|Home | About | Journals | Submit | Contact Us | Français|
Men are increasingly the heads of single parent households, yet are often excluded from child welfare research and practice. To better serve all families in the child welfare system, it is necessary to understand the impact of primary caregiving men on children’s wellbeing. In this study we investigated the longitudinal effects of primary caregiving fathers’ mental health and substance use on child mental health, and examined possible differences by child age and gender. Regression analyses were conducted with the sample of 322 youth living with a male primary caregiver at the first wave of data collection from the National Survey of Child and Adolescent Wellbeing-II (NSCAW-II). We found that father depression at baseline consistently predicted child mental health outcomes three years later, even after accounting for demographics and baseline child mental health. Surprisingly, fathers’ substance use did not predict child mental health, and interactions with child age and gender were not significant. Our findings are consistent with a small but growing literature suggesting that efforts to improve engagement of and attention to fathers within research, clinical and policy efforts are likely to be worthwhile.
Men are increasingly the heads of single parent households. In fact, in 2011 single fathers headed 8% of households1 – a nine-fold increase from 1960 (Livingston, 2013). Yet, fathers are often excluded from child welfare research and practice (Brown, Callahan, Strega, Walmsley, & Dominelli, 2009; Strega et al., 2008). While this gap is slowly decreasing, it still remains despite several national initiatives calling for improved efforts to engage fathers in parenting (e.g., National Fatherhood Initiative, Promoting Responsible Fatherhood and Strong Communities, and National Responsible Fatherhood Clearinghouse). Furthermore, fathers, who make up 8% of primary caregivers in the child welfare population, are especially understudied in the child welfare field (Bellamy, 2009). It is necessary to better understand the experiences of child welfare-involved children who are in the care of their father. The purpose of this study is to examine which father characteristics are most strongly associated with child mental health outcomes three years post child protective services investigation.
Child welfare-involved children are at high risk for mental health problems such as depression and anxiety (Burns et al., 2004; Gewirtz & August, 2008) and caregivers influence these child outcomes (Lovejoy, Graczyk, O’Hare, & Neuman, 2000) through multiple potential pathways. One such pathway is their own psychopathology, which can shape their children’s experiences and resulting mental health outcomes throughout development. Parents of child welfare-involved children experience mental health problems at rates greater than that of the general population. The past year incidence of major depression among adult males in the general population is 5.1% (SAMHSA, 2014). The rate among fathers identified as primary caregivers in the child welfare population is three times this (17.9%) (Ayer, Woldetsadik, Malsberger, Burgette, & Kohl, under review). Additionally, nearly one in ten men in the child welfare population has a substance abuse or dependence disorder (Ayer et al., under review). These numbers signify further risk to the already vulnerable child welfare population.
It is well-established that children of depressed mothers are at increased risk of internalizing and externalizing problems (Bender, 2010; Downey & Coyne, 1990; Kim-Cohen, Moffitt, Taylor, Pawlby, & Caspi, 2005; Lyons-Ruth, Wolfe, & Lyubchik, 2000; Weissman et al., 2006). While much less attention has been given to the role of paternal depression and children’s mental health, there is an emerging body of evidence demonstrating its contribution to child outcomes. In a large representative sample of US children living in two-parent households, researchers found that those living with depressed fathers were 70% more likely to have emotional or behavioral problems than those living with nondepressed fathers, after controlling for maternal depression (Weitzman, Rosenthal, & Liu, 2011). A meta-analysis found that overall, child externalizing problems were impacted similarly by mother and father psychopathology, but that child internalizing problems were most strongly related to mothers’ psychopathology (Connell & Goodman, 2002). However, the body of work meta-analyzed mostly consisted of studies on two-parent households and those with a female primary caregiver, with no studies on fathers in the child welfare system. Contrary to findings in the general population where women are at greater risk for depression compared to men (Eaton et al., 2012), a recent study found that there were no significant differences in the prevalence of depression among male and female primary caregivers in the child welfare system (Ayer et al., under review). In addition, the same study found no differences in the use of mental health services between these groups, counter to findings in the general population suggesting that fathers are less likely than mothers to seek medical care for themselves or for their children (Moore & Kotelchuck, 2004). Thus, it is likely that paternal depression is as detrimental to child mental health as maternal depression; however, this issue has not yet been explored in a child welfare population. Given the complex problems encountered by child welfare-involved families, including those with fathers in the role of primary caregiver, it is essential to better understand the impact of paternal depression on children’s mental health.
Even less is known about the role of paternal substance dependence and children’s mental health. Despite a call to build out this research (McMahon & Rounsaville, 2002), this remains a major gap in the literature. Given the well-known association between maternal substance dependence and child outcomes (America, 2001; Osborne & Berger, 2009), it is hypothesized here that a similar association will exist for paternal substance dependence.
