Background Until now, child mental health promotion efforts have focused primarily on reducing the prevalence and severity of problems; yet the absence of mental health problems does not necessarily imply the presence of healthy psychosocial functioning. We aimed to investigate the epidemiology of child mental health competence in a full national population of school entrants.
Methods The data source was the 2012 Australian Early Development Index, a national census of early childhood development completed for school entrants by teachers across Australia (n=275 800). The mental health competence outcome measure was derived from constructs that focused on children's social and emotional strengths. Children with mental health competence scores in the top quintile were compared with the standard population across individual and community characteristics.
Results Average age at assessment was 5 years 7 months. Higher odds of mental health competence were observed for children who lived in more advantaged areas (OR 1.62; 99% CI 1.49 to 1.75), had attended preschool (1.38; 1.25 to 1.51) and demonstrated effective oral communication skills in the classroom (19.01; 15.62 to 23.13). Indigenous children had lower odds compared with non-Indigenous children (0.59; 0.54 to 0.64). Children in disadvantaged areas who attended preschool did not ‘catch up’ with their more advantaged peers.
Conclusions Mental health competence is unequally distributed across the Australian child population at school entry and is strongly predicted by measures and correlates of disadvantage. Effective oral communication and attendance at preschool warrant further investigation as potentially modifiable factors that may support mental health competence in new school entrants.
- CHILD HEALTH
- MENTAL HEALTH
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The WHO states, ‘there can be no health without mental health’,1 acknowledging the growing body of evidence demonstrating the close interdependence of physical and mental well-being.2 Mental health in childhood has important implications for lifecourse development, including future emotional health, physical well-being and work productivity.3 ,4 Thus, promoting child mental health, already recognised as an important public health goal,5 ,6 may have far-reaching benefits that extend beyond the mental health sphere, supporting children's capacity to grow into healthy adults leading fulfilling lives and contributing socially and economically to society.7
Despite this rhetoric, the success of child mental health promotion efforts has largely been reported in terms of the degree to which they reduce the prevalence or severity of mental health disorders and related problems such as youth crime.8 ,9 The assumption implicit in this approach is that good mental health in children equates to the absence of mental health disorder. This is in direct contradiction of the WHO principle that health is more than merely the absence of disease.10
Instead, empirical evidence11 strongly supports the dual continuum model,12 in which positive mental health and mental health disorder are understood to be correlated to a degree, but are nevertheless distinct aspects of health.13 Adoption of the dual continuum model in public health terms requires robust constructs and measures of positive child mental health to complement those that measure mental disorder.14 ,15 One such construct is mental health competence which refers to ‘adaptational success in the developmental tasks expected of individuals of a given age in a particular cultural and historical context’.16
National and international bodies have recognised the potential of intervening at the population level to modify positive mental health,17 ,18 noting that ‘there is now compelling evidence for the need to promote positive mental health through interventions that promote competence and psychological strengths’.6 This requires a new public health approach with evidence that focuses on children's mental health competence and its determinants.14
Our aim in this investigation was to describe the epidemiology of positive mental health in a full national cohort of children as they started school. We hypothesised (1) that mental health competence would be unequally distributed across child characteristics, in particular measures of socioeconomic disadvantage; and (2) that attendance at preschool might narrow the gap between advantaged and disadvantaged children. We focused particularly on identifying modifiable factors and social determinants that may present opportunities for population-level policies and interventions.
Data sources and sample selection
The data source for this study was the 2012 Australian Early Development Index (AEDI),19 ,20 a national census of children's development in their first year of full-time schooling. The 2015 census and future iterations will be known as the Australian Early Development Census (AEDC). The AEDI/AEDC is an Australian adaptation of the Canadian Early Development Instrument.20 Teachers report on every child in their class using a structured survey with five developmental domains: physical health and well-being; social competence; emotional maturity; language and cognitive development; and communication skills and general knowledge. Approval for implementation of the AEDI was granted by the Royal Children's Hospital Human Research Ethics Committee (RCH HREC 24051).
