Introduction

In 2009, child protection agencies received an estimated 3.3 million reports of child maltreatment, and of these, an estimated 702,000 children were found to be victims of maltreatment (USDHHS 2010a, b). Studies have indicated that in addition to maltreatment, many parents in the child welfare system also experience a range of problems such as intimate partner violence, substance abuse, and mental health problems (English et al. 2005; Kohl et al. 2005; Marsh et al. 2006). The co-occurrence of these issues is a significant problem for the health and safety of many children in the United States.

The youngest children, from birth to age 3, are at the highest risk of maltreatment (Herrenkohl et al. 2008; USDHHS 2010a, b) and maltreatment can have serious, and often life-long, adverse consequences. Cumulative risk research asserts that the greater the number of risk factors present in a child’s life, the greater the prevalence of developmental and health problems (Appleyard et al. 2005; Sameroff 2000). These negative consequences for children include internalizing and externalizing behavior problems; increased potential for depression, anxiety, and PTSD; difficulty with peer relationships; and cognitive problems (Herrenkohl and Herrenkohl 2007; Herrenkohl et al. 2008; Margolin and Gordis 2000; Osofsky 2003). In addition, attachment studies have shown that children raised in stressful home environments may be less able to bond to their adult caregivers (Herrenkohl et al. 2008), which has in turn been associated with difficulty forming relationships in both childhood and adulthood and the perpetuation of the intergenerational cycle of violence (Baer and Martinez 2006; Herrenkohl et al. 2008; Morton and Browne 1998). Exposure to environments that produce stress and fear in children have also been linked to changes in children’s brain architecture, which can lead to lifelong consequences for how children learn and solve problems (National Scientific Council on the Developing Child 2010).

The accurate assessment of co-occurring stressors in maltreating families is critical in order to intervene effectively. Although intimate partner violence (IPV) is frequently associated with child maltreatment and negative child outcomes, it is often under-identified by child welfare agencies by 20–30% (Kohl et al. 2005). There is also relatively little known about the types of intimate partner conflict that occur in families in the child welfare system (English et al. 2009). In addition, the identification of substance abuse in the primary caregiver often overshadows issues of IPV (Kohl et al. 2005). Estimates of the percentage of parents with serious substance abuse problems involved in the child welfare system have ranged from one-third to two-thirds of parents (DHHS 1999) to as high as 80% (Marsh et al. 2011). Further, studies have found caregiver substance abuse to be the single most potent kind of caregiver vulnerability factor in predicting child maltreatment substantiation (Wekerle et al. 2007). In addition, mental health issues, particularly for mothers, may also have serious consequences for children’s psychological health outcomes (Beckman et al. 2010). It is clearly critical to understand the range of stressors faced by each family, in order for interventions to be based upon a more comprehensive understanding of the family.

Within the child welfare system, parent training programs are among the most frequent interventions assigned to parents (Barth 2009; Barth et al. 2005; Hodnett et al. 2009). Despite this prevalence, there is not good national data available about the types of parent training interventions that are routinely used in child welfare settings (Barth et al. 2005). There have been even fewer studies of parenting interventions that are targeted to the family as a unit, despite the fact that stressors that impact the entire family are related to maltreatment (Beckman et al. 2010). In addition, there is limited information on the effectiveness of parenting education programs, particularly the Nurturing Parenting Program, for maltreating parents with specific co-occurring problems. Only one recent study (Hodnett et al. 2009) has examined the relationship between parent demographic characteristics on outcomes in that program.

The purpose of this study was to explore the nature and co-occurrence of family stressors, particularly violence, substance abuse, and mental health problems, in a sample of parents involved in the child welfare system who have been referred to an intensive therapeutic parent training program. In addition, this study sought to identify whether parenting outcomes, as measured by the Adult/Adolescent Parenting Inventory, differed according to whether or not partner abuse or conflict, substance abuse, or mental health issues were identified in the program files.

Methods

Sample

All information for this study was derived from families who have participated in the Nurturing Parents Program (NPP), an intensive, 15-week therapeutic parenting program for individuals who are involved in or at risk of becoming involved in the child welfare system. The program is administered in a group format; adult groups meet weekly for 2.5 h, and are facilitated by two therapists. Topics covered in the group include empathy; effective communication; problem-solving; addressing and managing anger; expectations of children’s behavior and child development; and types of discipline, including the use of corporal punishment, and others. A children’s group, for children ages 2–12, meets concurrently. Adults participate in the program as a parenting dyad, defined as two adults who are involved with the children’s daily life; individuals can be married, co-parenting, or have a familial relationship (e.g., mother and grandmother of a child). Participants have been screened to not currently be in a physically violent relationship, as measured by the presence of an injunction or disclosure of current physical abuse. However, this does not exclude other types of abuse or a history of physical violence in the current or a past relationship.

