Maternal postnatal depressive symptoms and offspring emotional and behavioral development at age 7 years in a U.K. birth cohort: The role of paternal involvement.

Maternal Postnatal Depression, Offspring Development and Mechanisms of Transmission

A large body of research has now documented the high prevalence of maternal postnatal depression (PND; Ertel et al., 2011 ; Horowitz & Goodman, 2004 ), the link to adverse offspring emotional, behavioural and cognitive development across the lifespan (Dachew et al., 2021; Neri et al, 2020 ), and associated deleterious health and societal consequences ( Bauer et al., 2016 ). The rapid emotional and cognitive changes that children undergo at this stage of development may make them particularly vulnerable to the adverse effects of maternal PND (Bagner et al., 2020; Goodman & Garber, 2017 ).

However, outcomes of children whose mothers experience PND are not consistently poor, and effect sizes for the associations are mostly moderate or small ( Stein et al., 2014 ). Insights into factors that accentuate or ameliorate the risks may improve our understanding of intergenerational transmission of mental health difficulties in families and contribute to the development of targeted interventions to reduce adverse impact of maternal PND on offspring development. Parenting is a key modifiable environmental mechanism ( Kaminski et al., 2008 ) that explains some of the adverse effects of maternal PND on the child, representing an intervention target of particular importance ( Goodman & Garber, 2017 ). Evidence suggests that maternal PND disrupts maternal sensitivity and involvement ( Murray et al., 2010 ), which, in turn, are associated with offspring emotional and behavioural problems ( Giallo et al., 2014 ). The importance of paternal mental health ( Ramchandani et al., 2005 ) and involvement ( Diniz et al., 2021 ) for offspring development in the context of ‘normative’ family functioning is increasingly recognised. However, longitudinal studies that examine the role of paternal involvement in the association between maternal PND and offspring development ( Goodman et al., 2014 ) are lacking. Paternal involvement may be of particular importance for mothers and infants in families with mothers experiencing PND through support of the mother and care of the infant ( Goodman et al., 2014 ).

Paternal Involvement and Offspring Development in the Context of Maternal Postnatal Depression

Both ‘compensatory/buffering’ and ‘spillover’ hypotheses have been articulated to describe possible roles of fathers in the context of maternal PND ( Goodman et al., 2014 ). In line with a family-level resilience perspective ( Patterson, 2002 ), fathers may try to compensate for maternal PND by becoming more involved, ‘buffering’ the adverse effects of maternal PND on the child ( Mezulis et al., 2004 ). Conversely, maternal PND may have a ‘spillover’ effect by negatively affecting paternal involvement and increasing the risk of adverse offspring outcomes ( Goodman et al., 2014 ). Evidence for ‘compensatory/buffering’ effects of paternal involvement on offspring outcomes is inconsistent. Some studies report an increase in positive, warm and sensitive father-child interactions in families where mothers experience PND compared to families where mothers are not affected by PND ( Edhborg et al., 2003 ; Vakrat et al., 2018 ), but other studies report no evidence that paternal involvement is protective (i.e., moderates the association between maternal PND and offspring development; Carro et al., 1993 ; Chabrol et al., 1996 ). The ‘spillover’ hypothesis has also been substantiated with some evidence linking maternal PND to lower levels of paternal involvement with their infants (Paulson et al., 2011).

Family Systems Theory, where the family is viewed as systems of relationships with each individual/subsystem in the family influencing and being influenced by the others ( White et al., 2018 ), is the theoretical framework that underlies both hypotheses. Arguably, with the birth of the child and emergence of substantial postnatal caregiving needs, interdependency in the parental couple system increases, with maternal PND influencing paternal involvement, which, in turn, may influence the relationship between maternal PND and offspring development. Furthermore, mothers who experience PND engage in caregiving tasks to a greater extent than fathers, who are more likely to withdraw from caregiving and reduce involvement when experiencing PND. Thus, it is not always the case that fathers assume a compensatory/buffering role when mothers are affected by depression ( Paulson et al., 2006 ). It is important to provide more insights into the characteristics of paternal involvement when mothers experience PND. However, evidence regarding the impact of maternal PND on the nature and quality of paternal involvement is scant, and, to the best of our knowledge, no large population-based studies to-date have examined a possible mediating role of paternal involvement for child development in the context of maternal PND. Lack of research attention on paternal involvement as a potential mechanism in this context is an important omission given growing evidence emphasising the importance of paternal involvement in numerous aspects of offspring development ( Lamb, 2010 ) and persistent calls for including fathers in scientific and clinical approaches to family functioning affected by maternal PND ( Masten & Monn, 2015 ).