The proposed study will identify father-related risk factors for child mental health using existing data from a longitudinal, national probability study of children and families investigated for CM (National Survey of Child and Adolescent Wellbeing-II; NSCAW-II). Specifically, this study answers the following research questions:
Secondary data analyses of the National Survey of Child and Adolescent Wellbeing-II (NSCAW-II) were conducted to determine whether baseline (i.e., wave 1) father depression and substance dependence predicted child mental health outcomes at 3-year (wave 3) follow up. Further, we assessed whether these associations differed by child age and gender. NSCAW-II is a longitudinal, national probability study of children and families investigated for CM (see (Dolan, Smith, Casanueva, & Ringeisen, 2011) for a detailed description of the study design). Demographic characteristics, as well as mental health and substance dependence needs of fathers have already been described using the NSCAW-II baseline data (Ayer et al., under review).
The NSCAW-II sample includes 5,872 children aged birth to 17.5 years old whose investigations were closed during a 15-month period beginning February 2008. There were 322 (5.5%) youth who lived in their home with a male primary caregiver (“father”) at baseline.
Demographic variables included caseworker report of number of days in out of home care between baseline and wave 3, whether the father was the perpetrator of the index maltreatment, severity of the index maltreatment, and type of maltreatment; child age, gender, and race (White, Black, Hispanic) based on combined child, father and caseworker reports; and father report on his age, education of high school or above, employment (employed vs. not employed), family income below the federal poverty level, and whether he is the child’s biological parent.
The parent-report Child Behavior Checklist (CBCL) (T.M. Achenbach & Rescorla, 2001) internalizing (32 items; baseline α=.99) and externalizing problems (35 items; baseline α=.99) raw scores were used to measure child mental health at baseline and at 3-year follow up (wave 3). The CBCL preschool version was used for youth ages 1.5 - 5 and the school age version for ages 6 – 18. The CBCL is a widely used quantitative and empirically based measure of psychopathology in children and adolescents. The preschool version contains 100 items and the school age version contains 120 items (e.g., “there is very little that he/she enjoys”) on which parents rate their child’s behavior in the preceding 6 months on a three-point scale, where 0=not true, 1=somewhat or sometimes true, and 2=very true or often true. The psychometric properties of the CBCL are strong and well established (T. M. Achenbach & Rescorla, 2000).
Primary caregiver depression was measured by the Composite International Diagnostic Interview-Short Form (CIDI-SF) depression module (Kessler, Andrews, Mroczek, Ustun, & Wittchen, 1998). The CIDI-SF is a diagnostic interview that screens for psychiatric disorders as per the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (Association, 2000). The CIDI-SF is valid and reliable, and classifies respondents as having major depressive disorder with 93% accuracy (Kessler et al., 1998). The CIDI-SF assessed past-year depression with stem-branch logic where fathers were first asked if there was a 2-week period during the past 12 months where they felt sad, blue or depressed. If the father answered “yes”, they were then asked additional questions about the worst 2-week period. A “probably depression” diagnosis was assigned if the father reported three or more symptoms of depression (as defined by the DSM-IV) in addition to experiencing two or more weeks of dysphoric mood, two or more weeks of anhedonia, or using antidepressant medication.
Primary caregiver substance dependence was measured by the Alcohol Use Disorders Identification Test (AUDIT; total score 8 or higher indicates alcohol dependence) (Babor, Higgins-Biddle, Saunders, & Monteiro, 2001) and Drug Abuse Screening Test (DAST-20; total score 6 or higher indicates drug dependence) (Skinner, 1982). These self-report measures have good psychometric properties (Yudko, Lozhkina, & Fouts, 2007). The AUDIT includes ten items, which ask respondents about alcohol use, alcohol dependence, and alcohol-related problems (e.g., “have you or someone else been injured as a result of your drinking?”). The DAST-20 contains 20 yes/no items, which measure problematic substance use (e.g., “are you always able to stop using drugs when you want to?”). Using the aforementioned clinical cutoffs, we created a dichotomous “any substance dependence” variable to encompass both drug and alcohol dependence.
We used IVEware version 0.2, imputation and variance estimation software developed by the Survey Research Center at the University of Michigan, to perform multiple imputation to replace missing values of covariates. Multiple imputation accounts for the uncertainty of the imputed values by creating multiple (in our case, 10) copies of the original dataset with missing variables replaced. Missing elements in one variable are imputed by estimating the distribution of the observed values of that variable, conditional on all other variables included in the analysis.