The AEDI was completed for 289 973 children in 2012, capturing 96.5% of the estimated 5-year-old population of Australia in that year.19 Data were collected in the middle of the Australian school year. Owing to concerns about mixing exposures and outcomes, children whom the teacher had identified as having recognised special needs at school based on a pre-existing medical diagnosis21 were excluded from the current analysis, as these diagnoses included difficulties in social and emotional development, behaviour and language. The final size of the study population was 275 800 children.
Outcome measure: mental health competence
The competence measure for this study was derived from existing data in the AEDI and we have previously described the development of this measure,15 which is designed to report child strengths, not just the absence of difficulties. It incorporates five subdomains (32 data items in total) from the social competence and emotional maturity domains of the AEDI that reflect key positive mental health constructs: overall social competence; responsibility and respect; approaches to learning; readiness to explore new things; and prosocial and helping behaviour (table 1). These subdomains were selected to operationalise positive mental health as a multidimensional measure of developmental tasks.16 The measure was validated using the Strengths and Difficulties Questionnaire (SDQ).15 There was a positive association with the prosociality subscale (OR 1.21; 95% CI 1.05 to 1.40) and an inverse relationship with the peer problems subscale (OR 0.68; 95% CI 0.56 to 0.83). The measure also shows good predictive validity. For example, in a subsample of the Longitudinal Study of Australian Children B Cohort with linked AEDC data (N=2271), the mental health competence measure had a strongly positive association with the prosocial subscale of the SDQ at 8–9 years (regression coefficient=0.45; 95% CI 0.29 to 0.62) and a strongly inverse relationship with the peer problems subscale (regression coefficient=−0.34; 95% CI −0.49 to −0.19; data not shown).
For this investigation, the outcome of interest was optimal positive mental health. Children with a competence score equal to or above the 80th centile for the cohort were regarded as demonstrating a high level of competence.
Demographic characteristics: Age, gender, language background and Indigenous (Aboriginal or Torres Straits Islander) status are recorded in the AEDI for all children. Children were categorised as bilingual if their teacher indicated that they spoke a language other than English at home and/or were considered to have English as a second language. Indigenous status was included in this study to report outcomes in this historically and currently disadvantaged subpopulation. Teachers are instructed to base their report of Indigenous status on the school enrolment record. For 35% of the Indigenous children in this study, assessments were made jointly by the teacher and a cultural consultant (eg, an Aboriginal education worker).
Oral communication in the classroom: Teachers rated oral communication in English (the language of instruction in Australian schools) according to their assessment of children's ‘ability to use language effectively in English’. They were instructed to base their answer on children's ‘effective use of appropriate words and expressions at appropriate times, as well as the contribution to conversations’. ‘Effective use’ was defined as the sufficient conveyance of a desired message. Oral communication in the classroom (OCC) was rated on a three-point Likert scale (very good/good, average or poor/very poor; referred to henceforth as good, average and poor, respectively).
Non-parental care: Information about non-parental care prior to school entry is routinely recorded by Australian schools at enrolment, and is thus available to teachers completing the AEDI. This includes information about the type, for example, a preschool programme (play-based education attended in the year before compulsory schooling)22 or daycare centre. These categories are not mutually exclusive because children frequently attend a mix of care types. For this analysis, the exposure of interest was the potential effect of preschool on children's mental health; accordingly, children were categorised as having exposure to preschool, non-parental care other than preschool, or parental care only.
Community-level socioeconomic position: In the AEDI, the socioeconomic position of the child's local community of residence is assigned using the Australian Bureau of Statistics (ABS) Index of Relative Socioeconomic Disadvantage (Socio-Economic Indexes for Areas (SEIFA)) score, categorised into quintiles. This score reflects characteristics such as income, educational attainment and levels of employment in each area.23 Other standard ABS socioeconomic measures reported in this study were the proportions in the child's local community of labour force unemployed, of individuals who had completed year 12 schooling or equivalent, and of young people who were single parents. In 2012, there was an average of 62 school entrants per local community.
