Article Text
PDF
Abstract
Background Lower-quality employment, characterised by excessive or part-time hours, irregular schedules and inadequate earnings, is a key social determinant of health among adults. Research examining parental employment quality in relation to the mental health and school performance of children is lacking. The study objective was to measure the associations between parental employment quality and child mental health symptoms and school performance.
Methods We conducted a secondary analysis of the cross-sectional 2014 Ontario Child Health Study. Dependent variables were parent-reported child mental health symptoms and school performance. We used latent class analysis (LCA) to characterise employment status, hours, scheduling and earnings of parents. We used linear and multinomial regression to model the associations between parental employment quality, mental health symptoms, and school performance.
Results Our study sample consisted of 9,927 children. The LCA of dual-parent households yielded three classes of parental employment quality, which we labelled ‘Dual Parent, High Quality’, ‘Dual Parent, Primary Earner Model’ and ‘Dual Parent, Precarious’. The LCA of single-parent households yielded two further classes, which we labelled ‘Single Parent, High Quality’ and ‘Single Parent, Precarious’. Compared with children in the ‘Dual Parent, High Quality’ group, children in all other groups had higher-level mental health symptoms and lower school performance. Children with ‘precarious’ parental employment in both groups showed the least favourable outcomes.
Conclusions Lower-quality parental employment was associated with increased mental health symptoms and poorer school performance among children. A clearer understanding of these relationships and their underlying mechanisms can help inform relevant policies and interventions.
- MENTAL HEALTH
- Health inequalities
- CHILD HEALTH
Data availability statement
Data may be obtained from a third party and are not publicly available. Data are stored and managed by Statistics Canada and are accessible through Statistics Canada Research Data Centres.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Parental employment is an important determinant of family socio-economic status and therefore child health. Multidimensional measures of job quality, including hours and schedules, are associated with adult health outcomes and may have spillover effects on child health and development.
WHAT THIS STUDY ADDS
Precarious parental employment, characterised by long or part-time hours, unpredictable schedules and lower income, is associated with greater parent report of mental health symptoms and poorer school performance among a population-representative sample of children and youth.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This research can inform the design and implementation of policies and programmes aimed at promoting high-quality employment for parents. In the clinical setting, clinicians caring for children can learn to incorporate multiple dimensions of parental employment in taking social histories and understanding the role employment might play in the presentation and treatment of child behaviour problems.
Introduction
Rising prevalence of poor mental health, diagnosed mood and anxiety disorders, and mental health-related resource use among Canadian children has been observed since 2011, long predating the COVID-19 pandemic, although the pandemic may have exacerbated existing trends.1 2 Similarly, school performance has been declining among Canadian children for some decades, including significant drops in literacy and mathematics performance.3 The causes of these challenges among children are likely complex, shaped to a large degree by biological, social and contextual factors, including exposures to socio-economic hardships such as low-income, family relationships, parenting and community contexts.4–6 Theories of child health and development suggest that socio-economic hardship is a powerful predictor of child mental health and development and, indeed, many indicators of socio-economic conditions have declined over this time as well.7
Socio-economic conditions in childhood are determined in large part through parental employment status and quality of parental employment, referring to hours, scheduling, earnings and other aspects of employment that influence well-being.8 The employment landscape in Canada and other wealthy nations has shifted over the last few decades, with fewer jobs approximating the ‘standard’ model of full-time, permanent employment and an increasing share of jobs characterised by contingent and precarious work arrangements.9 As a result, more parents are having to take up insecure and unstable jobs to make ends meet.10 The declining quality of parental employment has the potential to have a population-level effect on the ability of parents to meet the needs of their children.11
There are multiple mechanisms through which parental employment quality can affect children’s health. One mechanism is directly material. When employment quality is low, parents may have insufficient income for housing, food and other health-promoting resources. Additionally, both low-quality employment and financial hardship can lead to poorer child health through a variety of indirect mechanisms. It can act as an additional stressor for parents, affecting parental mental health and parenting practices. It can also affect family access to material resources such as childcare, educational and healthcare resources, healthy food and other activities that promote child well-being.12 Another mechanism can be characterised in terms of the effects on family relationships and the home environment. High-quality employment can promote family processes, including those related to parenting, the parent-child relationship, family routines and the home environment more broadly.13 14 Finally, employment quality also affects parental health, including physical illness (occupational injury, cardiometabolic health), mental health (stress, mood and anxiety disorders) and substance use, all of which impact children’s well-being.5 6 15–17
Employment quality can be improved through targeted policy and programmatic interventions.18 These include laws requiring predictable scheduling, employment protection legislation and minimum wage legislation. Employment quality thus holds significant potential as a modifiable, upstream intervention target for improving child health. Early research in this area found that individual indicators of lower employment quality were associated with higher rates of family poverty,19 poorer child cognitive development and greater incidence of externalising and internalising behaviours.13 However, research on the contributions of multiple dimensions of parental employment quality, and configurations of employment quality in two-parent households, to children’s mental health and academic performance is extremely scarce.
