Much is known about risk factors for depression, but further research is needed to test models of how multiple biological and psychosocial factors work together and to clarify the mechanisms by which stressful experiences lead to depressive reactions in individuals and in the family context. Resiliency, despite exposure to parental depression and other adverse conditions is complex, and research will benefit from developmentally sensitive (i.e., age of the child) and integrative models that can be tested over a longitudinal course. We need to know more about optimal timing and methods of intervention to prevent the development and escalation of depression in those at greatest risk—especially young people during their formative family and career years.

Although strong evidence now supports the breadth and extent of associations between depression in parents and adverse outcomes in children, there remain many unanswered questions. In particular, many questions remain regarding mediation and moderation of those associations. In terms of mediation, more studies are needed to test specific aspects of parenting and other potential mediators of associations between depression in parents and child functioning. In this regard, the committee noted the strong potential of studies designed to test the effectiveness of interventions aimed at reducing the level of constructs that have been found to mediate associations between depression in parents and outcomes in children, for example, particular aspects of parenting. Such experimental designs can be strong tests of mediation.

In terms of moderation, more studies are needed to reveal which children of depressed parents are more or less likely to develop problems and which parents with depression are more or less likely to have problems with parenting. Moderators might include parent characteristics, including severity, duration, and impairing qualities of their depression, social context variables, child characteristics, and others. For example, the moderating roles of the child’s sex and ages at times of exposure are still not well understood, with findings suggesting that boys and girls may be affected differently depending on their ages at the times of exposures. More broadly, we need more studies to quantify percentages of children who are affected (with specific outcomes) and those who are not and what distinguishes them. The committee noted the potential knowledge to be gained by further studies that target interventions to subsets of children with greater or lesser risk (degree of presence of moderators) to determine whether interven-