Remoteness: The Australian Standard Geographic Classification24 was used to classify the remoteness of communities where children lived. Remoteness is reported in this study as a categorical measure with three levels: major city, regional and remote.
The frequency distribution of the measured characteristics was estimated in the study population and in the children with missing outcome data. Children with competence scores in the top quintile were compared with the standard population across individual and community characteristics. (Note: associations with age were not reported as the outcome is age adjusted.)
Associations were estimated using logistic regression to allow adjustment for likely confounding by gender, Indigenous status and socioeconomic position; crude and adjusted estimates were then compared.
To explore the potential for attendance at preschool to reduce inequalities in developmental outcomes,25 we stratified the association of preschool attendance and mental health competence by quintiles of socioeconomic position, and created a composite variable from categories of these two variables to compare preschool outcomes for children in advantaged and disadvantaged communities.
Missing data analysis
In the study population, 265 013 children (91.4%) had sufficient data to calculate the outcome measure (mental health competence). We considered whether complete case analysis may have introduced selection bias, and conducted sensitivity analyses to determine the likely bounds of this bias. Selection bias was explored under two potential scenarios: (1) none of the children with missing outcome data had high competence, and (2) all of the children with missing outcome data had high competence.
Most of the independent variables had missing data on fewer than 1% of records, except Early Childhood Education and Care (ECEC) where 8.7% of children had missing data, mostly due to ‘don't know’ responses.
Software and reporting
Descriptive statistics and logistic regression models were generated using Stata/SE V.13.1 for Windows (copyright 1985–2013 StataCorp LP, revision 30 October 2013). ORs refer to exponentiated coefficients obtained in logistic regression analyses. Robust SEs were used, clustered on teachers. Owing to the large number of records analysed, CIs are reported at the 99% limit throughout.
Characteristics of the study population
Mean age at the time of AEDI assessment was 5 years 7 months (table 2). There was an even distribution of gender (50.1% boys/49.9% girls) and 5.1% of the children were identified as Indigenous. The majority (81.2%) had attended preschool in the year before starting school, while very few children (5.1%) had been in the care of their parents only. The children who were excluded from the analyses because of missing data on the competence outcome were more likely to be male, Indigenous or living in a less advantaged local area, and were less likely to be assessed as having good OCC.
Table 3 shows crude and adjusted associations of individual and community characteristics with mental health competence. Adjusting for potential sociodemographic confounders (gender, Indigenous status and area socioeconomic position) did not substantially alter the results, except in the case of remoteness where the OR increased to the null after adjustment.
The strongest association was seen with oral communication: children who were rated as having good OCC had markedly higher odds of high mental health competence than those with a poor rating (OR 19.0; 99% CI 15.6 to 23.1). Indigenous children had lower odds of high mental health competence than non-Indigenous children, and bilingual children had lower odds of competence than monolingual children. We note, however, that Indigenous children with good OCC had lower odds of mental health competence relative to their non-Indigenous counterparts with good OCC (OR 0.75; 99% CI 0.68 to 0.82), while bilingual children with good OCC had similar odds of mental health competence to the monolingual population with good OCC (OR 1.07; 99% CI 1.00 to 1.13; data not shown). Among Indigenous children, the odds of high competence were similar whether or not a cultural consultant was present at the assessment (OR 0.92; 99% CI 0.76 to 1.11).
The odds of high mental health competence increased across the SEIFA quintiles as socioeconomic advantage increased, and children who had attended preschool had higher odds of high mental health competence compared with children who had not (OR 1.38; 99% CI 1.25 to 1.51). When the association of preschool with the outcome was estimated separately for each quintile of area socioeconomic position, it could be seen that within each quintile, children who had attended preschool had higher odds of competence than children who had been in parental care only. However, the ORs of this association were similar whether children lived in disadvantaged or advantaged areas (table 4), suggesting that disadvantaged children did not experience additional benefit from attendance. Among the preschool population, children in advantaged areas still had substantially higher odds of mental health competence than children in disadvantaged areas (OR 1.56; 99% CI 1.43 to 1.70; not shown).