Program files were reviewed for all classes of participants who completed the Nurturing Parenting program from initiation of the program in July, 2008 until October, 2010. In total, 62 cases were reviewed (each consisting of a parenting dyad), for a total of 124 individual adults. Of these, 81% completed the program. While parenting dyads do not have to be a mother-father couple, only data for mothers (or mother figures) and fathers (or father figures) are presented throughout this paper, because the number of other types of adults was too small to analyze. The final sample consisted of 56 fathers (46 biological fathers, 6 mother’s paramours, and 4 step-fathers) and 61 mothers (58 biological mothers and 3 step-mothers). Participants were referred to the program by the following sources: 42% through the child welfare system as a requirement for a case plan; 24% through Child Protection Investigations as a preventive diversion effort before opening a child welfare case; 26% at risk for child welfare or diversion involvement through other agencies, such as Healthy Start; and 8% were self-referred.

A sub-sample of 21 program participants participated in in-depth, in-person interviews: 12 mothers, 7 fathers or boyfriends, and 2 grandparents who acted as caregivers. Fourteen of these participants were referred through child welfare services and 7 through the diversion program.

Data Collection

Data were collected primarily from reviews of program files, and were supplemented with in-person interviews for a sub-sample of program participants. Program files contained multiple sources from which data were extracted (See Table 1).

Table 1 Data collection matrix for program files
Full size table

File Review

Files were reviewed using a structured data collection instrument. Abstracted information included: (1) demographic information about both parents and children; (2) type of referral; (3) reason for referral to child welfare system and nature of child maltreatment allegations, if they were made; (4) identification of past or present intimate partner violence or conflict issues within the couple; (5) other stressors faced by parents, including mental health and substance abuse as well disabilities, history of abuse in the family of origin, teen parenthood, and prior termination of parental rights; and (6) results of the pre- and post-test Adult-Adolescent Parenting Inventory.

Table 1 illustrates the component of data collection; location in the file; operational definition; and type of analysis for each item.

To capture attitudes toward parenting and child-rearing, data from the Adult-Adolescent Parenting Inventory (AAPI-2) (Bavolek and Keene 1999) were abstracted. Each participant completed the AAPI-2 before and after the parenting program. The AAPI-2 is a 40-item, Likert-scale inventory that provides an index of child maltreatment risk using five constructs: (1) Inappropriate Parental Expectations of Children (do parents inaccurately perceive the skills and abilities of their children); (2) Parental Lack of Empathy Towards Children’s Needs (do parents perceive children’s every day, normal demands as unrealistic, resulting in increased stress); (3) Strong Belief in the Use of Corporal Punishment (do parents use physical punishment as the preferred means of discipline); (4) Reversing Parent–Child Family Roles (do parents interchange some of the traditional role behaviors of parent and child, so that parents act like children looking to their own children for care and comfort); and (5) Oppressing Children’s Power and Independence (do parents demand obedience and complete compliance to parental authority without allowing children to have choices or voice opinions) (Bavolek and Keene 1999). The AAPI-2 has an internal consistency (Cronbach’s alpha) ranging from .83 to .98 (Hodnett et al. 2009).

The information for each parent in the couple is kept in separate, but related files, although in each case, some principal information pertaining to both parents is stored in the file of only one member of the dyad. For example, the father’s file contains the referral to the program, but the information in the referral may apply to either or both parents. Therefore, data were collected so that the parenting dyad and children in the family formed a “case,” and data for each case was collected on one instrument to form a cohesive picture of the family unit. Children were listed as “Child 1,” “Child 2,” etc., based on their age, with the oldest child in the family listed first. All data were entered into a password-protected database on a secured computer.

Interviews

Participants were recruited for interviews through two mechanisms: brief presentations by the researcher to ongoing groups, and mailed letters to previous program participants. Potential participants recruited through either method contacted the researcher by phone or email; participants recruited through presentations were also invited to provide their contact information to the researcher if they were interested in an interview. Parents were excluded if they did not speak English fluently, were not comfortable interviewing in English, were younger than 18 years, or did not have any involvement with the child welfare system or diversion services.

Semi-structured, in-depth, in-person interviews were conducted with 21 participants of the parenting program who had been referred to the program either through child welfare services or diversion services. Interviews lasted approximately 1 h and utilized a flexible interview guide. The interview guide was developed using a family-centered framework and contained questions about the parent’s involvement in the child welfare system. For the purposes of this manuscript, only participants’ comments regarding the parenting program will be presented. All interviews were conducted at a public location. If both members of the couple were interested in interviewing, interviews were held separately. A $25 incentive was provided for participation.

