Does Parents’ Own History of Child Abuse Moderate the Effectiveness of the Promoting First Relationships® Intervention in Child Welfare?

Abstract

To better understand how and for whom parenting intervention may improve family outcomes in child welfare services, we examined whether parents’ own history of child abuse moderated the indirect effects of the Promoting First Relationships® (PFR) intervention on toddlers’ secure base behavior via parental sensitivity. Parents (N = 247) and their toddlers (10–24 months) involved with child protective services were randomized to PFR or a control intervention. Results showed that the PFR group demonstrated greater parental sensitivity at postintervention than the control group, which in turn led to higher levels of toddler secure base behavior at 6-month follow-up. Findings from a moderated mediation model indicated that these intervention effects were only evident for parents who experienced physical abuse in their childhood. Parents’ history of sexual or emotional abuse did not significantly moderate outcomes. These results provide evidence for a key mechanism of change in PFR informed by attachment theory and suggest that PFR intervention effects may be stronger in parents at higher risk of the intergenerational transmission of abuse.

Keywords

toddlers home visiting child protective services parental sensitivity attachment

Most parents involved with child protective services (CPS) carry their own history of adverse childhood experiences that can undermine the quality of the parent–child relationship and contribute to poor child outcomes. For instance, many CPS-involved parents have been abused as children (Widom, Czaja, & DuMont, 2015), and these early traumatic experiences can adversely impact their caregiving behavior and increase risk of the intergenerational transmission of child abuse (Dixon, Hamilton-Giachritsis, & Browne, 2005). Indeed, parents’ own history of child maltreatment predicts repeated referrals to CPS (Hindley, Ramchandani, & Jones, 2006), perpetuating CPS involvement across generations. To help break this cycle and promote child well-being, more understanding is needed on early intervention services that may benefit these vulnerable families.

Although the majority of parents with prior exposure to child abuse do not abuse their own offspring (Thornberry & Henry, 2013), they tend to experience greater caregiving difficulties during children’s early development. These challenges may be partly underpinned by insecure representations of early attachment relationships, and low reflective functioning about past traumatic experiences, that impair parents’ ability to identify and reflect on, and sensitively respond to, their child’s emotional needs (Berthelot et al., 2015; George, Solomon, Cassidy, & Shaver, 2008). Moreover, the neuropsychological consequences of early abuse experiences, involving deficits in the “social brain,” may also compound parents’ difficulties in providing emotionally attuned caregiving (DeGregorio, 2013). Ten of the 14 studies included in a recent review of observed caregiving toward infants and toddlers in parents with abuse histories showed a significant direct or indirect effect of physical and/or sexual abuse history on lower parental sensitivity (Vaillancourt, Pawlby, & Fearon, 2017). Mothers with these prior abuse experiences may fail to show the typical trajectory of improved sensitivity from infancy to toddlerhood (Fuchs, Möhler, Resch, & Kaess, 2015) and are more likely to have an insecurely attached child (Berthelot et al., 2015; Pasalich, Cyr, Zheng, McMahon, & Spieker, 2016). Furthermore, parents’ history of emotional abuse may also negatively impact parent–child interactions; however, this association has received less empirical attention (Hughes & Cossar, 2016). In sum, past abuse experiences in childhood appear to be a significant risk factor for a broad continuum of adverse caregiving behaviors, ranging in severity from a lack of sensitivity to children’s needs to abusive behaviors that necessitate CPS involvement.

Parenting interventions are increasingly being disseminated to prevent the occurrence, or reoccurrence, of child abuse via improving positive caregiving behaviors. The most promising model of parenting intervention for families with young children who have experienced, or are at risk of, abuse focuses on enhancing the caregiver–child relationship, particularly through home visits (Levey et al., 2017; Toth, Gravener-Davis, Guild, & Cicchetti, 2013). Promoting First Relationships® (PFR; Kelly, Sandoval, Zuckerman, & Buehlman, 2008) is a 10-week home-based intervention informed by attachment theory and was designed to strengthen caregivers’ awareness of and sensitive responding to their child’s social–emotional needs to promote the behavioral and emotional well-being of toddlers. Although PFR is similar to nurse home-visiting services (e.g., Olds et al., 1998) in terms of its delivery setting, it can also be distinguished by its shorter duration and more narrow focus on enhancing parental sensitivity and the parent–child relationship. To this end, PFR providers use video-based feedback with caregivers to help them reflect on how they and their child may be responding during interactions. A recent randomized controlled trial (RCT) of PFR in families referred to CPS, showed greater effects of PFR versus a control condition on improved caregiver sensitivity and understanding of toddler developmental needs, reductions in child atypical affective communication, and lower likelihood of entering foster care through 1-year postintervention (Oxford, Spieker, Lohr, & Fleming, 2016). Although these findings are encouraging in light of the complex challenges in working with CPS-involved families, interestingly, PFR did not demonstrate significant main effects on secure base behavior.

Through the lens of attachment theory, the impact of PFR on improving parents’ sensitive responding toward children’s emotional cues should subsequently shift the emotional dynamics of the parent–child relationship and, over time, promote toddlers’ felt security in the caregiver as a source of safety and comfort (Bowlby, 1969/1982). Thus, a test of PFR’s theoretical process of change is whether the program indirectly improves toddlers’ attachment security through strengthening caregivers’ sensitivity. Although Oxford et al. (2016) did not demonstrate a main effect of PFR on secure base behavior, from both a conceptual and statistical standpoint, the program may have an indirect effect on attachment security if it has a stronger, more proximal effect on parental sensitivity (i.e., the intervening variable; Bakermans-Kranenburg, van Ijzendoorn, & Juffer, 2003; Rucker, Preacher, Tormala, & Petty, 2011). Indeed, the two prior RCTs of PFR showed the strongest intervention effects on caregivers’ understanding of and sensitive responding to toddlers’ social–emotional needs (Oxford et al., 2016; Spieker, Oxford, Kelly, Nelson, & Fleming, 2012), consistent with the program’s focus on improving child developmental outcomes through directly enhancing parenting behavior.