When the logistic regression analysis for the effect of preschool was rerun under assumptions that (1) all or (2) none of the children with missing outcomes had high competence, the adjusted ORs of high competence for children were similar to the observed OR using complete case analysis. Using area disadvantage as an example, when the logistic regression analysis was rerun under an assumption that none of the children with missing outcomes had high competence, the adjusted OR of high competence for children living in the most advantaged area quintile (compared with the least advantaged quintile) was 1.63 (1.51 to 1.77). Under the opposite assumption that all of the children with missing outcomes had high competence, the adjusted OR of high competence for children in this quintile was 1.43 (1.32 to 1.54). These results are comparable to the observed OR using complete case analysis (OR 1.62; 1.49 to 1.75). This implies that missing outcome data were unlikely to have introduced sufficient bias to alter the conclusions of the study.
This is the first study to measure and report positive child mental health across a full national cohort. We found higher odds of mental health competence for children demonstrating effective OCC, girls, non-Indigenous children, children from more advantaged areas and children who had attended preschool. Thus, a range of determinants (some potentially modifiable) predicted inequalities in mental health competence for Australian children as they started school.
At the individual level, the association of competence with OCC in our study was so strong that it raises the question of whether oral communication in the language of instruction is a necessary condition of being assessed as having high competence. This finding is consistent with previous work showing that oral communication and mental health are closely linked,26 but it is not possible in this cross-sectional sample to identify the direction of effect with certainty. If further investigation indicates that good oral communication promotes positive mental health, addressing oral communication, particularly through whole of classroom strategies delivered by teachers, might be an effective point of intervention at a population level. While not possible to differentiate in this study, it is likely that OCC reflects both language development and exposure. Health (eg, speech pathology) and education (eg, classroom-based learning) can have substantive and convergent roles in addressing oral communication27 and therefore mental health competence.
Children living in more advantaged areas experienced better positive mental health. Previous individual-level research shows that maternal education and family socioeconomic advantage predict child competence.15 ,28 Mechanisms that may be operating at the area level include greater social support, a safer community environment, and better access to health services and recreational facilities.29 ,30
Children with exposure to preschool were more likely to demonstrate a high level of competence than children who had been in parental care only. While this association is likely to reflect some self-selection of families into preschool because we could not control for family characteristics,31 attendance at a high-quality ECEC facility has been shown to promote positive outcomes at school entry, including better social and communication skills and better relationships with teachers and peers.32 ,33 Children from disadvantaged areas have the most to gain from ECEC experiences34 but, in our population, attendance per se did not appear to close the disadvantage gap. A recent Australian-based study has shown that in the ECEC setting, the quality of relationships between carers and children strongly modifies the association between income and developmental outcomes,35 suggesting that promoting high-quality relationships may be an effective strategy for closing this gap.
Indigenous children were assessed by teachers as having markedly poorer levels of mental health competence in the school setting than non-Indigenous children, in line with highly concerning gaps in general health outcomes for Indigenous children and their families.36 In our analyses, systematic differences in area disadvantage and in effective OCC, although influential, did not fully account for the inequalities in outcomes between Indigenous and non-Indigenous children. We considered whether the mental health competence of Indigenous children may have been underestimated by teachers who did not share a cultural background with their students, but we found no evidence in our data for this: the OR for competence in Indigenous children with a cultural consultant present was similar to that in assessments of Indigenous children without a cultural consultant.
Australian Indigenous children face severe additional threats to their mental health that were not measured in this study, such as racially based bullying and the intergenerational effects of forced removal of children from their families.37 ,38 Of particular concern is the possibility that if the education environment is not sufficiently responsive to Indigenous linguistic and cultural identities, schools and preschools will not be able to support the positive mental health of young Indigenous children.39 An important next step will be to establish the extent to which the competence measure used in this study reflects values and norms of child mental health held by Indigenous communities.37 ,40
In summary, the study results were consistent with our first hypothesis in that we observed inequalities in mental health competence over a number of measured characteristics. However, the results did not support our second hypothesis, instead suggesting that preschool attendance did not close the disparity gap between advantaged and disadvantaged children.