Analysis

File Review

Data were entered into a password-protected database. Descriptive statistics were conducted on the overall sample as well as on the sub-samples of cases that were classified as having violence, substance abuse, or mental health problems.

Bivariate correlations were performed to examine relationships between the co-occurring risk factors (i.e., violence, substance abuse, and mental health). Related-samples t tests were performed to examine pre- and post-test differences on the AAPI-2 across eight groups of parents: (1) all fathers; (2) all mothers; (3) fathers with violence issues; (4) mothers with violence issues; (5) fathers with substance abuse issues; (6) mothers with substance abuse issues; (7) fathers with mental health issues; and (8) mothers with mental health issues. Mean standardized scores were also examined for each AAPI-2 subscale in each subgroup for both pre-test and post-test to examine overall change. All analyses were conducted using IBM SPSS Statistics version 19.0.

Interviews

Interviews were transcribed verbatim and analyzed using Atlas.ti. Prior to analysis, an a priori codebook was developed consisting of general, flexible themes, based on the interview guide. Analysis was ongoing as interviews were conducted and transcribed. Identification of emergent codes occurred throughout the iterative open coding process, and those codes were added to the codebook or used in place of the a priori codes as appropriate. A “second pass” through the data examined conditions and interactions, and determined which categories clustered together (Neuman 2003). Once major themes were identified and organized, data were selectively coded in order to illustrate themes and make comparisons. For the purpose of reliability and validity, a second coder coded the majority of transcripts. After the researcher and the second coder coded each transcript independently, they met and discussed each transcript in depth, and came to consensus on the way codes were applied to the interviews.

This study was approved by the Institutional Review Board at the University of South Florida.

Results

Demographics

Parents

All data are presented by relationship to the child (i.e., fathers and mothers). “Fathers” are defined as biological fathers, step-fathers, or mother’s paramour; “mothers” are defined as biological mothers, step-mothers, or father’s paramour. Other program participants, such as grandparents, were excluded from the analysis because the numbers were too small to be meaningful. The final sample consisted of 56 fathers (46 biological fathers, 6 mother’s paramour, and 4 step-fathers) and 61 mothers (58 biological mothers and 3 step-mothers).

Table 2 illustrates the demographic characteristics of fathers and mothers.

Table 2 Parent demographics
Full size table

Children

Most cases (44%) involved one child; 19% two children; 22% three children; and 15% 4–5 children. Children ranged in age from less than 6 months to 18 years. The median age for the oldest child in the family (Child 1) was 6.5 years; median ages for subsequent children were 6.0 years (Child 2) and 4.5 years (Child 3). Over half of the children identified in the study were male: Child 1 was male in 52% of cases, Child 2 in 61% of cases, and Child 3 in 46% of cases.

Child Maltreatment Allegations

The reason for referral to the child welfare system, when available, is presented for all participating fathers and mothers in Table 3. Allegations were not always available in the file, particularly for those families that were self-referred (8%) or who were referred to the program through a non-child welfare agency (26%). In addition, the referral form in some child welfare or diversion files did not contain the nature of the allegations. The percentage of fathers with physical abuse allegations was higher than for mothers, and the percentage of substance abuse allegations was higher for mothers than fathers.

Table 3 Reason for referral to child welfare system
Full size table

Prevalence of Family Stressors

Table 4 illustrates the percentage of mothers and fathers experiencing various stressors. The files indicated a higher percentage of many issues for mothers compared to fathers, including mental health, substance abuse, violence and conflict, teen parenthood, and depression.

Table 4 Co-occurring issues for fathers and mothers
Full size table

Violence and conflict were significant issues for the participants in this program. In addition to the percentages of current abuse or conflict noted in Table 4 (25% of mothers and 22% of fathers), approximately two-thirds of participants had been involved in at least one past incident of abuse or conflict. Past incidents included both relationship abuse in prior romantic relationships as well as abuse in the family of origin (both exposure to IPV and child maltreatment as a child).

Relationships Between Co-Occurring Stressors

Files indicated that many parents experienced more than one co-occurring stressor. Of all mothers, 39% experienced concurrent substance abuse and mental health issues; 30% experienced current violence and substance abuse; and 29% violence and mental health. Similarly, of all fathers, 27% experienced substance abuse and mental health, 26% experienced current violence and substance abuse issues; and 23% current violence and mental health issues. This overlap in stressors was also described by participants in the in-person interviews, in which one-third of participants experienced at least two problems. Co-occurring stressors were also evident in the reasons for referral to the parenting program, as abstracted from the file. For example, families experiencing violence or conflict were referred to the program for various reasons, not all of them involving relationship violence. For those families, other reasons for referral most often included issues of substance abuse, mental health problems, physical abuse or corporal punishment of children, and the specific need for young, first-time parents to gain additional parenting skills. Table 5 provides the results for the Pearson Correlations for selected family stressors.

