Maternal Depression and Child Problems 1 Improvements in Maternal Depression as a Mediator of Intervention Effects on Early Childhood Problem Behavior Maternal Depression and Child Problems

Maternal depression has been consistently linked to the development of child problem behavior, particularly in early childhood. Despite this link, interventions typically address either adult depression, or child behavior, as separate foci. The current study examines the possibility that an intervention can address both maternal depression and child behavior problems in unison, using a sample of 731 high risk families receiving services from a national food supplement and nutrition program. Families with toddlers between ages 2 and 3 were screened and then randomized to a brief family intervention, the Family Check Up, which included linked interventions that were tailored and adapted to the families needs. Follow-up intervention services were provided at age 3 and follow-up of child outcomes occurred at ages 3 and 4. Following a previous report showing the intervention to be effective for reducing child externalizing problems, latent growth models also revealed intervention effects for early internalizing problems from 2 to 4, and reductions in maternal depression from ages 2 to 3. More importantly, reductions in maternal depression mediated improvements in both child externalizing and internalizing problem behavior. The results are discussed with respect to possibility of linking psychosocial preventive interventions for adult depression and early childhood adjustment problems within a comprehensive framework. Maternal Depression and Child Problems 3 Several types of parental psychopathology have been associated with increased risk of child psychopathology (Connell & Goodman, 2002; DelBello & Geller, 2001; Goodman & Brumley, 1990; Lapalme, Hodgins, & LaRoche, 1997). One of the most highly researched relationships of this type has been between maternal depression and different forms of child adjustment. This is not surprising as women more often serve as primary caregivers compared to men, and the incidence of depression is quite high among females beginning during adolescence. Moreover, being the primary caregiver of several young children is highly related to depression in western cultures (Strickland, 1992). The link between maternal depression and childhood problem behavior is unlikely to be incidental. Findings in the extant literature provide substantial evidence for a relation between maternal depression and negative child outcomes across development stages of childhood and adolescence, including both externalizing and internalizing child problem behaviors (for reviews of this literature, see Beardslee, Versage, & Gladstone, 1998; Cummings & Davies, 1994; Gelfand & Teti, 1990). These associations have been found to be particularly robust during early childhood when mothers and children spend more time together than at later ages (Marchand, Hock, & Widaman, 2002; Shaw et al., 1994, 1998). Despite the consistency of associations between maternal depression and child adjustment during early childhood, most intervention programs aimed at reducing child problem behavior have been explicitly focused on changing parenting practices rather than maternal depression per se. More recent versions of parenting programs for young children have included components dedicated to parental well being and social support (Baydar, Reid, & Webster-Stratton, 2003; Olds, 2002); however, the vast majority continue to focus on modifying caregiving practices (Brinkmeyer & Eyberg, 2003; WebsterStratton & Reid, 2003). Maternal Depression and Child Problems 4 A focus on parenting practices has substantial face and empirical validity, especially during early childhood when children undergo dramatic changes in cognitive and emotional development from infancy to the preschool period (Shaw & Bell, 1993) and elicit many challenges to caregivers (Fagot & Kavanagh, 1993). Related to parenting models, several theorists have noted how maternal depression might compromise a parent’s ability to be consistently and actively engaged with children and be attentive and responsive to their socioemotional needs (Belsky, 1984; Conger, Patterson, & Ge, 1995; Patterson, 1980), yet relatively few studies have directly examined whether child behavior might be improved by reducing maternal depressive symptoms, with the notable exception of research on parent management training among recently divorced families (DeGarmo, Patterson, & Forgatch, 2004; Patterson, DeGarmo, & Forgatch, 2004). The present study sought to address this issue by examining whether a parent-centered intervention to prevent the emergence of early conduct problems also proved successful in improving levels of maternal depressive symptoms, and if so, whether reductions in maternal depression mediated improvements in subsequent levels of child problem behavior. Maternal Depression and Child Adjustment The association between maternal depression and poor child outcomes is one of the most robust findings in psychological research (Gross, Shaw, & Moilanen, 2007). Both maternal clinical depression and sub-clinical, elevated levels of depressive symptoms have been found to be related to child maladjustment (Cummings, Keller, & Davies, 2005; Farmer, McGuffin, & Williams, 2002). As a result, the term maternal depression will be used throughout this paper to describe both criteria. Studies of children of depressed mothers across both narrowly-defined developmental periods and broad age spans (Goodman & Gotlib, 1999) have yielded consistent Maternal Depression and Child Problems 5 findings linking maternal depression to disruptions in both socio-emotional and instrumental functioning (Elgar et al., 2004; Gelfand & Teti, 1990; Hay, Pawlby, Angold, Harold, & Sharp, 2003; Leve, Kim, & Pears, 2005; Sinclair & Murray, 1998). These associations have been corroborated most consistently during early childhood, when maternal depression has been linked to fussiness and difficult child temperament (Cutrona & Trouman, 1986; Whiffen & Gotlib, 1989), insecure attachment (Campbell et al., 2004; Field et al., 1988), behavior problems (Marchand, Hock, & Widaman, 2002; Shaw, Keenan, & Vondra, 1994), and reduced mental and motor development (Murray, Fiori-Cowley, Hooper, & Cooper, 1996a; Sharp et al., 1995). Also, there