In addition to examining how PFR may work, to inform a targeted or tailored approach to early intervention in vulnerable families, it is important to investigate for whom this proposed mechanism of change may benefit the most. Given their significant challenges in caregiving and higher risk of perpetuating child abuse, it is surprising that little research has investigated parenting intervention outcomes in parents with prior abuse experiences (Levey et al., 2017). In a home-visiting trial, parents’ history of abuse and neglect predicted worsening parenting stress at 18-month postenrollment via higher depressive symptoms and lower social support (Shenk et al., 2017). Although this finding suggests that parents with abuse histories may respond less favorably to home visiting, the lack of a comparison condition in the study precludes making inferences about whether these at-risk parents still benefit from receiving intervention services delivered in the home. This knowledge is vital for decision-making in policy and practice regarding families with complex needs related to child welfare concerns.

Findings from three prior RCTs help shed light on potential moderating effects of parents’ abuse history on intervention-induced improvements in parenting behavior. In a study of home-visiting outcomes in depressed mothers of infants, mothers’ prior experiences of physical abuse (but not other maltreatment types) moderated program effects on parenting (Ammerman, Peugh, Teeters, Putnam, & Ginkel, 2016). Specifically, mothers with these histories who were randomized to standard home visiting showed less sensitive parenting at pretreatment than their counterparts assigned to home visiting plus cognitive behavior therapy but made greater gains in this parenting domain across treatment and showed comparable levels at 3-month follow-up. Moreover, in a trial of a home-based intervention targeting sensitive caregiving in teen mothers, there was a trend toward weaker program effects on sensitivity in mothers with an unresolved state of mind regarding loss or abuse in childhood (Moran, Pederson, & Krupka, 2005). Finally, in an RCT of parent management training for preschoolers in Head Start, families of parents with and without an abuse history generally showed similar benefits following the treatment, despite parents with abuse histories demonstrating greater parenting deficits at preintervention (Hurlburt, Nguyen, Reid, Webster-Stratton, & Zhang, 2013). Importantly, there were larger intervention effects on nurturing/supportive parenting in parents with abuse histories.

Findings from two of the three studies just reviewed suggest that parenting behaviors—particularly responding to children’s emotional needs—in at-risk parents with abuse histories may improve following parenting intervention to reach levels at least comparable with their counterparts without abuse histories. However, this is a tentative conclusion based on only a few studies that examined interventions targeting different components of parent–child interaction. Only one of these studies (Moran et al., 2005) reported on changes in child outcomes following intervention. Thus, little is known about the mechanisms through which parenting intervention may potentially benefit the families of parents with experiences of child abuse. As parenting interventions often show stronger effects in at-risk families with greater deficits in specific parenting domains targeted by the program (e.g., Oxford, Marcenko, Fleming, Lohr, & Spieker, 2016; Theise et al., 2014), arguably parents with and without abuse histories may differentially respond to a relationally oriented parenting intervention that largely focuses on improving caregiver sensitivity.

This study aimed to contribute new understanding into how parenting intervention outcomes in early childhood may vary between vulnerable families of parents with versus without a history of significant abuse. To this end, extending on prior results from an RCT with families referred to CPS (Oxford et al., 2016), we tested whether parents’ abuse history moderated a theory-driven process of change in PFR, involving indirect program effects on toddler secure base behavior via parental sensitivity. Considering that parents with an abuse history tend to show significant challenges in sensitive caregiving (Vaillancourt et al., 2017), we expected them to demonstrate greater improvements in this parenting domain from participating in PFR versus a brief resource and referral (R&R) service (control condition) than parents without an abuse history. Accordingly, we predicted that the impact of PFR on enhanced parental sensitivity, and, in turn, increases in secure base behavior (Hypothesis 1), would be evident or more pronounced in the families of parents with versus without an abuse history (Hypothesis 2). Given the prevalence of polyvictimization in survivors of abuse (Dong et al., 2004), we controlled for co-occurring forms of abuse to examine the specificity of abuse types as possible moderators of the process of change model tested in this study. Our focus on specific abuse dimensions was motivated by the need for continued investigation into unique effects of child abuse types, given that prior results suggest differential associations between forms of child abuse and adult outcomes (Widom, 2017) such as parenting behavior (e.g., Barrett, 2009; Lyons-Ruth & Block, 1996). Moreover, histories of physical and/or sexual abuse may be most strongly associated with lower parental sensitivity (Vaillancourt et al., 2017), and a history of physical abuse in particular has been shown to moderate home visitation effects on parenting (Ammerman et al., 2016).

Method

Sample

We used data from the Supporting Parents Program (SPP; Oxford et al., 2016). With approval from the Washington State Institutional Review Board, SPP enrolled 247 birth parents with infants and toddlers between January 2011 and January 2014. Participant recruitment, random assignment to intervention condition, and completion of assessments, as well as the sociodemographic composition of the sample (reported for both intervention conditions), are described in detail in Oxford et al. (2016). For study eligibility, parents (Mage = 26.73 years, SD = 5.72) needed to be conversant in English, have housing, and live in Snohomish, Southern Skagit, or Northern King County in Washington State. Furthermore, they had to have a toddler aged between 10 and 24 months (Mage = 16.37 months, SD = 4.46) and an open case with an allegation of any type of maltreatment recorded in the database of one of the six collaborating CPS offices at least 2 weeks prior to initial recruitment contact. Over 90% of the parents in the study were mothers. Most (77%) parents were White and 10% reported being of mixed race, while 31% of the children were identified as being biracial. Nineteen percent of parents and 32% of children were Hispanic. Most families were low income, 79% receiving food stamps, and 54% were headed by only one parent. Study families were compensated between US$50 and US$100 after each research visit and could receive a cumulative total of US$300 for completing all four research visits.