Our primary objective was to identify profiles of parental employment quality among dual- and single-parent households in a representative sample of children in Ontario. Our secondary objectives were to measure the associations between profiles of parental employment quality and mental health symptoms and school performance among those children. We hypothesised that profiles consistent with lower-quality employment would be associated with worse mental health and school performance.
Methods
Design and study sample
We conducted a secondary analysis of the 2014 Ontario Child Health Study (2014 OCHS), a provincially representative cross-sectional survey of households with children ages 4–17 in Ontario—Canada’s most populated province. Using the 2014 Canadian Child Tax Benefit file as the sampling frame, households were selected using cluster sampling of residential areas stratified by income to ensure the representation of households and neighbourhoods across the full-income distribution. The analytic sample includes 10 802 children aged 4–17 from 6537 households (50.8% response rate).20 Our study sample consisted of all participants in the OCHS who had a mother or father working in paid employment and who had responded to questions for main study variables. Households where both the mother and father were not in the labour force (ie, neither working nor looking for work) were excluded (n=125), given that these households were lacking information on the main exposure (ie, employment quality). We also excluded households with same-sex parents or who did not have a mother or father in the household (n=20) because the numbers did not make the minimum count criteria required for vetting and release by Statistics Canada. We accessed the data through the Statistics Canada Research Data Centre at McMaster University, which granted ethical approval.
Measures
Exposure variables
The primary independent variable was parental employment quality, assessed in terms of the following indicators: employment status (employed, unemployed or out of the labour force),21 employment hours (less than 30 hours per week, 30–49 hours per week or 50 or more hours), employment schedule (regular day schedule, regular evening or night schedule, rotating or split shift, or irregular schedule) and employed earnings (proxied by household income quintile as we could not differentiate employment income by family members). Each of these indicators was assessed separately for mothers and fathers in dual- and single-parent households, yielding one set of indicators describing mother’s employment quality and a second set of indicators describing father’s employment quality.
Outcome variables
Our dependent variables were parent-reported child mental health symptoms and school performance. Child mental health symptoms were measured using the OCHS Emotional Behavioral Scales (EBS),22 a 52-item valid and reliable parent-rated scale measuring emotional and behavioural symptoms of the following seven disorders from the Diagnostic and Statistical Manual of Mental Disorders: generalised anxiety disorder, separation anxiety disorder, major depressive disorder, social phobia, attention-deficit hyperactivity disorder, oppositional-defiant disorder and conduct disorder. For each symptom, parents are asked to report how often in the past six months it was ‘never or not true’, corresponding to a score of 0 for that item, ‘sometimes or somewhat true’, corresponding to a score of 1, or ‘often or very true’, corresponding to a score of 2. The total EBS score was calculated by summing the score across all 52 items, where a higher score indicates a higher report of symptoms.
Child and youth school performance was defined using a single question from the OCHS, asking parents, “Based on your knowledge of [your child’s] school work and most recent report card, how well is [your child] doing in the following subjects and overall”, with responses of ‘A/80–100%’, ‘B/70–79%’, ‘C/60–69%’ or ‘D/<60%’. We created a categorical variable based on parent-reported overall performance, defined as ‘A’, ‘B’ or ‘C or lower’, with A as referent.