are some data to suggest that elevated rates of maternal depression during the toddler years may be more predictive of later child adjustment problems than when assessed in the preschool period. For example, Shaw and colleagues (2000) found a direct link between maternal depressive symptoms when children were 1.5 and 2 years of age and clinically-elevated reports of school-based conduct problems (CP) when children were age 8 (d = .73 at age 1.5), associations that were appreciably stronger than parent reports of CP at ages 1.5 and 2. In addition, the magnitude of effects of maternal depression on age-8 CP decreased with the child’s increasing age (d = .27 when maternal depression was measured at age 5.5). As evidence links maternal depression during early childhood to subsequent child problem behavior, it follows that targeting changes in maternal depression during the toddler period might lead to reductions in later child problem behavior. The Toddler Years as a Period of Transition for Children and Parents The concept of socio-emotional transitions has been a focus for developmental and interventions scientists, as critical periods to understand, and possible opportunities for prevention (Sameroff, 1981; Cicchetti & Toth, 1995). The toddler period is one such Maternal Depression and Child Problems 6 developmentally critical transition for both parents and children. The toddler years represent a time of marked change for children in terms of cognitive, emotional, and physical maturation. Despite growth in all of these areas, children’s developing cognitive abilities are not well matched to the challenges afforded by their newfound physical mobility. Their new mobility permits children to ambulate quickly but without the cognitive appreciation to anticipate the consequences of violating other’s personal space, understanding the principles of electricity or gravity, or considering the potential hazards of straying too far from caregivers in novel settings (e.g., shopping malls). Thus, toddlers require proactive involvement and monitoring to literally keep them out of harm’s way (Gardner, Sonuga-Barke & Sayal, 1999). For parents dealing with this transformation (Shaw, Bell, & Gilliom, 2000), the nature of the parent-child relationship changes from a focus on responsivity and sensitivity to the immobile infant’s emotional needs to monitoring a mobile and naive toddler. As a result, parental pleasure in childrearing has been shown to decrease from the first to second years (Fagot & Kavanagh, 1993). Previous research suggests that how caregivers respond to these changes and how involved they are during this period has been shown to have important repercussions for early CP (Gardner et al., 1999; Shaw et al., 2000; 2003), as the course of CP has been shown to be moderated by controlling, uninvolved, and rejecting parenting (Aguilar et al., 2000; Campbell et al., 1996). As noted above, similar associations have been found between maternal depression and subsequent child CP, and several studies have explicitly attempted to address post-partum maternal depression during infancy in the hopes of improving the quality of the parent-child relationship (e.g., Murray, Cooper, Wilson, & Romaniuk, 2003). It would seem prudent to design interventions that address these unique demands as a preventive strategy to reduce child mental health problems such as CP and internalizing Maternal Depression and Child Problems 7 symptoms. Examples of successful preventive interventions that target early childhood include Olds’ (2002) Nurse-Family Partnership for first-time parents with newborns, Webster-Stratton’s Incredible Years Program (Baydar et al., 2003) for children approaching formal school entry, and Dishion’s Family Check Up (FCU, Shaw et al., 2006), which has also been previously applied to adolescent populations (Dishion & Kavanagh, 2003). To date, few interventions initiated in early childhood have specifically examined whether reductions in maternal depressive symptoms are a potential mechanism underlying improvements in early child problem behavior. Where such changes substantially account for the intervention effect, then maternal depression would qualify as a mediating mechanism (Kraemer, Wilson, Fairburn, & Agras, 2002; Rutter, 2005). To fill this void, the purpose of the current study was to examine the efficacy of a family-centered intervention in improving maternal depression and test whether such changes if found, accounted for reductions in both child externalizing and internalizing problem behaviors. Although the study was designed to specifically target child CP, we also tested the same issues with child internalizing problems to evaluate the generalizability of the intervention’s effects. To assess the viability of this approach, we selected 731 families with toddlers participating in Women, Infants and Children Nutritional Supplement Program (WIC) service systems in urban (Pittsburgh, PA), suburban (Eugene, OR), and rural (Charlottesville, VA) locations. Toddlers were screened to be at risk for showing early-starting pathways of CP and those meeting eligibility requirement were subsequently randomly assigned to the intervention condition. Follow-up results on maternal depression and child CP and internalizing problems were available one and two years after initial contact. Barriers to Family Interventions and the Family Check Up One of the barriers to implementing family interventions within service settings is parent Maternal Depression and Child Problems 8 motivation (Dishion & Stormshak, 2006). Many of the efficacy and effectiveness trials that form the backbone of our empirical literature are based on high levels of funding for both the research component and the families’ participation. For example, Dishion and colleagues developed the Adolescent Transitions Program (ATP), which comprised 12 parent group sessions that emphasized family management practices. Within the context of a clinical trial, the intervention reduced observed coercive parent-adolescent interactions, decreased antisocial behavior and subsequent substance use (Dishion, Andrews, Kavanagh, & Soberman, 1996). However, in implementing the program outside the context of a well funded intervention trial, parent participation was difficult to obtain, and dependence on parent groups as the exclusive delivery format deemed