Paternal Involvement as a Multidimensional Construct

Paternal involvement is multidimensional, as conceptual definitions long ago departed from traditional breadwinning roles to ‘intimate’ fatherhood that prioritises emotional closeness and the quality of the father-child relationship ( Dermott, 2014 ). Paternal involvement is also no longer solely defined as time spent together ( Parke & Cookston, 2019 ), with more nuanced facets, including direct nurturant caregiving and disciplining, as well as emotional and practical support of the mother, being linked to improved offspring emotional and behavioural outcomes ( Lamb, 2010 ). New conceptualisations of paternal involvement include child-focused behavioural (e.g., direct involvement in caregiving), affective (e.g., quality of the father-child relationship, such as enjoyment, warmth and conflict), and cognitive (e.g., parenting confidence and beliefs regarding caregiving) dimensions ( Marsiglio et al., 2000 ; Pleck, 2010 ). Aspects of paternal involvement that are shaped by maternal influences and mother-father relationship (e.g., ‘gatekeeping’, a complex phenomenon defined as maternal beliefs and practices that discourage or facilitate paternal involvement; Miller, 2018 ) have also been described. Supporting the mother through sharing household tasks and responsibilities is another important dimension of paternal involvement, with significant consequences for the child ( Parke & Cookston, 2019 ). Similar to mothers, fathers are also under pressure to balance caregiving with demands of paid work, childcare and family life ( Lamb, 2013 ; Miller, 2010 ) often accompanied by mothers’ perceptions that fathers are not ‘pulling their fair share’ with this benchmarking influencing paternal involvement ( Pleck, 2012 ).

Recognition that child-focused and mother-influenced paternal involvement are important for offspring development represents another marked shift in fatherhood scholarship ( Lamb, 2010 ), explicitly acknowledging that the father-child relationship is embedded in a network of relationships among fathers, mothers and children ( Cabrera et al., 2014 ). However, few studies have examined paternal involvement as a multidimensional construct that captures cognitive, affective and behavioural aspects as well as maternal influences ( Schoppe-Sullivan et al., 2004 ). Failures to do so preclude establishing the individual relative contributions of various dimensions of paternal involvement to offspring development ( Lamb, 2010 ). We address this limitation by examining several dimensions of child-focused and mother-influenced paternal involvement in relation to offspring emotional and behavioural outcomes in the context of maternal PND.

Importance of Paternal Postnatal Depression

Parenting is transactional and dynamic ( Bornstein, 2015 ). Existing research acknowledges reciprocal influences and the interplay between paternal involvement and the child, with offspring response to parenting shaped by temperamental and biological dispositions, which may evoke variations in parenting behaviours ( Avinun & Knafo, 2014 ). In addition, paternal PND and its impact on offspring development in the context of maternal PND is often overlooked. Maternal depression in late pregnancy and during the postnatal period are associated with increased risk of paternal PND ( Fredriksen et al., 2019 ), which in turn influence paternal parenting ( Wilson & Durbin, 2010 ) and offspring emotional and behavioural development ( Ramchandani et al., 2005 ), thus potentially influencing the association between paternal involvement (mediator) and offspring development (outcome); Figure 1 presents conceptual pathways between maternal and paternal PND, paternal involvement and offspring development. Paternal and maternal PND may increase offspring risk of adverse development ( Goodman et al., 2014 ), while absence of paternal PND may be a protective factor reducing adverse effects of maternal PND on paternal involvement and offspring outcomes. Mezulis et al. (2004 ) found that paternal PND exacerbated adverse effects of maternal PND on offspring behavioural problems, but this moderating effect was limited to those fathers affected by depression who spent considerable time caring for their children. To the best of our knowledge, no longitudinal study has examined the role of paternal PND in the associations between maternal PND, paternal involvement and offspring development. We focused on the first year of the child’s life, which may be a sensitive period of increased vulnerability to the adverse effects of maternal PND and parenting on offspring development ( Bagner et al., 2010 ). This is a time when most infants are highly dependent on parental care and involvement. In order to satisfy temporal precedence assumptions that underly mediational models, the earliest time point of maternal PND was chosen to align with the earliest assessment of paternal involvement items (8 weeks), while paternal PND was assessed at 8 months and offspring development at 7 years.

Conceptual diagram for exposure (maternal PND), mediator (child-focused and mother-influenced dimensions of paternal involvement), outcome (offspring development), baseline confounders (child, socioeconomic, maternal and family characteristics) and intermediate confounder * (paternal PND)

* Intermediate confounder: variable induced by the exposure, affecting both the mediator and the outcome, thus confounding mediator-outcome association

The Current Study

The current study aims to build and expand on existing research by examining the role of paternal involvement as a potential mediating mechanism that could transmit the adverse effects of maternal PND on offspring development. We took advantage of rich longitudinal data on paternal parenting collected in a large UK-based birth cohort study, the Avon Longitudinal Study of Parents and Children (ALSPAC) to (1) model child-focused and mother-influenced dimensions of paternal engagement and (2) estimate the extent to which the association between maternal PND at 8 weeks and offspring emotional and behavioural difficulties at 7 years is mediated by child-focused and mother-influenced dimensions of paternal involvement during the first four years of the child’s life. Quantifying this association has important implications for the design and development of preventative and intervention programmes, given the modifiable nature of parenting ( Kaminski et al., 2008 ). The unique nature of the rich longitudinal ALSPAC data enabled us also to account for a range of confounding factors, including child polygenic score for neuroticism (PGS) as a proxy for genetic confounding that may indicate possible child evocative effects, and paternal PND as a possible intermediate confounder. It is important to note that the study focused on parents with postnatal depressive symptoms (as assessed by the Edinburgh Postnatal Depression Scale; Cox et al., 1987 ), and not a clinical diagnosis of depression.