All analyses described below were then run in two steps: 1) Each of the 10 completed datasets were analyzed using the standard statistical procedures; 2) The 10 sets of parameter estimates and associated covariance matrices produced by the standard procedures were then analyzed using the MIANALYZE procedure. The MIANALYZE procedure combines these sets of results to calculate valid inferences for the parameters, accounting for both the within variance of each imputed value and between variance for each imputed dataset. All analyses were run with SAS version 9.4.
For each outcome measure, we used forward stepwise regression to identify the independent and demographic variables to retain in each model. The independent variables were dichotomous measures of baseline father depression and substance dependence, which we hypothesized to predict child internalizing and externalizing problems at 3-year follow-up. Covariates included number of days in out of home care between waves 1 and 3, baseline internalizing and externalizing, child age, child gender, child race, caregiver age, caregiver education, caregiver employment, family income, whether the father was the perpetrator of the index maltreatment, whether the father is the biological parent of the child, severity of the index maltreatment, and type of maltreatment.
The demographic characteristics of fathers and their children are shown in Table 1 and have been reported previously (Ayer et al., under review). At baseline, 17.0% of fathers met criteria for depression, and 7.5% of fathers met criteria for probable alcohol or drug dependence. A minority (2.5%) met criteria for both depression and substance dependence. The mean CBCL child internalizing score at baseline was 7.0 (SD=6.0, range=0–33), and the mean externalizing score at baseline was 10.7 (SD=9.3, range=0–45). The mean number of days the child was in out of home care between waves 1 and 3 was 44.0 (SD=131.8, range=0–900). Twenty-one percent of youth were in out of home care for one or more days between waves 1 and 3. The types of maltreatment leading to investigation at baseline were: physical abuse (27.4%), sexual abuse (8.2%), and neglect (36.7%). The levels of maltreatment severity as rated by caseworkers were: none (35.9%), mild (28.2%), moderate (24.2%) and severe (11.7%).
Table 2 shows the regression results for the model testing whether baseline father depression and substance dependence predicted wave 3 child internalizing after accounting for relevant covariates. Only covariates retained in the final regression model are shown in Table 2. These results suggest that even after accounting for baseline demographics, father depression predicts child internalizing problems three years later. Not surprisingly, child internalizing at baseline also predicted internalizing at wave 3. Father substance dependence was not a significant predictor of child internalizing problems, nor were the other covariates entered into the model. Child age and gender did not significantly predict internalizing, and regression models testing the interactions of baseline father depression and substance dependence with child age and gender revealed no significant interaction effects.
Table 3 shows the regression results for the model testing whether baseline father depression and substance dependence predicted wave 3 child externalizing after accounting for relevant covariates. Only covariates retained in the final regression model are shown in Table 2. Again, after accounting for baseline externalizing levels and other covariates, father depression predicted child externalizing three years later. Baseline father substance dependence was not a significant predictor of wave 3 child externalizing. The only significant covariates that predicted wave 3 externalizing were baseline externalizing, number of days in an out of home placement since baseline, and whether the family income was below the federal poverty line. As expected, baseline externalizing was positively associated with wave 3 externalizing; children with more days in out of home placements and those in families living below the federal poverty line also had higher levels of wave 3 externalizing problems. Child age and gender did not significantly predict externalizing, and regression models testing the interactions of baseline father depression and substance dependence with child age and gender revealed no significant interaction effects.
The growing body of literature on fathers has focused primarily on non-resident fathers and fathers in two-parent households, leaving families with male primary caregivers largely ignored. This study sought to fill this gap by examining the relationship between paternal mental health and child mental health among one of the country’s most vulnerable populations – those involved with the child welfare system.
We focused specifically on understanding the impact of paternal depression and substance dependence – two well-established maternal risk factors for child problems – on child mental health outcomes. We found father depression predicted both child internalizing and externalizing problems three years later, even after accounting for baseline levels of internalizing and externalizing problems and other covariates. This is consistent with previous studies on fathers in female-led households (Connell & Goodman, 2002; Weitzman et al., 2011). It also underscores the importance of attending to fathers’ psychological needs when conducting research or treatment focused on child outcomes.