impractical. Stormshak and colleagues (2002) also report difficulty relying exclusively on parent groups with a set format and agenda as a barrier to changing parenting practices. The problem is that many of our family-centered interventions do not explicitly target the parents’ motivation to change. Parent resistance to change has been programmatically studied by investigators who who study parent management training (e.g., Patterson & Chamberlain, 1994; Patterson & Forgatch, 1985). Forgatch and colleagues (2006) found that training therapists to effectively address families’ resistance to change was prognostic of positive intervention outcomes (Forgatch, DeGarmo et al., 2005). In general, therapist training in developing collaborative relationships with parents and working through motivation issues in therapy is a key to the change of parenting practices. This is even true when parents voluntarily come to the treatment settings and request help, such as in the studies of parent management training. Miller and colleagues have developed the technique of motivational interviewing to encapsulate the therapist-client dynamics that are most likely to result in productive change. For Maternal Depression and Child Problems 9 example, in the field of alcohol misuse, The Drinkers Check Up is a direct application of motivational interviewing designed to promote change in adults who drink heavily (Miller & Rollnick, 2002). Two of the key strategies of the Drinker’s Check Up are to use assessment data in a feedback interview to elicit interactions between the client and therapist that influence change, and provide a flexible menu of change strategies for the client to select to achieve reductions in drinking. Several studies reveal that random assignment to the brief Drinkers Check Up was as effective as 28 days of costly inpatient treatment for reducing problem drinking in adults (Miller & Rollnick, 2002). The Family Check Up (FCU) was directly inspired by the work of Miller and colleagues on motivational interviewing. The FCU is a brief intervention that contains a broad assessment of the family context and parenting practices, an initial get-to-know-you meeting with the family, and a formal feedback session. We see the FCU as the foundation of an ecological approach to child and family interventions, the first step in a menu of empirically-supported child and family interventions that reduce problem behavior and promote emotional well being in children and families. In contrast to the standard clinical model, the ecological approach is seen as a health maintenance model, which explicitly promotes periodic contact with families (at a minimum yearly) over the course of key developmental transitions. The current study focuses primarily on the FCU for families and toddlers at-risk for early CP engaged in the WIC service system. Previous research with the FCU involved random assignment of young adolescents in public middle schools to a family resource room in contrast to a ‘middle school as usual’ control condition. The family resource rooms were staffed by trained personnel focused on engaging families in the FCU and a variety of other linked family interventions (see Dishion & Kavanagh, 2003). Using an intention to treat design, the authors found that proactive parent engagement Maternal Depression and Child Problems 10 reduced substance use among high-risk adolescents, and prevented substance use among typically developing youth (Dishion, Kavanagh, Schneiger, Nelson, & Kaufman, 2002). Significant reductions in these problem behaviors resulted from, on average, six direct contact meetings with parents over the course of three years. Complier Average Causal Effect models support the notion that the FCU was the key intervention strategy, and that receipt of the FCU and linked services as needed lead to significant long-term reductions in substance use and antisocial behavior, including decreased substance use diagnoses and fewer arrests by the end of high school (Connell, Dishion, Yasui, & Kavanagh, in press). We have previously applied the Family Check Up to high risk families of toddlers involved in WIC, a national program for nutrition support for income eligible families with children ages 0 to 5. Randomly assigning 120 families of toddlers to WIC as usual, versus WIC with a Family Check Up at age 2 was found to result in reductions in problem behavior and improvement in parent involvement at ages 3 and 4, respectively (Shaw et al., 2006). In addition, intervention effects were evident for those families with a risk profile for early-starting CP, including aboveaverage levels of maternal depressive symptoms and child fearlessness. Those families assigned to the intervention group with this risk profile showed a sharp decline on child CP between ages 2 and 4 compared to families in the control group with same risk profile at age 2. This previous study of the FCU with families of toddlers was limited by a small sample size, the use of only male children recruited from an urban community, and the extent of intervention services offered to the families. The current study, which we refer to as the Early Steps Multisite Study (ESMS), remedies these three limitations and provides a broader perspective on possible mediating mechanisms of change. First, the sample size includes 731 atrisk families, half of whom were randomly assigned to the Eco FIT, versus WIC as usual. The Maternal Depression and Child Problems 11 families were recruited from three geographically and culturally unique regions, including metropolitan Pittsburgh, Pennsylvania suburban Eugene, Oregon, and rural Charlottesville, Virginia. The sample also reflects cultural diversity, including African American, European American, and Latino families. In addition, we utilized the entire Eco FIT model, in that families were provided additional services following the Family Check Up, consistent with an adaptive, tailored approach to intervention. Extending an earlier report on the ESMS that demonstrated the FCU to be associated with improvements in child CP and positive parenting (Dishion et al., 2007), the current study tests whether the intervention was also successful in reducing levels of maternal depressive symptoms and child internalizing problems, and whether changes in maternal depression mediated reductions in child CP and internalizing symptoms.