Outcome: Offspring Emotional and Behavioural Development

Offspring emotional and behavioural development were assessed using the Strengths and Difficulties Questionnaire ( Goodman, 1997 ) completed by the mothers of the study children at age7 years). The scale consists of 25 questions with five subscales (prosocial [i.e., engagement in positive social interactions], emotional symptoms, conduct problems, hyperactivity and peer problems), which have been extensively validated, demonstrating high consistency, reliability and diagnostic predictability ( Goodman, 2001 ). Consistent with previous research, the first 4 subscales were combined to derive a total difficulties score used in all analyses ( Gutierrez-Galve et al., 2015 ).

Mediators: Child-Focused and Mother-Influenced Dimensions of Paternal Involvement

Full details pertaining to item selection and theoretical development of factors are presented in Methods S1. In summary, potential parenting items (>150) were extracted from paternal self-reported questionnaires administered on 5 occasions from birth to age 3 years 11 months. Conceptual factor underpinnings were drawn from extensive empirical and sociological literature on paternal engagement in infancy ( Dermott, 2014 ; Marsiglio et al., 2000 ; Palkovitz, 2019 ; Pleck, 2012 ). In line with a revised conceptualisation of paternal involvement ( Pleck, 2010 ; Schoppe-Sullivan et al., 2004 ), individual items fell into two theoretically distinct sources of paternal involvement: (1) child-focused paternal involvement capturing behavioural (e.g., direct involvement in caregiving), affective (e.g., quality of father-child relationship, such as enjoyment, warmth and conflict, worries about the child) and cognitive (e.g., parenting confidence and beliefs regarding caregiving) dimensions directed at the child; and (2) mother-influenced paternal involvement with the child through the lens of maternal expectations (e.g., maternal ‘gatekeeping’, managing employment and parenthood), mother-father relationship (e.g., paternal beliefs regarding mother-father relationship and its impact on parenting) and indirect material care through support of the mother (e.g., paternal help with household tasks and responsibilities). The hypothesised latent factor (CFA) models capturing child-focused (6 dimensions), and mother-influenced (4 dimensions) paternal involvement are represented in Figure 3a-b , with derived factors, items, standardised factor loadings and fit indices for the measurement models presented in Tables S1-S2 . Two separate measurement models including CFA-based factors ( Figure 3a-b ) and structural mediation models including pathways between constructs ( Figure 4a-b ) were fitted to capture potentially differential mediating roles of child-focused and mother-influenced paternal involvement in the association between maternal PND and offspring development.

Structural equation mediation model estimating the direct effect of maternal postnatal depression on offspring development at age 7 years and the indirect effect through specific factors capturing child-focused and maternal-influenced dimensions of paternal involvement, adjusted for child PRS, gender, antenatal baseline confounders, and paternal PND as an intermediate confounder

Note: Observed variables are represented by squares. Latent variables are represented by circles. Error terms covariances, correlations between the factors and individual items loading onto each specific factor are not represented to reduce figure complexity.

Pathways that constitute total indirect effect are represented by bold lines; pathways that constitute direct effect are represented by dashed lines.

Potential Baseline Confounders: Child Polygenic Score for Neuroticism, Socioeconomic, Parental and Family Characteristics

Analyses were adjusted for child polygenic score (PGS) for neuroticism to account for possible child genetic correlations including evocative associations with parenting ( Mills-Koonce et al., 2007 ). However, it should be noted that polygenic scores capture only a small fraction of heritability, thus they do not comprehensively account for possible genetic confounding ( Pingault et al., 2021 ). Genotyped data were available for 8,237 children in the ALSPAC cohort (full details in Methods S1). Disadvantaged socioeconomic status, marital status and conflict are also strong risk factors for maternal PND ( Milgrom et al., 2008 ), less optimal paternal parenting practices ( Leigh & Milgrom, 2008 ), and adverse offspring emotional and behavioural development ( Stein et al., 2014 ). In consequence, we adjusted for a range of antenatal prospectively measured potential confounding factors extracted from maternal questionnaires, including: highest maternal educational attainment (minimal education or none/compulsory secondary level [up to age 16 years; O-Level], non-compulsory secondary level [up to age 18 years; A-Level/university level education]), maternal age in years, presence of financial difficulties or no financial difficulties, homeownership status (owned/mortgaged, private/council [subsidised public housing] rented), marital status (married, never married) and a continuous score capturing parental conflict (higher scores representing higher levels of conflict). Given the primary exposure of interest was maternal PND and lower response rates for fathers compared to mothers, as well as moderate to strong correlations between maternal and paternal education (r=0.6, p≤0.001) and age (r=0.7, p≤0.001), analyses were adjusted for mother-reported baseline confounders only to avoid introducing missing data.