Interventions

PFR

Parent–child dyads (n = 124) randomized to the PFR intervention were offered 10 weekly 60- to 75-min in-home visits by one of the two PFR providers from a community agency who had a master’s degree in social work or counseling. Training of providers and procedures to ensure implementation fidelity are described in detail in Oxford et al. (2018). Attachment theory–informed and strength-based consultation strategies focused on increasing parental sensitivity to child cues. Reflective video feedback was included in five sessions using taped episodes of caregiver–child play or caregiving behavior, wherein the PFR provider guided discussion concentrating on parenting strengths and interpretation of the child’s cues. Eighty-six percent of families received all 10 sessions, and the mean number of weeks to complete the intervention for these families was 14.3 (SD = 5.4).

R&R

Participants (n = 123) randomized to the comparison condition received an R&R intervention delivered over the phone in three sessions. A social service provider conducted a 30-min needs assessment, mailed a packet of personalized information, and followed up with two 10-min check-in calls. A resource list of over 150 local services was mailed to each family. Based on the initial assessment of family needs, the social service provider included an average of six specific resources for each individual family. The three contacts were spaced across approximately 14 weeks to mirror the average length of the PFR intervention. Eighty-nine percent of families in the R&R condition received all three telephone sessions.

Measures and Procedure

The parent–child dyad who enrolled in the study was assessed at four time points: baseline, postintervention, and 3- and 6-month follow-up. All 247 dyads completed the baseline assessment. Completion rates were 91% at postintervention, 87% at 3-month, and 85% at 6-month follow-ups. These completion rates reflect both the child staying with the birth parent home (i.e., not experiencing a child welfare services removal) and completion of the follow-up assessments. Completion rates did not significantly differ by intervention condition or any sociodemographic variable examined. Assessment visits included interviewer-administered questions, self-report questionnaires, videotaped caregiver–child interactions of free play, a teaching task, and a brief separation procedure. The mean number of weeks spent completing the intervention did not differ between the conditions, but the number of weeks between baseline assessment and the postintervention assessment was greater for dyads in the PFR condition, PFR: M(SD) = 18.47 (6.82); R&R: M(SD) = 16.64 (3.51); t (177) = 2.54, p < .05, because it took longer to schedule and complete the first PFR in-home intervention session compared to the first R&R phone session. Thus, we included time between baseline and postintervention follow-up as a covariate in analyses.

Parents’ abuse history was assessed using the Childhood Trauma Questionnaire–Short Form (CTQ-SF; Bernstein et al., 2003), which was administered at baseline assessment and filled out by parents on their own. The CTQ included 25 items that ask about different forms of maltreatment experienced before the age of 18. Each item had a 5-point response option ranging from never to very often true. The CTQ dichotomous cutoff scores were used to measure significant (i.e., “moderate to severe”) levels of physical, sexual, and emotional abuse (Bernstein et al., 2003).

We used a modified total score on the Nursing Child Assessment Teaching Scale (NCATS; Barnard, 1994) to measure parental sensitivity. The NCATS instrument uses scoring of a videotaped interaction to assess caregiver sensitivity, stimulation of the child, and emotional responsiveness and contingent interactions between the parent and child. The scale was modified from the original measure, excluding items with low variability. A total score was based on a sum of 45 items scored yes (1) or no (0) (α = .68–.72). A blinded coder was trained to reliability by a certified NCATS instructor and passed regular reliability checks.

Secure base behavior was measured with the Toddler Attachment Sort-45 (TAS-45; Kirkland, Bimler, Drawneek, McKim, & Schölmerich, 2004). Four research study visitors were trained over the course of the study to score the TAS-45 by taking into consideration the child’s behavior during home visits. The visits included a separation procedure in which the caregiver was instructed to play with the child for a minute, say goodbye, leave the home for 3 min, and then return. The TAS-45 includes 39 items from the Attachment Q-Sort (Waters, 1987), plus 6 additional items concerning atypical affective communication. After research visits, research visitors sorted cards for 45 descriptive statements of child attachment behavior into five piles representing “most like” to “least like” the child, with a maximum of 18 cards placed in one pile. Item rankings were then used to calculate an overall secure base behavior score. To assess reliability across the project, 88 videotaped sessions were coded by pairs of raters. Interrater reliability was r = .75 for secure base behavior.

Analysis Plan

The analysis was conducted in Mplus 7.4 (Muthén & Muthén, 2015) using a series of steps. First, we examined change trajectories for parental sensitivity across the three time points after intervention (i.e., postintervention, 3-month follow-up, 6-month follow-up) using growth modeling with two growth factors: intercept (reflecting scores at postintervention) and linear slope (the rate of change). Second, we examined the indirect effect of intervention condition on secure base behavior at 6-month follow-up via parental sensitivity after the intervention (Hypothesis 1). Indirect effects computed as product of coefficients were reported, together with 95% bias-corrected bootstrap confidence intervals (CI) with 5,000 bootstrap draws which takes into account the nature of asymmetrical distribution of the indirect effect (MacKinnon, 2008). Estimates of the indirect effects were deemed significant if the 95% CI does not include zero. The product of coefficients approach tests jointly the effects of indirect paths and is not contingent on a significant association between the independent variable (intervention condition) and the outcome (secure base behavior; MacKinnon, 2008; Rucker et al., 2011). For instance, intervention condition may exert a stronger effect on parental sensitivity than secure base behavior, which could lead to a stronger indirect effect than total effect (total effect is the sum of indirect and direct effects). The standardized indirect effect can be used to interpret the size of the mediation effect (Preacher & Kelley, 2011) using Cohen’s (1988) standards of effect sizes, where small effect size = .01, medium effect size = .09, and large effect size = .25 (an indirect effect is a product of two effects, thus Cohen’s typical effect size standards are squared).

Third, we employed conditional process analysis—using Hayes’s (2013) PROCESS macro—to examine whether parental abuse history moderated the indirect effect described above (Hypothesis 2). Conditional process analysis is a conceptual and analytic approach to examining whether the strength of indirect effects is conditional on levels of a moderator and combines principles of mediation and moderation analysis. Parents’ prior exposure to each type of child abuse (i.e., physical, sexual, and emotional abuse) was included as a moderator of the association between intervention condition and parental sensitivity in separate models, while controlling for the effects of the other two abuse variables.