Additional study variables
We included child, parent and household sociodemographic characteristics as covariates. For children, these were age (years), sex assigned at birth (male/female), history of preterm birth (due to potential confounding effects for school performance) and race/ethnicity (white, non-white).23 24 For parents, these were education (high school or less, undergraduate degree, more than undergraduate degree), parental immigrant status (born in Canada, migrant to Canada).25 For households, these were urbanicity (large urban area, medium/small urban area, rural area) and number of household members.
Statistical analysis
We used descriptive statistics to calculate means and proportions to characterise our study sample. Separately for dual-parent and single-parent households, we applied latent class analysis (LCA) to the indicators of parental employment quality defined above. LCA is recognised as an important strategy to capture the potential configurations of employment indicators that make up job quality.19 26 For dual-parent households, we required that at least one parent be active in the labour force (either working or actively seeking work). For single-parent households, we required that the parent be working, and so this LCA included only three indicators: employment hours, employment schedule and income quintile. Conducting separate LCA for dual- and single-parent households enabled us to include these additional indicators for dual-parent families, which would be missing in single-parent households. We were not able to assess single mothers and single fathers separately due to the small sample of single fathers. We evaluated models for the most appropriate class solution using fit indices, including Akaike information criterion, Bayesian information criterion, Vuong-Lo-Mendell-Rubin adjusted LRT and entropy, as well as assessing interpretability of the models. On selecting the most appropriate class solution, we used descriptive statistics to summarise the prevalence and key features of the latent employment quality groups.
Once we had identified the final LCA models, we used multivariate regression models to measure associations between parental employment quality (assessed in terms of latent class membership) and the two dependent variables, linear regression for mental health and multinomial logistic regression for school performance. We adjusted for child, parent and household sociodemographic variables. We tested for interaction by child age (4–11 vs 12–17) using interaction terms. Finally, to provide more information about the specific effects of employment variables, we conducted a sensitivity analysis measuring associations between individual indicator variables and each outcome excluding and including income. This was to assess whether employment indicators were meaningfully accounting for the effects of income. The LCA was estimated using full information maximum likelihood (FIML) with robust SEs in MPlus V.8.3.27 FIML estimates parameters based on the available complete data as well as the implied values of the missing data given the observed data. The multivariate regression models were applied in STATA 18.0, and bootstrapping technique with replicate weights was used to account for the complex survey design. All models were estimated with sampling weights to account for the unequal probability of selection.
Results
Study sample
Our study sample consisted of 9927 children ages 4–17 years. They had a mean age of 10.6 years, 48% were identified as female, 63.7% identified as white race/ethnicity and 40.3% were born prematurely. Among parents, 52.8% reported education of high school or less, 43.0% were not born in Canada and 69.1% lived in a large, urban area (table 1).
Participant characteristics
Latent profiles of parental employment quality
The LCA of dual-parent households supported a three-class solution (Vuong-Lo-Mendell-Rubin adjusted LRT p=0.07 2 vs 3 classes). Results of LCA are presented in table 2: (1) both parents working full-time jobs with regular schedules and incomes distributed across quintiles 2–5, which we labelled ‘Dual Parent, High Quality’; (2) fathers working full-time jobs, often involving long hours and irregular schedules, and mothers working part-time jobs of varying quality, with incomes distributed across quintiles 1–3, which we labelled ‘Dual Parent, Primary Earner Model’; and (3) both parents working lower quality jobs involving short hours and irregular schedules, with incomes distributed across quintiles 1–2, which we labelled ‘Dual Parent, Precarious’. The LCA of single-parent households supported a two-class solution (Vuong-Lo-Mendell-Rubin adjusted LRT p=0.22 for 1 vs 2 classes), presented in table 3: (1) single parents working high-quality jobs with moderate income, which we labelled ‘Single Parent, High Quality’; and (2) single parents working lower quality jobs with low income, which we labelled ‘Single Parent, Precarious’. The final independent variable in our model was a single categorical variable including the five and a sixth category for single parents who were unemployed (n=264). Detailed fit statistics can be found in online supplemental file 1.