Surprisingly, we found no significant relationship between father substance dependence and later child internalizing and externalizing problems after accounting for other predictors and covariates. This finding is inconsistent with literature on mothers, where maternal substance dependence is a key predictor of child outcomes, including in child welfare samples (Blanchard et al., 2005; Manly, Oshri, Lynch, Herzog, & Wortel, 2013). There are multiple potential explanations for this result. For one, the prevalence of drug or alcohol dependence among fathers in this study (7.5%), which is lower than the prevalence among male adults nationally (11.4%) (SAMHSA, 2013), may have limited our ability to detect an effect within our sample of 322 families. Given the low rate of co-occurrence of depression and substance dependence (2.5% in this sample), this is unlikely to explain the lack of a significant unique contribution of substance dependence above and beyond the effect of depression, but could be an area for exploration in future studies with larger samples. It is also possible that fathers’ substance dependence simply does not have a strong impact on child mental health outcomes for reasons that could be explored in future studies. For example, children with male primary caregivers may have access to other resources that serve as protective factors. While we found no difference between primary caregiving men and women in terms of whether another caregiver was present in the home (Ayer et al., under review), other supports outside of the home (e.g., teachers, neighbors) could buffer children from the negative effects of paternal substance use. The involvement of these external supports could be greater for children with a male primary caregiver given commonly held views that parenting is primarily a woman’s responsibility and thus fathers may need more assistance (Maxwell, Scourfield, Featherstone, Holland, & Tolman, 2012). If this is the case, however, it is unclear why depression affected child mental health but drug and alcohol dependence did not. More research is needed in order to better understand the role of father substance use on child mental health, and how it differs from mothers.
We also found that child age and gender did not significantly impact the relationship between baseline father depression or substance dependence and child internalizing and externalizing three years later. This could be explained by our relatively small sample, where we had few fathers endorsing substance dependence in particular and thus limited power to detect small effects. Thus, the lack of significant interaction findings here does not necessarily mean that the effect of father depression on child internalizing and externalizing is similar for children of different ages and across both genders; replication with larger samples is required to draw any such conclusions.
This longitudinal study is the first, to our knowledge, to examine the impact of caregiver depression and substance dependence within families with a male primary caregiver in the child welfare system. While the study can help to advance the field and hopefully spur additional investigations within this understudied population, there are a number of limitations to consider. We recognize that caregiver mental health contributes to child mental health through multiple mechanisms, including genetic, neuroregulatory, emotional, cognitive, behavioral, and environmental (Goodman & Gotlib, 1999). However, due to the availability of the data within NSCAW-II, we limited our measurement to the present or absence of self-reported depression or substance dependence, and did not include any mediation modeling to examine potential pathways through which parental mental health impacts child mental health. These pathways are worthy of rigorous exploration in future studies. We also relied on fathers’ self-report of depression and substance use. While fathers were assured their participation was confidential and their responses were not shared with anyone (e.g., child welfare workers), some may have underreported symptoms in these areas. Furthermore, as mentioned earlier, we had weak statistical power to detect effects, particularly interactive effects. Replication of these analyses in larger samples is necessary. There may also be other father-related predictors, or interaction variables, worthy of exploration in future studies. For instance, we did not examine other forms of psychopathology (e.g., anxiety disorders) or substance abuse behavior other than dependence (e.g., binge drinking, misuse/abuse). Relatedly, we combined alcohol and drug dependence into a “substance dependence” measure, whereas different types of substance use may reveal different relationships with child outcomes in larger samples. Future studies should also empirically test whether and how much - after accounting for demographic and other systematic differences (Ayer et al., under review) - the relationship between parental depression and substance dependence with child mental health outcomes differs in families with male versus female primary caregivers in the child welfare system.
Our findings that father mental health predicts child mental health three years later are consistent with a small but growing body of empirical and theoretical work (Connell & Goodman, 2002; Weitzman et al., 2011) suggesting that efforts to improve engagement of and attention to fathers within research, clinical, and policy endeavors are likely to be worthwhile. The U.S. Preventive Services Task Force recently released a statement recommending depression screening pregnant and postpartum women, in part due to the impact of maternal depression on children (Siu et al., 2016). Results from this study indicate that such screening standards should be applied to fathers in the child welfare system as well. Early detection and treatment of paternal depression using evidence based interventions such as antidepressant medication or cognitive behavioral therapy (CBT) may not only improve the mental health of the father, but also of his children (Weissman et al., 2006). Given the increasing presence of male primary caregivers in the child welfare system and in the general population (Livingston, 2013; Zanoni, Warburton, Bussey, & McMaugh, 2014), and the gap in knowledge about these families, this study underscores the need to acknowledge the impact of fathers on children’s wellbeing and attend to their mental health concerns.
Funding source: This study was supported by National Institutes of Health (NIH) grant R03MH101542 to the first author.
1Although male primary caregivers can include relatives other than fathers (e.g., grandfathers, uncles, stepfathers), we will use the term “fathers” to refer to male primary caregivers in this study for simplicity, and to reflect that male primary caregivers are in fathering roles regardless of biological relations to the child.
Financial Disclosures: The authors have no financial relationships relevant to this article to disclose.
Conflict of interest: The authors have no potential conflicts of interest to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.