All models were estimated using full information maximum likelihood with robust standard errors, which provides estimates of the variance–covariance matrix for all available data, including those individuals who have incomplete data (assumed to be missing at random) on some measures. Data for postintervention and follow-up assessments come from visits with parent–child dyads that remained intact since baseline. Less than five cases had missing data at baseline and there was minimal missing data (2%) on particular measures for intact dyads that completed follow-up assessments. In terms of evaluating model fit, a nonsignificant χ², root mean square error of approximation (RMSEA) less than .05, and comparative fit index (CFI) greater than 0.90 were considered an adequate fit to the data (Hu & Bentler, 1999). In addition to covarying the effects of types of abuse, consistent with the prior analysis examining main effects of PFR (Oxford et al., 2016), we included the following covariates in the models: child age, months between baseline and postintervention assessment, and baseline measures for parental sensitivity and secure base behavior. The intervention conditions did not significantly differ on sociodemographic variables (see Oxford et al., 2016).

Results

Descriptive Statistics

Table 1 shows descriptive statistics and bivariate correlations among the main study variables (these data were not reported in Oxford et al., 2016). Intervention condition (R&R = 0; PFR = 1) was significantly negatively associated with parental sensitivity scores at baseline and parental history of moderate to severe sexual abuse. Parental sensitivity scores were significantly associated at all-time points. Parental sensitivity scores at baseline and the three postintervention time points were positively associated with levels of secure base behavior at 6-month follow-up. Secure base behavior at baseline was positively associated with parental sensitivity at 3-month follow-up and at 6-month follow-up. Parental history of moderate to severe physical abuse was negatively associated with parental sensitivity at 6-month follow-up. There were significant associations among parents’ reports of the three types of abuse.

Table 1.

Descriptive Statistics and Intercorrelations for the Key Study Variables.

Variable	M (SD) or %	1	2	3	4	5	6	7	8	9
1. R&R = 0; PFR = 1	—	—
2. Parental sensitivity (T1)	36.01 (4.54)	−.13*	—
3. Parental sensitivity (T2)	35.37 (4.94)	.06	.23***	—
4. Parental sensitivity (T3)	36.25 (4.83)	.04	.20**	.25***	—
5. Parental sensitivity (T4)	36.38 (5.05)	.11	.30***	.14*	.25***	—
6. Secure base behavior (T1)	0.11 (0.09)	−.01	.12	.10	.19**	.08	—
7. Secure base behavior (T4)	0.15 (0.08)	.08	.14*	.19**	.25***	.29***	.39***	—
8. Physical abuse history	20%	−.12	−.07	−.13	−.11	−.17*	−.09	−.09	—
9. Sexual abuse history	14%	−.13*	.00	−.06	.05	−.05	.02	−.07	.53***	—
10. Emotional abuse history	28%	−.09	.00	−.09	−.10	−.06	−.08	−.10	.61***	.46***

Note. Range: n = 201–247. R&R = resource and referral; PFR = promoting first relationships; T1 = baseline; T2 = postintervention; T3 = 3-month follow-up; T4 = 6-month follow-up; abuse history = parental history of moderate to severe levels of child abuse.

*p < .05. **p < .01. ***p < .001.

Proportions of parents who reported abuse histories are displayed in Table 1. Overall, 33% (n = 81) of parents reported a history of one or more type(s) of abuse in the moderate to severe range, including the following: physical, sexual, and emotional abuse (n = 22); physical and emotional abuse (n = 18); physical and sexual abuse (n = 3); sexual and emotional abuse (n = 5); physical abuse only (n = 6); sexual abuse only (n = 5); and emotional abuse only (n = 22).

Growth Model of Parental Sensitivity

An unconditional growth model of parental sensitivity after the intervention was first fit to examine individual differences in the intercept (with fixed factor loadings of 1) and linear slope (coded 0, 1, and 2). Both the intercept (M = 35.49, SE = .31, p < .001) and the slope (M = .50, SE = .22, p = .026) differed from zero, showing that, on average, parents increased in sensitivity levels across the three postintervention time points. The intercept had significant variability (Var = 8.07, SE = 3.38, p = .017), indicating that parents differed in their levels of sensitivity immediately postintervention. Variability for the slope was not significant (Var = 2.39, SE = 1.88, p = .21), suggesting that parents changed in their levels of sensitivity over time at approximately the same rate. The growth model of parental sensitivity after intervention, with a random intercept and fixed linear slope (in other words, we freely estimated variance around the intercept but fixed the variance of the slope at 0), provided the best fit to the observed data, χ²(3) = 2.76, p = .43, CFI = 1.00, RMSEA = .00.

Indirect Effect of PFR on Secure Base Behavior via Parental Sensitivity

We tested our first hypothesis by examining the indirect effect of intervention condition on secure base behavior at 6-month follow-up via parental sensitivity at postintervention. As there was no significant variability in the slope for sensitivity, predictors were modeled only to the intercept (reflecting scores immediately postintervention). The model provided satisfactory fit to the data, χ²(19) = 26.41, p = .12, CFI = 0.93, RMSEA = .040, and explained 48% and 36% of variance in the intercept for parental sensitivity and scores for secure base behavior at 6-month follow-up, respectively. The total effect of intervention condition on secure base behavior at 6-month follow-up was not statistically significant, B(SE) = 0.020(0.010), 95% CI [−0.002, 0.039], in line with prior results from the Oxford et al. (2016) study. In terms of direct effects, intervention condition (0 = R&R; 1 = PFR) was positively associated with parental sensitivity at postintervention, B(SE) = 1.14(0.43), p = .008, β = .25, and higher levels of parental sensitivity were associated with greater scores on secure base behavior at 6-month follow-up, B(SE) = 0.020(0.008), p = .014, β = .57. Intervention condition was not significantly associated with secure base behavior at 6-month follow-up, B(SE) = −0.003(0.014), p = .82, β = −.02. There was a significant indirect effect of PFR (vs. R&R) on greater secure base behavior via higher levels of parental sensitivity, B(SE) = 0.023(0.019), 95% CI [0.004, 0.068].

