018#block-views-block-file-attachments-content-block-1 (accessed 25 November 2025).

47.

Nylund-Gibson

Choi

(2018) Ten frequently asked questions about latent class analysis. Translational Issues in Psychological Science 4: 440–461.

48.

Petersen

Humphrey

Qualter

(2020) Latent class analysis of mental health in middle childhood: Evidence for the dual-factor model. School Mental Health 12: 786–800.

49.

Ravens-Sieberer

Auquier

Erhart

, et al. (2007) The KIDSCREEN-27 quality of life measure for children and adolescents: Psychometric results from a cross-cultural survey in 13 European countries. Quality of Life Research 16: 1347–1356.

50.

Reeve

(2018) Confidence intervals of difference of proportions in logistic regression in presence of covariates. Statistical Methods in Medical Research 27: 451–465.

51.

Rosenthal

(1996) Qualitative descriptors of strength of association and effect size. Journal of Social Service Research 21: 37–59.

52.

Sacco

Camilleri

Eberhardt

, et al. (2024) A systematic review and meta-analysis on the prevalence of mental disorders among children and adolescents in Europe. European Child & Adolescent Psychiatry 33: 2877–2894.

53.

Schonert-Reichl

Guhn

Gadermann

, et al. (2013) Development and validation of the middle years development instrument (MDI): Assessing children’s well-being and assets across multiple contexts. Social Indicators Research 114: 345–369.

54.

Shevlin

McElroy

Murphy

(2017) Homotypic and heterotypic psychopathological continuity: A child cohort study. Social Psychiatry and Psychiatric Epidemiology 52: 1135–1145.

55.

Slade

Chapman

Halladay

, et al. (2024) Diverging trends in alcohol use and mental health in Australian adolescents: A cross-cohort comparison of trends in co-occurrence. JCPP Advances 4: e12241.

56.

Solmi

Radua

Olivola

, et al. (2022) Age at onset of mental disorders worldwide: Large-scale meta-analysis of 192 epidemiological studies. Molecular Psychiatry 27: 281–295.

57.

Suldo

Shaffer

(2008) Looking beyond psychopathology: The dual-factor model of mental health in youth. School Psychology Review 37: 52–68.

58.

Tein

Coxe

Cham

(2013) Statistical power to detect the correct number of classes in latent profile analysis. Structural Equation Modeling: A Multidisciplinary Journal 20: 640–657.

59.

Thomson

Richardson

Gadermann

, et al. (2019) Association of childhood social-emotional functioning profiles at school entry with early-onset mental health conditions. JAMA Network Open 2: e186694.

60.

Tkacz

Brady

(2021) Increasing rate of diagnosed childhood mental illness in the United States: Incidence, prevalence and costs. Public Health in Practice 2: 100204.

61.

Vermunt

Magidson

(2002) Latent class cluster analysis. In: Hagenaars

McCutcheon

(eds) Applied Latent Class Analysis, vol. 11. Cambridge: Cambridge University Press, pp. 89–106.

62.

Watkeys

O’Hare

Dean

, et al. (2024) Patterns of health service use for children with mental disorders in an Australian state population cohort. Australian and New Zealand Journal of Psychiatry 58: 857–874.

63.

Weller

Bowen

Faubert

(2020) Latent class analysis: A guide to best practice. Journal of Black Psychology 46: 287–311.

64.

World Health Organisation (2025) Mental Health. Available at: https://www.who.int/news-room/fact-sheets/detail/mental-health-strengthening-our-response (accessed 25 November 2025).

65.

You

Furlong

Dowdy

, et al. (2014) Further validation of the social and emotional health survey for high school students. Applied Quality of Life Research 9: 997–1015.