Supplemental material
Profiles of parent employment quality for dual-parent families: latent class analysis
Profiles of parent employment quality for single-parent families: latent class analysis
Parental employment quality and child and youth mental health
We observed a greater number of mental health symptoms among children and youth whose parents reported lower quality employment (table 4). Compared with children in the ‘Dual Parent, High Quality’ group, children in all other groups had significantly higher EBS scores after adjustment. Children in the ‘Dual Parent, Precarious’ and ‘Single Parent, Low Quality’ groups had the highest scores (ß: 4.91, 95% CI 3.35 to 6.47; ß: 3.95, 95% CI 2.31 to 5.95). These were followed by lower scores in the ‘Dual Parent, Primary Earner’ and ‘Single Parent, High Quality’ groups (ß: 1.89, 95% CI 0.87 to 2.91; ß: 2.16, 95% CI 0.61 to 3.72). For children in single-parent households, ‘low-quality’ parental employment was associated with higher EBS score (ß: 3.95, 95% CI 2.31 to 5.59) than unemployment (ß: 2.60, 95% CI 0.14 to 5.07). There were no significant interactions by age (4–11 vs 12–17).
Associations between parental employment quality and child mental health†
Parental employment quality and school performance
Children in households with lower quality employment had worse parent-reported school performance outcomes (table 5). Compared with children in the ‘Dual Parent, High Quality’ group, children in the ‘Dual Parent, Primary Earner Model’, the ‘Dual Parent, Precarious’ group and both single-parent groups were more likely to have lower grades (ie, B, C or lower grade compared with A) after adjustment. A school performance of C or lower was most prevalent among children in the ‘Dual Parent, Precarious’ group (prevalence ratio (PR): 2.51, 95% CI 1.60 to 3.91) and children in the ‘Single Parent, Low Quality’ group (PR: 1.93, 95% CI 1.10 to 3.40). These were followed by the ‘Dual Parent, Primary Earner’ (PR: 1.74, 95% CI 1.23 to 2.47) and ‘Single Parent, High Quality’ (PR: 1.75, 95% CI 1.10 to 2.81) groups.
Associations between parental employment quality and parent-reported school performance
Sensitivity analysis
Our sensitivity analysis entered employment quality variables manually into the model and is presented in online supplemental files 2 (mental health symptoms) and 3 (school performance). While individual dimensions of parental employment quality are associated with mental health and school performance, they are less predictive than our newly derived latent indicators.
Discussion
In this study, we identified profiles of parental employment quality among households in Ontario, Canada, and then examined their associations with child mental health symptoms and school performance. We found that being in a household in which parents reported lower quality employment was consistently associated with worse mental health and school performance outcomes among children. This was particularly true for children in either of the two ‘Precarious’ employment groups. Even in the ‘Dual Parent, Primary Earner Model’ group, however, children had more adverse mental health and school performance outcomes than their peers with the highest quality of parental employment.
As a predictor of child health and family socio-economic status, parental employment is often captured as a dichotomy of employed versus unemployed. Our study highlights that this distinction does not appear to capture some of the important drivers of child mental health and development. There is increasing evidence that even heterogeneity within employment is important, with recent literature focused primarily on the health of working-age adults.8 However, our study identifies multiple configurations of employment within households, particularly dual-parent households, where variations in the quality of each parent’ employment may have differential effects on the well-being of children, including whether each parent is employed, their working hours and job stability.
Our findings extend the research that shows an important distinction between employment quality and employment status (ie, lack of employment). A systematic review in 2023 suggested that maternal employment was associated with greater report of externalising behaviours, but lower internalising and greater prosocial behaviour.28 Configurations other than two parents working high-quality jobs were associated with worse mental health for children, including fathers working extended hours potentially as a trade-off for mothers working part time. This finding was somewhat surprising in the context of the literature around maternal employment. It is possible that the stress of long hours or precarious employment of fathers has meaningful effects on child mental health even when another parent is at home. Perhaps the financial implications of this trade-off are more harmful than the benefits of part-time employment for families, or the employment configuration may have been driven by childcare needs, work availability or other factors. Regardless of the motivations for labour force participation, interventions to support high-quality employment are needed. Focusing on employment-related interventions offers an opportunity to target policy-amenable upstream causes. These may include policies which promote a living minimum wage, job stability and predictable scheduling.
