Journal of Childhood & Developmental Disorders

Introduction

Maternal mental health difficulties, such as symptoms of anxiety or depression, are well-established risk factors for childhood mental health problems, including externalising problems (e.g. conduct problems) and internalising problems (e.g. depression). Potential explanations include transmission of genetic risk factors, impacts of prenatal maternal depression on foetal neurodevelopment, dysfunctional parenting and insecure attachment. Several studies have highlighted that familial factors such as genetic risks and shared family environment explain considerable variance in the associations between child and maternal mental health.

In transactional models of child development, parent–child transactions contribute to associations between parent and child mental health, with children viewed as active participants in shaping their social environment, including their parents’ behaviour and mental health. However, existing research has mostly focused on how maternal mental health problems affect child mental health. The few studies to address bidirectional effects have yielded mixed findings: child-driven effects on parental psychopathology are reported in some studies but not others. Moreover, despite meta-analytic evidence suggesting effects of paternal mental health problems on child adjustment problems, especially externalising problems, most investigations of parent–child transactions focus on maternal mental health problems, leaving fathers comparatively overlooked. Including fathers is also important for investigating the reciprocal impacts of mental health between parents. Recent studies to examine reciprocal relations between child, paternal and maternal mental health have yielded mixed findings, highlighting the need for further research in this area.

Prevalence of Mental Health Problems

Notably absent are longitudinal investigations: existing studies of relations between child and parental mental health rely heavily on cross-sectional data, precluding analyses of directional effects. Studies adopting a longitudinal approach have mostly focused on just two time points; or a limited developmental span or employed long time intervals between measurement occasions or only investigated parent-to-child effects.

Another methodological limitation of existing research relates to the statistical operationalisation of dynamics that are presumed to play out within families. Most previous studies have relied on comparing group-level differences, that is between-person differences in mental health problems that reflect the effects of time-stable risk factors differing across families (e.g. the effects of genes). However, the developmental effects of primary interest in transactional theories reference within-person and within-family processes. For example, increased paternal mental health problems may be hypothesised to lead to increases child internalising problems within the same family. Illuminating these within-family mechanisms can yield more suitable targets for intervention.

Previous longitudinal research has commonly used methods such as the cross-lagged panel model. However, these models do not appropriately disentangle within- from between-family effects and so yield ambiguous results with regard to the withinfamily dynamics most commonly of interest in developmental theories of child–parent mental health transactions. Two exceptions deserve note. Tyrell et al. tracked 392 Mexican– American and European–American adolescents (ages 13, 15, 20 and 22 years), using a trait and time-varying cross-lagged model to investigate within-family relations of parental and adolescent mental health. Xerxa et al. tracked 5,536 Dutch children (1.5, 3 and 10 years), using an autoregressive latent trajectory model with structured residuals (ALT-SR) to investigate family mental health dynamics. Both studies reported reciprocal relations between parent and child mental health.

Early Identification and Intervention

Bridging the developmental scope of these two studies, the current study uses data from the Millennium Cohort Study, a large UK representative birth cohort study (N = 10,746) to explore the within-family developmental relations between both maternal and paternal mental health problems, such as symptoms of anxiety and depression, and child internalising and externalising problems across childhood and adolescence (ages 3, 5, 7, 11, 14 and 17). Rather than using a clinical sample, we focus on a general population sample. Although this limits our ability to identify parent–child mental health transactions as they play out for those with serious mental health issues, our community-ascertained sample has some advantages. This includes being less likely than a clinically ascertained sample to be influenced by Berkson’s bias (i.e. overestimation of symptom co-occurrence due to different symptoms independently affecting treatment seeking; Berkson) or range restriction (i.e. underestimation of symptom co-occurrence due to participants only showing symptoms on the upper extreme; Murray, McKenzie, Kuenssberg, & O’Donnell) Meaningful variation in mental health problems exists both below and above clinical cut-offs, making it important to study parent–child mental health relations in representative samples that include the full spectrum of continuously distributed mental health symptoms. We focus on child internalising and externalising as key outcomes as these are the most common mental health issues experienced over this developmental period.

To appropriately operationalise within-family dynamics, we built an autoregressive latent trajectory model with structured residuals (ALT-SR) to disaggregate within- and between-family effects. As associations between parental and child mental health may differ between boys and girls, we stratified by gender. There is limited research to guide specific hypotheses regarding effects at particular developmental stages and involving parents and children of particular genders. Based on transactional models that assume reciprocal effects, we investigated both parent-to-child and child-to-parent effects and considered whether these span the entirety of child and adolescent development, involve both mothers and fathers and hold for both boys and girls.

Supporting a transactional model of child development, our results demonstrate child-to-parent, parent-to-child and parentto- parent effects (with child-to-mother effects being especially consistent). These findings indicate that the whole family system should be included in intervention efforts. Family-focused interventions such as family system therapy may prove effective in reducing both parental and child mental health difficulties. Future studies are needed to investigate potential mediating factors in the transactions between parental and child mental health such as coercive parenting and parent–child relationship quality, as well as moderating factors such as family structure or socio-economic status.