Table 5 Pearson correlations for co-occurring stressors
Full size table

Parents’ Attitudes Toward Parenting

Standardized scores and paired-samples t tests for the AAPI-2 instrument are presented for parents who completed both a pre-test and post-test. Table 6 presents results for all fathers and mothers. Results are also presented for each sub-group: parents with identified violence or conflict (Table 7), substance abuse (Table 8), and mental health (Table 9). Standardized pre- and post-test scores range from 1 to 10; higher scores represent more desirable outcomes (e.g., a score of 8 on the Inappropriate Expectations sub-scale indicates that the parent has more appropriate expectations of their child than a score of 4). Wilcoxon Signed-Rank tests were also performed for all sub-groups, and results supported those of the t tests. Only the t test results are presented for direct comparison between groups.

Table 6 Paired sample t tests for AAPI results: all parents who completed program
Full size table
Table 7 Paired samples t tests for AAPI results: parents with violence issues
Full size table
Table 8 Paired samples t tests for AAPI results: parents with substance abuse issues
Full size table
Table 9 Paired samples t tests for AAPI results: parents with mental health issues
Full size table

The results of these analyses indicate that parents in all sub-groups improved on each sub-scale. However, while fathers in the violence and substance abuse sub-groups improved on the oppressing power sub-scale, their improvements were not significant. In fact, fathers’ lowest scores in all sub-groups were on the sub-scale that measured oppressing their children’s power and independence. Mothers in the substance abuse sub-group scored the highest on the oppressing power measure, which was the highest of any measure for either mothers or fathers in any group. Mothers in all groups scored lowest on the sub-scale that measured empathy for their children’s needs.

Parents’ Perceptions of the Parenting Program

While the AAPI results provided information on the parents’ change in parenting attitudes from pre- to post-test, in the qualitative interviews (N = 21), parents expressed their opinions about the Nurturing Parenting Program, including perceived benefits and challenges. Overall, parents had positive comments regarding this program. One-third of parents described that the program helped them change the method used to discipline their children because they learned alternatives. Reduction in corporal punishment was one benefit described by participants:

Yeah, I have calmed down. I will still spank every now and then, but I just grit my teeth…and deal with it differently. So yeah, it made a positive change and still is. [Father]

Several parents also identified the impact of their parenting styles and specifically noted the impact that corporal punishment can have on children, such as teaching them violence. In addition to changing discipline practices, 9 of the parents also described that the program helped them with communication with their partner. Parents specifically discussed that they had a better understanding and tools for how to cope with anger, and how to more effectively show respect for their partner.

In addition to specific comments on parenting practices, one-third of parents also mentioned that the facilitators of the parenting program were supportive, helpful, and concerned. These responses are in contrast to their general perceptions of child welfare caseworkers. Parents mentioned that the parenting program was the first place where they felt providers were interested in their perceptions and concerns. For example, one father said:

Well the caseworker didn’t [ask about our concerns], but when they referred us to those people, [the parenting program provider], they were really interested in the family life, and our concerns with the kids, what goes on with them and stuff like that. [Father]

Several parents reported challenges associated with the program, particularly with implementing new parenting practices. For example, one mother described that the information provided in the program was a lot to take in all at once, and therefore difficult to implement. Two other mothers, whose children were in foster care, faced challenges because they wanted to implement the new skills immediately, but were unable to do so. One mother said:

It do be kinda hard learning stuff in the class, cause I can’t go home and put into effect what I learned right then. I got to wait, but it’s helping me, you know, ‘cause I let it sink in and everything. [Mother]

Overall, parents expressed that they learned new skills from this parenting program, regardless of whether they initially did not want to be there, or whether they initially felt they did not need assistance with their parenting.

At first, like the first couple of weeks, I was like I can’t believe I have to do this, and it’s ridiculous. But it was all right. I mean the group, we got to know the people in our group and stuff, and they were people just like us. There was a couple that was our age, couples that were older. I liked the group thing, the way it was set up like that. [Mother]

The supportiveness of the group facilitators and group format of the class contributed positively to parents’ perceptions, and perhaps made them feel less stigmatized, thus better meeting their needs.

