Unintended Intervention Effects: Family Foundations Reduces the Influence of Maternal Depression on Infant Sleep

Introduction:

Prior research finds that maternal depression is related to worse infant sleep. The direction of effects between these is likely bidirectional. Infant night waking can cause distress and fatigue for mothers, and distressed mothers may engage in inappropriate infant sleep practices causing poor sleep outcomes. Prior research also shows that positive couple relationships can buffer the impact of depression on maternal parenting. Therefore, we examine whether an intervention aimed at improving coparenting (i.e., an intervention not aimed at infant sleep) might buffer against the effects of maternal depression on infant sleep outcomes.

Methods:

Participants were 169 couples in a randomized intervention testing a program aimed at improving coparenting in first-time parents. Maternal depression and infant sleep hours were measured post-intervention at 7 months and 1 year post-birth.

Results:

In prior work, we found significant program effects on maternal depression and some aspects of infant sleep. Here, however, we seek to understand whether the intervention affected the association between these two variables. We examined bivariate correlations between depression at 7 months and 12-month infant sleep and found the expected negative correlation (r = -.37, p < .001). We then examined this correlation by intervention (r = -.09, p = .43) and control condition (r = -.54, p < .001). We also examined a multiple regression model predicting 12-month infant sleep by 7-month maternal depression, including the interaction of depression with condition. We controlled for pretest variables collected during pregnancy: income, marital status, mother age, depression, and social desirability. We also included 12-month child health as a covariate. The model was significant (F (9, 134) = 8.20, p < .001, R² = .36). As hypothesized, the interaction between 7-month depression and intervention status was significant (β = .20, p < .05): Follow-up showed that only the control group showed a relationship between maternal depression and infant sleep (t = -3.49, p < .001).

Conclusions:

Results support the hypothesis that the intervention disrupted the link between maternal depression and infant sleep, buffering infant sleep from the negative impact of depression. As families are complex systems, it is crucial to examine how family interventions may also alter family processes other than only those specifically targeted by the program. It is also important to extend our scope from examining effects on the level of the outcomes to examining how our interventions may modify relations between risk factors and outcomes.