Regarding associations with the covariates, postintervention parental sensitivity was significantly positively associated with child age (B = 0.13) and baseline scores for parental sensitivity (B = 0.25) and secure base behavior (B = 5.09). Furthermore, levels of secure base behavior at 6-month follow-up and baseline were significantly associated (B = 0.23).

Indirect Effect of PFR on Secure Base Behavior via Parental Sensitivity as Moderated by Parental Abuse History

To test our second hypothesis, we examined the magnitude of the indirect effects of intervention condition on secure base behavior at 6-month follow-up, through parental sensitivity at postintervention, as a function of parental history of abuse. Regarding the moderating effects of physical abuse, the conceptual model for the conditional process analysis and unstandardized path estimates with 95% CI are shown in Figure 1. The model provided a good fit to the data, χ²(24) = 19.30, p = .74, CFI = 1.00, RMSEA = .00. The interaction term between intervention condition and physical abuse history significantly predicted postintervention parental sensitivity, B(SE) = 1.65(0.84), p = .049, β = .20. As illustrated in Figure 2, in parents with a history of physical abuse, those randomized to PFR versus R&R demonstrated greater levels of parental sensitivity at postintervention. The conditional indirect effect of PFR on secure base behavior via parental sensitivity was significant in the “moderate to severe physical abuse” group, B(SE) = 0.056(0.026), 95% CI [0.007, 0.126], but not in the “no to minor physical abuse” group, B(SE) = 0.017(0.011), 95% CI [−0.005, 0.071].

Figure 1.

Conceptual model showing unstandardized estimates and 95% confidence intervals (CI) for the primary tested paths and 95% CI for the conditional indirect effects of intervention condition on secure base behavior through parental sensitivity, moderated by parental physical abuse history. Significant paths are indicated with solid lines and nonsignificant paths are represented by dashed lines. Covariates are not shown in the figure and include child age, months between baseline and postintervention assessment, baseline measures for parental sensitivity and secure base behavior, and parental history of sexual and emotional abuse. R&R = resource and referral; PFR = promoting first relationships.

Figure 2.

Parents’ history of physical abuse as a moderator of PFR intervention effects on parental sensitivity at postintervention. The y-axis has been adjusted to focus on the group differences. R&R = resource and referral; PFR = promoting first relationships.

Results from the conditional process model testing the moderating effects of sexual abuse history, model fit statistics: χ²(24) = 20.99, p = .64, CFI = 1.00, RMSEA = .00, demonstrated a nonsignificant interaction term between intervention condition and sexual abuse history in predicting parental sensitivity, B(SE) = 0.45(0.95), p = .64, β = .04. Similarly, results from the conditional process model examining emotional abuse history as the moderator, model fit statistics: χ²(24) = 19.64, p = .72, CFI = 1.00, RMSEA = .00, showed a nonsignificant interaction between intervention condition and emotional abuse history, B(SE) = 0.52(0.81), p = .52, β = .07.

Follow-Up Analyses

In a follow-up analysis, we reran the conditional process analysis for the moderating effect of physical abuse without controlling for the effects of sexual and emotional abuse and observed similar findings. We also examined the same conditional process model but with parents’ amount of exposure to multiple types of moderate to severe levels of abuse (physical, sexual, and/or emotional) as the moderator. This moderator variable (“multiple abuse”) was coded as 0 = no abuse, 1 = one type of abuse, 2 = two types of abuse, and 3 = three types of abuse. Estimation of this model provided a good fit to the data, χ²(20) = 15.47, p = .75, CFI = 1.00, RMSEA = .00. The interaction effect—Intervention Condition × Multiple Abuse—was not significant, B(SE) = 0.40(0.34), p = .25, β = .12.

Discussion

Past research demonstrates that parents’ own experiences of child abuse impacts the quality of their current interactions with their offspring and may predispose risk of the intergenerational transmission of abuse. Accordingly, parents with abuse histories appear to be a particularly vulnerable subpopulation of parents in child welfare who require parenting services. However, little research has investigated whether these parents respond differently to parenting intervention (Levey et al., 2017). This study addresses this gap by examining the moderation of parents’ abuse history on an attachment-informed process of change in the PFR intervention in families involved with CPS. Our findings show how PFR leads to improvements in the parent–child relationship and suggest that this change process may best account for intervention benefits in families of parents with a physical abuse history. These first findings regarding how the families of parents with and without an abuse history differentially respond to a home-based parenting intervention cast new light on approaches to targeting and tailoring early intervention services for vulnerable families.

Consistent with our first hypothesis, PFR indirectly influenced secure base behavior at 6 months via impacting parental sensitivity at postintervention. The effect of PFR on improved parental sensitivity is in line with the previously reported project results in the Oxford et al. (2016) study, which also found that PFR did not have a significant main effect on secure base behavior. The present findings help clarify the nature of PFR effects on the quality of the parent–child relationship by demonstrating that PFR predicts improvements in parental sensitivity, which in turn lead to increases in toddlers’ secure base behavior. Importantly, by influencing parental sensitivity—via the PFR program—and observing subsequent shifts in attachment quality, this study provides further experimental evidence for a major developmental process in attachment theory (Bakermans-Kranenburg et al., 2003). Moreover, this finding supports a key mechanism of change in PFR and, more broadly, helps fill a significant gap in understanding how evidence-based parenting interventions operate for high-risk families.

Findings from the moderated mediation analysis contribute new understanding into a subpopulation of CPS-involved families that may benefit the most from this process of change in PFR. Consistent with our second hypothesis, PFR-induced improvements in parental sensitivity leading to improved secure base behavior were only evident in parents with a history of physical abuse. This is most likely due to the relatively greater improvements in parental sensitivity observed in parents with a physical abuse history who participated in PFR versus R&R. Moreover, following the intervention, their levels of sensitivity were almost comparable with observed levels in parents without a physical abuse history. This is encouraging given recent results suggesting that mothers with a history of abuse and neglect may be less responsive to home visiting (Shenk et al., 2017). Here, we observed the lowest levels of parental sensitivity at postintervention in the group of parents with a physical abuse history who were assigned to R&R. Furthermore, parents’ prior exposure to sexual and emotional abuse did not significantly interact with intervention condition in predicting postintervention sensitivity.