Discussion

The review of co-occurring family issues in this population indicated that more mothers have issues with violence, mental health, and substance abuse that are identifiable in the program file than fathers. The reason for this is unclear, although it is likely at least in part due to the increased scrutiny of mothers compared to fathers by the child welfare system. It is difficult to make generalizations, though, because the files contained limited information. However, between 23 and 39% of parents were currently experiencing at least two of those issues at the time they entered the program, and this data was reinforced by one-third of interview participants who disclosed multiple stressors. This is particularly of concern given what research has indicated about the cumulative negative effect of family stressors on child health, behavior, development, and learning (Anda et al. 2006; National Scientific Council on the Developing Child 2010), especially over the lifespan.

Although this parenting program specifically aimed to screen out families with active physical intimate partner violence, the files indicated that approximately 25% of parents were currently experiencing some type of abuse or conflict in their relationship, and this may be an underestimate because not all abuse may be captured in the files. Kohl et al. (2005) demonstrated that nationally, child welfare workers primarily screen for physical abuse, and they often significantly under-identify intimate partner violence compared to what parents reported during interviews. The sample of parents in this program experienced a wider array of abusive and conflict behaviors in their relationships that have the potential to impact their ability to be effective parents, yet these behaviors were not formally recognized by the referring agency or were not noted in the referral to the parenting program. It was also particularly noteworthy that approximately two-thirds of parents in this study revealed past abuse in the psychosocial interview with program facilitators, yet few disclosed taking part in interventions for these problems. This suggests that different types of abusive behaviors may not be routinely addressed through child welfare interventions, which was confirmed by interviews of parents in this study.

Barth et al. (2005) determined that the Nurturing Parenting Program (NPP), while not the most commonly used parent-training program in child welfare, was a “possibly efficacious” intervention, for which the currently available research designs “have the capacity to show substantial likelihood of benefit” (p. 360). Little research to date has examined the program’s effectiveness for sub-groups of parents in the child welfare population, including those experiencing other issues concurrently with maltreatment. The results of this study indicate that for this sample, although the group sizes were small for sub-group analysis, participants’ parenting attitudes improved significantly from pre-test to post-test for the entire population of parents who completed the program, as well as for members of each of the sub-groups. This suggests that even parents who faced family stressors, such as mental health and substance abuse, experienced positive changes in their parenting attitudes. In fact, significant improvements were found in all AAPI-2 sub-scales with the exception of the Oppressing Children’s Power and Independence sub-scale for fathers with violence and substance abuse problems. It is difficult to determine why this is the case, but it may be due to the small sample size or the usefulness of this particular sub-scale in samples of fathers with co-occurring stressors.

Qualitative interviews with parents confirmed that parents felt that they were learning from the program, particularly in reference to changing attitudes towards corporal punishment and changing communication behaviors within the couple. Further, parents indicated that they felt supported by the group facilitators and sometimes the other parents in the group. In other qualitative results of this study not included here, this type of support and empowerment was generally not noted by parents in regard to the child welfare system as a whole, but did characterize other mandated therapeutic services to which parents were referred. These findings indicate that there may be some advantage to the intensive, therapeutic nature of this intervention, especially as it is designed for the family unit and little research attention has been given to such programs to date (Beckman et al. 2010). In light of the knowledge that family stressors contribute significantly to maltreatment, more rigorous evaluations of the Nurturing Parenting Program with child welfare populations may be warranted, as well as more qualitative research to further understand how parents perceptions of the parenting interventions may impact their parenting attitudes.

This study was limited by a small sample size, which restricted the type of analyses that could be completed. In addition, the information was limited to what was available for abstraction in the program files, and files did not always contain the exact same information for every family. Without the information contained in the child welfare system records, it is not possible to draw conclusions about what types of problems were identified by the child welfare worker that may not have been included in the referral to the parenting program or communicated with the program facilitators directly. More detailed information from the child welfare worker, if available, would provide additional information on the circumstances of these families. Further, there was no comparison group with which to compare the AAPI-2 results, limiting the conclusions to only the participants in this particular intervention program. While this limitation has also been identified for other studies on this program (e.g., National Registry of Evidence-based Programs and Practices 2010), the multi-method approach to data collection allowed triangulation of information to offset some of these issues. This was particularly true for the information obtained from the qualitative interviews, which demonstrated that parents’ perceptions of the parenting program were supportive of their change in attitudes toward parenting.

This study contributes to the knowledge base regarding parents in the child welfare system with co-occurring problems. Parents who experience multiple problems in addition to maltreating their children are unlikely to be able to parent effectively, and thus are likely to need significant support and intervention. A better understanding of the types of issues they face and whether particular intervention programs are effective for those groups would allow more targeted, individualized interventions that may lead to increased success for parents.