In this prospective birth cohort study including 536 families, we estimated five different classes of maternal employment trajectories 0–36 months after childbirth and their possible associations with social and emotional outcomes in children aged 4–6 years. For boys, trajectory classes 2, and 3 were associated with poorer mental health at ages 5 and 6, whereas for girls, class 5 was associated with better mental health. In addition, assessment of employment status 1 year postpartum revealed a positive association between part-time employment and subsequent mental health in boys. The findings suggest the possibility of effect modification by gender and highlight the complexity of the issue.
Mothers in trajectory classes 2 and 3 worked relatively few hours in the first 12 months postpartum, followed by weekly hours of around 20 and 30 hours respectively. Other studies have suggested that an early return to work by mothers (in the first 12 months postpartum) may be particularly detrimental to the child's mental health.1,9 Our results for trajectory classes do not support this conclusion for neither boys nor girls. However, as class 4 - mothers working less than 6 hours a week up to 24 months postpartum - was not associated with mental health in neither boys nor girls, our results seem to support the possibility that it's early, rather than later, maternal employment that may be associated with children's mental health.7,9,12,30,31 This is supported by the results for girls: Class 5, with a rapid increase in weekly hours in months 1–12 postpartum, was substantially associated with better mental health in girls aged 5 and 6. However, the latter findings do not support the hypothesis that early maternal employment is detrimental to children's mental health. One reason for this may be that class 5 was not suitable for identifying possible adverse effects in the first 12 months, because in this class working hours decreased between months 24 and 36 postpartum. Another reason may be that we did not distinguish between externalizing and internalizing behavioral problems, as we were interested in the overall effect on children's mental health. Nevertheless, the associations found among girls are consistent with findings from a large nationally representative cohort of British children born in the new millennium.24 The authors found no detrimental effect of maternal employment on child mental health, and even suggested a positive effect on the mental health of 5-year-old girls.24 Further research supports these findings.1,9,11,13,32
Employment at 12 months
Our results suggest that maternal part-time vs no employment at 1 year postpartum was associated with fewer social and emotional difficulties at 4 years only in boys. Similarly, Brooks-Gun et al. found that early part-time employment may have a positive indirect effect on externalizing problems, through differences in the home environment and maternal sensitivity.7 In a meta-analysis, Kopp et al. suggested that part-time employment, longer employment, and returning to work later than 12 months postpartum may be particularly beneficial for a child’s mental health.1 Other studies support this conclusion.32,33 However, as noted above, we also found that for boys, trajectory classes 2, and 3 were associated with adverse outcomes at ages 5 and 6. These results appear somewhat contradictory, but show that the assessment of a single point in time was associated with a positive effect at age 4 (near-term outcome), whereas assessing employment trajectories (reflecting maternal employment patterns 0–36 months postpartum) showed effects in the opposite direction for more distal outcomes.
Moderators, Mediators, and adjustment variables
Our finding that child gender moderates the association between maternal employment and child mental health is somewhat consistent with the current literature.7,24 Brooks-Gun et al. reported that in the few cases where effect modification occurred, the association was more detrimental for boys than for girls (analyzing n = 900 non-Hispanic white children).7 However, Hope et al, including n = 11,538 children, found no effect modification between child gender and maternal employment.11 In our study (as well as in the studies cited), reporting bias (perception bias) is possible: Interestingly, maternal employment status affects parents' perceptions of their child’s mental health differently for boys and girls, and may therefore contribute to gender differences in the effects of maternal employment on child mental health.34,35
In our multivariable regression analysis, parental stress and anxiety at/before childbirth were associated with child mental health (Supplementary Tables S7-S10), similar to results of previous studies.36,37 Given that elevated levels of depressive symptoms were found in mothers who worked during the first 12 months postpartum,7 an indirect negative effect of maternal employment on child mental health, mediated by maternal mental health, is possible. Interestingly, the presence of an older sibling seemed to be associated with better mental health for both genders.
Marginal part-time employment
For both gender there was a tendency for a negative association between maternal marginal part-time employment at 1 year postpartum and child mental health. Marginal part-time employment may be an indicator of an urgent need for money. The transition into poverty (within 11 years postpartum) was found to have a negative effect on child and maternal mental health.38 Compared with the findings for part-time employment 1 year postpartum (positive effect for boys), this shows that the type and reasons for employment may play a role. This needs to be taken into account when translating the findings (poorer mental health of boys in trajectory classes 2 and 3 vs 1) into advice for families. In the long term, these mothers prevent the transition into poverty. Participation in the labour market while having young children is recommended from a long-term perspective. This conclusion is supported by another study analyzing maternal employment trajectories over a much longer period than ours (9 months, 3, 5 and 7 years postpartum).11 The authors found that maternal employment was associated with fewer social and emotional difficulties at 7 years.11
Limitations
The interpretation of our results is limited by the sample size and the resulting lack of statistical power. In addition, we had a high proportion of families with high maternal educational attainment at baseline, which is representative of the local population, but families with low educational attainment and migrant backgrounds had higher loss to follow-up, especially during the first year of follow-up. Further, it is possible that those mothers who noticed a delay in their child's socio-emotional development did not return to work,19 and therefore the reference group (no employment) may be confounded. As mothers' employment decisions depend on many factors19 that also influence children's mental health, residual confounding is still possible.
Class 5 had the highest proportion of mothers having a second child 36 months postpartum (51.4% versus 24.0% in class 1, 8.6% in class 2, 5.3% in class 3, and 6.4% in class 4). As this variable was included in our analysis at each level of adjustment, only mothers who had or had not had a second child were compared. Therefore, the results can be considered independent of this circumstance.
The relatively low number of mothers working full-time at 1 year postpartum in our sample has to be taken into account. This may be the reason for the statistically non-significant associations when comparing full-time vs no employment at 1 year postpartum.