Our moderation findings are consistent with a prior study that found that only parents’ history of physical abuse, and not histories of other forms of maltreatment, moderated the impact of home visiting on parenting outcomes (Ammerman et al., 2016), and highlights the importance of examining the specificity of child abuse types in predicting adult outcomes (Widom, 2017). Although physical abuse, and not sexual or emotional abuse, significantly moderated the indirect effect of PFR, it is important to note that the majority of parents (88%) in this study who reported been physically abused as a child also reported exposure to moderate to severe levels of emotional and/or sexual abuse. Therefore, our findings are consistent with those from past studies in suggesting that physical abuse commonly co-occurs with other abuse types (e.g., Dong et al., 2004). However, we can have more confidence in the unique impact of physical abuse history on PFR outcomes, given that we controlled for the potentially confounding effects of co-occurring abuse types and that follow-up analysis showed that the indirect effects were not moderated more broadly by parents’ past experiences of multiple types of abuse.

Interestingly, although our bivariate associations demonstrated that physical abuse history predicted less parental sensitivity at 6-month follow-up, we did not find a significant association between these variables at preintervention. Given that parents with an abuse history fail to show the typical trajectory of improvement in parental sensitivity from infancy to toddlerhood (Fuchs et al., 2015), our findings may reflect this phenomenon in showing that physical abuse history more strongly predicts lower sensitivity as children enter toddlerhood (more children in this sample were toddlers at 6-month follow-up). Toddlerhood is marked by a rapid increase in children’s exploratory behavior and desire for autonomy, and parents with an abuse history might be challenged by this developmental change if it triggers a sense of loss of control or dysfunctional attributions regarding the child’s emerging independence (Fuchs et al., 2015). It is possible that PFR helps parents with abuse histories overcome some of this difficulty by enhancing their understanding of normative toddler behavior (e.g., motivations for exploration) and shaping more developmentally appropriate expectations regarding toddlers (Oxford et al., 2016). In this light, our moderation results suggest that PFR might protect against an atypical developmental trajectory of sensitivity in parents with a physical abuse history (recall that they showed levels comparable with the nonabused group) rather than boost levels of sensitivity relative to other CPS-involved parents.

This study’s results meaningfully add to the growing evidence base for PFR and home-visiting interventions for young at-risk children more generally. Findings from a prior RCT of PFR in toddlers who had recently entered a new foster/kinship placement or were reunified (Spieker et al., 2012) indicate potential benefits of PFR for reducing risk of poor behavioral outcomes in children who have experienced unstable placements. Specifically, a maladaptive cascade from a history of multiple placement changes to less secure base behavior to greater externalizing problems was mitigated in children whose caregivers participated in PFR versus a control intervention (Pasalich, Fleming, Oxford, Zheng, & Spieker, 2016). The current findings converge with, and extend, these past results, by demonstrating more potent effects of PFR in another high-risk subpopulation of families involved in child welfare—those headed by parents with a physical abuse history. Preventive interventions often show greater benefits in families with more severe levels of child and/or parent problematic functioning, arguably because the intervention effectively targets the particular domains of deficient or disrupted functioning (Cassidy, Woodhouse, Sherman, Stupica, & Lejuez, 2011; Theise et al., 2014). In addition to examining how parenting interventions may produce stronger effects in high-risk families through its targeted effects on improving domains of poor functioning, future studies should also investigate how these interventions may leverage families’ existing strengths or activate protective factors (e.g., social support), to help facilitate change in parent and child outcomes.

The current findings should be understood in the context of this study’s noteworthy strengths and limitations. The data are from an RCT of PFR in a “real-world” setting, involving community social service providers who were trained in implementing PFR in families’ homes. Moreover, outcomes were assessed via multiple informants using questionnaire and observational methods that used independent coders for parental sensitivity and toddler secure base behavior. Notwithstanding these strengths, given the size and setting of the trial, gold-standard assessments of toddler attachment using the strange situation paradigm were not feasible. Furthermore, as per most research examining parents’ abuse history, our study relied on parents’ retrospective reports of their abuse experiences in childhood, which may be subject to biases in their self-perceptions of these experiences that affect their accuracy in reporting (Shaffer, Huston, & Egeland, 2008). This study is also limited in that it did not examine parents’ reports of child neglect experiences. Past findings suggest that abuse but not neglect histories may moderate home-visiting outcomes (Ammerman et al., 2016) and that physical and sexual abuse histories are most strongly associated with parental sensitivity (Vaillancourt et al., 2017). Prior results also demonstrate effects of parents’ histories of abuse but not neglect on their victimization of their offspring (Appleyard, Berlin, Rosanbalm, & Dodge, 2011; Berlin, Appleyard, & Dodge, 2011). Thus, these past findings coupled with statistical power constraints guided our focus on examining parents’ abuse histories. Given the increasing prevalence of neglect, however, it will be important for future studies to investigate unique effects of this dimension of parents’ maltreatment history on parenting intervention outcomes in child welfare samples. Finally, our findings may not generalize to more diverse child welfare samples, as 77% of the parents in this study were White. An ongoing trial of PFR in American Indian communities is expected to generate new data about the program’s effectiveness in culturally diverse families. Although this study highlights the impact of parenting on child outcomes, the broader social–environmental context is also important, albeit less malleable (Belsky, 1984).

Our results offer new understanding regarding a promising evidence-based approach to enhancing parenting and child well-being in families involved with CPS. Prevention of adverse family outcomes is a major focus of early intervention services in child welfare; however, evidence-based parenting strategies to this end have only recently been implemented and rigorously evaluated in this setting (Horwitz, Chamberlain, Landsverk, & Mullican, 2010). Although the families of parents with a history of physical abuse may be at heightened risk of poorer outcomes (Vaillancourt et al., 2017; Widom et al., 2015), as suggested here, they may also benefit the most from a home-based parenting intervention targeting parental sensitivity. In this study, the magnitude of the indirect effect of PFR on secure base behavior was modest. Although this finding should be considered with caution, it should also be understood in the context of the practical importance of the intervention for preventing child removals by potentially fostering safe and healthy parent–child relationships (Oxford, Marcenko, Fleming, Lohr, & Spieker, 2016). If the present data are confirmed in future research, they will be valuable in guiding important policy and practice decisions in child welfare services with regard to targeted and tailored intervention for this at-risk subpopulation. Furthermore, our findings are even more encouraging when understood in the context of past results, showing that improved parent–child relationships can help reduce risk of the intergenerational continuity of child maltreatment (Thornberry et al., 2013). In this light, improving parental sensitivity and attachment in toddlerhood in CPS-involved families at high risk of repeated cycles of maltreatment may prove to be critical for ensuring healthy development across generations.