We found stronger associations between classes of parental employment quality and mental health than school performance. This may reflect different mechanisms of its effects as well as measurement differences (symptom scoring for mental health compared with subjective school performance in three categories). Among adults, there is evidence that low-quality or precarious employment is associated with poorer mental health.29–31 Parental mental health is a meaningful driver of child and youth mental health,32 which may explain the strength of this association to some degree. Long hours or unpredictable schedules may affect parenting practices or lead to reduced predictability in routines, all of which may be associated with child behaviour and development.33
Additionally, our latent profiles of parental employment quality likely illustrate some of the trade-offs that parents make in meeting material needs, against the background of labour market trends characterised in part by increasing prevalence of precarious jobs.9 Previous literature has focused extensively on the effects of maternal employment on child outcomes in the context of increasing female participation in the workforce, finding mixed associations between indicators of employment quality and child school performance,34 child well-being35 and child and adolescent mental health.13 36–38 Our LCA provides a more nuanced view of household configurations of multiple indicators of employment quality for dual- and lone-parent families. Examining employment quality at the household level, with a single, comprehensive and multi-dimensional indicator, was more predictive than individual indicators.
This study has several strengths. It draws on a large and representative survey with robust measures of child mental health. Additionally, this study fills a meaningful gap in the social determinants of health literature by addressing multiple indicators of parental employment as well as incorporating the family dynamics involved in labour participation. Our study was subject to certain limitations. This was a cross-sectional study precluding causal inferences, and all variables were by parental self-report which may be subject to bias. Parental employment may be affected by their child’s health which may interfere with their labour force participation, with the possibility of reverse causality.39 Additionally, latent class groupings may fluctuate over time, and this cross-sectional analysis was not able to assess temporal changes in the exposure. Though we did not identify significant interactions by age, there may be sensitive time periods both in terms of timing of the exposure and manifestation of the outcome. Importantly, parental mental ill-health may affect their perception and reporting of their child’s mental health symptoms.40 In an effort to avoid over-adjustment, however, we did not adjust for parental mental health or parenting as these are also hypothesised mechanisms through which parental employment quality is associated with children’s outcomes.41 Not all components of job quality were measured; additional factors such as choice of schedule or hours, or job satisfaction, should be assessed in future research. While these data reflect the best available at this time, the survey was collected in 2014, and there have likely been important changes in employment and mental health that should be considered in future work. Finally, due to limited numbers, we restricted our analysis of dual-parent households to those with an identified mother and father, and so other dual-parent configurations or families with multiple caregivers would not be reflected.
Conclusions
This study of parental employment quality and child mental health symptoms and school performance found that lower quality parental employment among mothers and fathers was associated with poorer mental health as well as poorer school performance. The study highlights the importance of examining multiple dimensions of employment quality as well as household configurations of employment in the context of two working parents. Design and implementation of policies and programmes aimed at promoting high-quality employment for parents should consider the downstream benefits for children. Clinicians caring for children should consider multiple dimensions of parental employment when applying family-centred approaches to preventing and addressing child mental health and behaviour concerns. Future research should incorporate longitudinal design to identify causal relationships, unpack mechanisms and determine potentially sensitive developmental periods for exposure.
Data availability statement
Data may be obtained from a third party and are not publicly available. Data are stored and managed by Statistics Canada and are accessible through Statistics Canada Research Data Centres.
Ethics statements
Patient consent for publication
Ethics approval
We accessed the data through the Statistics Canada Research Data Centre at McMaster University, which grants ethical approval. Participants gave informed consent to participate in the study before taking part.
References
Footnotes
AEF and FVS are joint first authors.
AEF and FVS contributed equally.
Contributors AEF conceptualised and designed the study, reviewed the analyses, drafted the initial manuscript and critically reviewed and revised the manuscript. FVS conceptualised and designed the study, reviewed the analyses and critically reviewed and revised the manuscript. JC, GW and AS conceptualised and designed the study and critically reviewed and revised the manuscript. LW conceptualised and designed the study, conducted the analyses and critically reviewed and revised the manuscript. JRD, HMcM and CB critically reviewed and revised the manuscript. KG conceptualised and designed the study, designed the data collection instruments, reviewed the analyses and critically reviewed and revised the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work. AEF is the guarantor.
Funding This research was supported by a Catalyst Grant from the Edwin S.H. Leong Centre for Healthy Children. Grant number: not applicable.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.