Strengths and limitations
The 2012 AEDI data set is a full national census of Australian children in their first year of school, and this breadth is a major strength of the study in that it provides an opportunity to report on disadvantaged subpopulations which are numerically small, as for Indigenous children. However, while the large sample size reduces the risk of random error, it does not address errors arising from bias.
Despite the fact that the 2012 AEDI data collection was able to assess over 96% of the estimated target population, missing outcome data in this study were predicted by several characteristics associated with disadvantage, indicating a potential for selection bias. However, sensitivity analyses indicated that the magnitude of any resulting bias was likely to be small.
No data are available in the AEDI about family-level measures of disadvantage, duration of exposure to preschool or preschool quality, as this is beyond the scope of information that is available to teachers. Imprecision in these measures may have resulted in attenuation of the observed relationships or residual confounding. Interpretation of the results could be further refined by additional research examining the relative contribution of family-level and community-level disadvantage to positive mental health outcomes.
Substantial inequalities in the positive mental health of Australian children are already present at school entry. The findings from this study suggest that effective population-level interventions to promote mental health competence may not necessarily be direct mental health interventions; intervening to modify the social determinants of mental health may also be a promising strategy.
Preschool attendance and OCC are two examples of potentially modifiable factors identified by this study which have a strong association with the outcome and which are already established targets for large-scale policy and intervention efforts.22 These results demonstrate the value of undertaking research that actively seeks to explore policy opportunities at the health/education interface. Indeed, they highlight the need for policy and professional thinking that more purposefully incentivises an integrated approach to health and education, particularly where there is mutual benefit and potential for improved outcomes. The results also indicate that closing the disparity gap will require additional effort that takes into account the need for equitable approaches if we are to address the disadvantages experienced by the most vulnerable children.41 ,42
What is already known on this subject
Positive child mental health is recognised as more than just the absence of mental illness. Effective mental health promotion requires a rigorous understanding of the epidemiology of positive mental health.
What this study adds
This study is the first to report positive child mental health in a full national population. Mental health competence in young Australian children is strongly predicted by measures and correlates of disadvantage. A positive association with oral communication skills and attendance at preschool programmes suggests that these should be further investigated as possible approaches for effective population-level promotion of positive mental health.
There are a number of key groups to be acknowledged for their support of the Australian Early Development Index (AEDI): including the Australian Government who funded the study; all schools, principals and teachers across Australia who participated in the AEDI; and each of the State and Territory AEDI Coordinators and their Coordinating Committees who helped to facilitate the AEDI data collection in their respective jurisdictions. The authors appreciate their time and commitment. Personnel support for this analysis was funded by the Australian Government, and was supported by the Victorian Government’s Operational Infrastructure Support Program. SG is supported by the Australian National Health and Medical Research Council (NHMRC) Career Development Fellowship 1082922.
Contributors SG contributed substantially to the study conceptualisation and design, interpretation of the data, and manuscript drafts and revisions. AK performed the data analysis and contributed substantially to the study conceptualisation and design, interpretation of the data, and manuscript drafts and revisions. AK and MO had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. EI contributed substantially to the study conceptualisation and design, data analysis, interpretation of the data, and manuscript drafts and revisions. MO contributed substantially to preparation of the data set, as well as to the interpretation of the data and manuscript drafts and revisions. All authors approved the final manuscript as submitted. EI and AK's substantial contributions to the study were undertaken while at the Centre for Community Child Health.
Funding Australian National Medical Research Council, 1082922, State Government of Victoria, 10.13039/501100004752, Australian Government.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Access to all unpublished data is subject to approval by the Department of Education and Training and the AEDC National Committee. Approval for access to unpublished data will be given for any legitimate research, analysis purpose or policy development. Applicants using AEDC data are bound by specific deeds of confidentiality to comply with all principles and procedures outlined in the AEDC Data Guidelines. AEDC Ethical Approval Guidelines are available to provide information around the need for ethics clearance to access AEDC data.