In sum, this study’s findings show that parents with a physical abuse history benefit more from PFR versus a control condition in terms of increased levels of parental sensitivity, which, in turn, result in improved secure base behaviors in toddlers. These initial findings demonstrating how families with parents with a physical abuse history may be impacted by a home-visiting parenting intervention offer new insight into targeting and tailoring therapeutic services to improve outcomes in a subpopulation of CPS-involved families at elevated risk of the intergenerational transmission of abuse. Our results augment a growing body of work highlighting the importance of fostering safe and nurturing parent–child relationships for promoting resilient child outcomes in families exposed to abuse.

Footnotes

Authors’ Note

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by Grants R01 HD061362 and U54HD083091 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development. Additional support for this study was provided by an Australian Research Council Discovery Early Career Researcher Award (Award Number DE170100078) to Dave S. Pasalich.

References

Ammerman

R. T.

Peugh

J. L.

Teeters

A. R.

Putnam

F. W.

Ginkel

J. B. V.

(2016). Child maltreatment history and response to CBT treatment in depressed mothers participating in home visiting. Journal of Interpersonal Violence, 31, 774–791.

Appleyard

Berlin

L. J.

Rosanbalm

K. D.

Dodge

K. A.

(2011). Preventing early child maltreatment: Implications from a longitudinal study of maternal abuse history, substance use problems, and offspring victimization. Prevention Science, 12, 139–149.

Bakermans-Kranenburg

M. J.

van Ijzendoorn

M. H.

Juffer

(2003). Less is more: Meta-analyses of sensitivity and attachment interventions in early childhood. Psychological Bulletin, 129, 195–215.

Barnard

K. E.

(1994). What the Teaching Scale measures. In Sumner

G. S.

Spietz

(Eds.), NCAST: Caregiver/parent–child interaction teaching manual (pp. 98–121). Seattle: NCAST, University of Washington.

Barrett

(2009). The impact of childhood sexual abuse and other forms of childhood adversity on adulthood parenting. Journal of Child Sexual Abuse: Research, Treatment, & Program Innovations for Victims, Survivors, & Offenders, 18, 489–512.

Belsky

(1984). The determinants of parenting: A process model. Child Development, 55, 83–96.

Berlin

L. J.

Appleyard

Dodge

K. A.

(2011). Intergenerational continuity in child maltreatment: Mediating mechanisms and implications for prevention. Child Development, 82, 162–176.

Bernstein

D. P.

Stein

J. A.

Newcomb

M. D.

Walker

Pogge

Ahluvalia

… Zule

(2003). Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child Abuse & Neglect, 27, 169–190.

Berthelot

Ensink

Bernazzani

Normandin

Luyten

Fonagy

(2015). Intergenerational transmission of attachment in abused and neglected mothers: The role of trauma-specific reflective functioning. Infant Mental Health Journal, 36, 200–212.

10.

Bowlby

(1982). Attachment and loss. Vol. 1: Attachment (2nd ed.). New York, NY: Basic Books. (Original work published 1969)

11.

Cassidy

Woodhouse

S. S.

Sherman

L. J.

Stupica

Lejuez

C. W.

(2011). Enhancing infant attachment security: An examination of treatment efficacy and differential susceptibility. Development and Psychopathology, 23, 131–148.

12.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

13.

DeGregorio

L. J.

(2013). Intergenerational transmission of abuse: Implications for parenting interventions from a neuropsychological perspective. Traumatology, 19, 158–166.

14.

Dixon

Hamilton-Giachritsis

Browne

(2005). Attributions and behaviours of parents abused as children: A mediational analysis of the intergenerational continuity of child maltreatment (Part II). Journal of Child Psychology and Psychiatry, 46, 58–68.

15.

Dong

Anda

R. F.

Felitti

V. J.

Dube

S. R.

Williamson

D. F.

Thompson

T. J.

… Giles

W. H.

(2004). The interrelatedness of multiple forms of childhood abuse, neglect, and household dysfunction. Child Abuse & Neglect, 28, 771–784.

16.

Fuchs

Möhler

Resch

Kaess

(2015). Impact of a maternal history of childhood abuse on the development of mother–infant interaction during the first year of life. Child Abuse & Neglect, 48, 179–189.

17.

George

Solomon

Cassidy

Shaver

(2008). The caregiving system: A behavioral systems approach to parenting. In Handbook of attachment: Theory, research, and clinical application (2nd ed., pp. 833–856). New York, NY: Guilford Press.

18.

Hayes

A. F.

(2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: Guilford Press.

19.

Hindley

Ramchandani

P. G.

Jones

D. P.

(2006). Risk factors for recurrence of maltreatment: A systematic review. Archives of Disease in Childhood, 91, 744–752.

20.

Horwitz

S. M.

Chamberlain

Landsverk

Mullican

(2010). Improving the mental health of children in child welfare through the implementation of evidence-based parenting interventions. Administration and Policy in Mental Health and Mental Health Services Research, 37, 27–39.

21.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1–55.

22.

Hughes

Cossar

(2016). The relationship between maternal childhood emotional abuse/neglect and parenting outcomes: A systematic review. Child Abuse Review, 25, 31–45.

23.

Hurlburt

M. S.

Nguyen

Reid

Webster-Stratton

Zhang

(2013). Efficacy of the Incredible Years group parent program with families in head start who self-reported a history of child maltreatment. Child Abuse & Neglect, 37, 531–543.

24.

Kelly

J. F.

Sandoval

Zuckerman

Buehlman

(2008). Promoting first relationships: A program for service providers to help parents and other caregivers nurture young children’s social and emotional development (2nd ed.). Seattle, WA: NCAST Programs.

25.

Kirkland

Bimler

Drawneek

McKim

Schölmerich

(2004). An alternative approach for the analyses and interpretation of attachment sort items. Early Child Development and Care, 174, 701–719.

26.

Levey

E. J.

Gelaye

Bain

Rondon

M. B.

Borba

C. P. C.

Henderson

D. C.

Williams

M. A.

(2017). A systematic review of randomized controlled trials of interventions designed to decrease child abuse in high-risk families. Child Abuse & Neglect, 65, 48–57.

27.

Lyons-Ruth

Block

(1996). The disturbed caregiving system: Relations among childhood trauma, maternal caregiving, and infant affect and attachment. Infant Mental Health Journal, 17, 257–275.

28.

MacKinnon

D. P.

(2008). Introduction to statistical mediation analyses. Mahwah, NJ: Lawrence Erlbaum.

29.

Moran

Pederson

D. R.

Krupka

(2005). Maternal unresolved attachment status impedes the effectiveness of interventions with adolescent mothers. Infant Mental Health Journal, 26, 231–249.

30.

Muthén

L. K.

Muthén

B. O.

(2015). Mplus (Version 7.4) [Computer software]. Los Angeles, CA: Author.

31.

Olds

Henderson

Kitzman

Eckenrode

Cole

Tatelbaum

(1998). The promise of home visitation: Results of two randomized trials. Journal of Community Psychology, 26, 5–21.

32.

Oxford

M. L.

Marcenko

Fleming

C. B.

Lohr

M. J.

Spieker

S. J.

(2016). Promoting birth parents’ relationships with their toddlers upon reunification: Results from Promoting First Relationships® home visiting program. Children and Youth Services Review, 61, 109–116.

33.

Oxford

M. L.

Spieker

S. J.

Lohr

M. J.

Fleming

C. B.

(2016). Promoting First Relationships®: Randomized trial of a 10-week home visiting program with families referred to child protective services. Child Maltreatment, 21, 267–277.

34.

Oxford

M. L.

Spieker

S. J.

Lohr

M. J.

Fleming

C. B.

Dillon

Rees

(2018). Ensuring implementation fidelity of a 10-week home visiting program in two randomized clinical trials. Maternal and Child Health Journal, 22, 376–383.

35.

Pasalich

D. S.

Cyr

Zheng

McMahon

R. J.

Spieker

S. J.

(2016). Child abuse history in teen mothers and parent–child risk processes for offspring externalizing problems. Child Abuse & Neglect, 56, 89–98.

36.

Pasalich

D. S.

Fleming

C. B.

Oxford

M. L.

Zheng

Spieker

S. J.

(2016). Can parenting intervention prevent cascading effects from placement instability to insecure attachment to externalizing problems in maltreated toddlers? Child Maltreatment, 21, 175–185.

37.

Preacher

K. J.

Kelley

(2011). Effect size measures for mediation models: Quantitative strategies for communicating indirect effects. Psychological Methods, 16, 93–115.

38.

Rucker

D. D.

Preacher

K. J.

Tormala

Z. L.

Petty

R. E.

(2011). Mediation analysis in social psychology: Current practices and new recommendations. Social and Personality Psychology Compass, 5, 359–371.

39.

Shaffer

Huston

Egeland

(2008). Identification of child maltreatment using prospective and self-report methodologies: A comparison of maltreatment incidence and relation to later psychopathology. Child Abuse & Neglect, 32, 682–692.

40.

Shenk

C. E.

Ammerman

R. T.

Teeters

A. R.

Bensman

H. E.

Allen

E. K.

Putnam

F. W.

Van Ginkel

J. B.

(2017). History of maltreatment in childhood and subsequent parenting stress in at-risk, first-time mothers: Identifying points of intervention during home visiting. Prevention Science, 18, 361–370.

41.

Spieker

S. J.

Oxford

M. L.

Kelly

J. F.

Nelson

E. M.

Fleming

C. B.

(2012). Promoting First Relationships: Randomized trial of a relationship-based intervention for toddlers in child welfare. Child Maltreatment, 17, 271–286.

42.

Theise

Huang

K.-Y.

Kamboukos

Doctoroff

G. L.

Dawson-McClure

Palamar

J. J.

Brotman

L. M.

(2014). Moderators of intervention effects on parenting practices in a randomized controlled trial in early childhood. Journal of Clinical Child & Adolescent Psychology, 43, 501–509.

43.

Thornberry

T. P.

Henry

K. L.

(2013). Intergenerational continuity in maltreatment. Journal of Abnormal Child Psychology, 41, 555–569.

44.

Thornberry

T. P.

Henry

K. L.

Smith

C. A.

Ireland

T. O.

Greenman

S. J.

Lee

R. D.

(2013). Breaking the cycle of maltreatment: The role of safe, stable, and nurturing relationships. Journal of Adolescent Health, 53, S25–S31.

45.

Toth

S. L.

Gravener-Davis

J. A.

Guild

D. J.

Cicchetti

(2013). Relational interventions for child maltreatment: Past, present, and future perspectives. Development and Psychopathology, 25, 1601–1617.

46.

Vaillancourt

Pawlby

Fearon

R. M. P.

(2017). History of childhood abuse and mother–infant interaction: A systematic review of observational studies. Infant Mental Health Journal, 38, 226–248.

47.

Waters

(1987). Attachment Behavior Q-sort (version 3.0). Unpublished manuscript, State University of New York at Stony Brook, Stony Brook.

48.

Widom

C. S.

(2017). Long-term impact of childhood abuse and neglect on crime and violence. Clinical Psychology: Science and Practice, 24, 186–202.

49.

Widom

C. S.

Czaja

S. J.

DuMont

K. A.

(2015). Intergenerational transmission of child abuse and neglect: Real or detection bias? Science, 347, 1480–1485.