Introduction

Researchers over several decades have pursued empirical evidence for the premise that children who are abused and neglected are at increased risk for maltreating their own children. Childbearing during adolescence has been studied both as an outcome of child maltreatment history and as a precursor to future maltreatment (Buek et al., 2019; Garwood et al., 2015; Pears & Capaldi, 2001). Numerous studies investigated whether a history of abuse and neglect increases girls’ risk for teen pregnancy, while other studies focused on the increased risk for maltreatment among children born to adolescent mothers. While the US teen birth rate (17.4 births per 1000 females aged 15 to 19 in 2018) declined in recent years, it remains considerably higher than in much of the developed world (Sedgh et al., 2015). Teen childbearing is associated with poorer educational, behavioral, and health outcomes for both the teen mother and her children (Martin et al., 2019; Ventura et al., 2014).

While methodologies and findings from these studies vary widely, some consensus supports the claim that children born to adolescent mothers with a history of abuse are at increased risk for maltreatment during their infancy and early childhood (McDonnell & Valentino, 2016; Widom et al., 2015). Because many studies included small samples, inconsistent definitions of maltreatment, and relied on self-reported maltreatment in one or both generations, research with improved methodologies remains necessary. In particular, reviewers of the literature on intergenerational maltreatment emphasize the need to measure maltreatment in adjacent generations via multiple reporters or official CPS records to increase reliability of findings and to avoid the biases inherent in retrospective recall (Child Welfare Information Gateway, 2016; Madigan et al., 2019; Thornberry et al., 2012).

In this study, we examine characteristics associated with adolescent mothers who have been identified as maltreatment perpetrators and the patterns of their prior maltreatment experiences. To do this, we linked adolescent mother perpetrators identified in FFYs 2016–2018 National Child Abuse and Neglect Data System (NCANDS) data to their alleged victim records in NCANDS FFYs 2005–2018. To our knowledge, this is the first study to employ this approach and thus extends the evidence base on intergenerational transmission of maltreatment and contributes an innovative methodology for studying the phenomena. Of adolescent mother perpetrators we identified, we explored how many had prior CPS experience as alleged victims and the nature of this experience, such as whether they were determined to be a maltreatment victim (i.e., the CPS investigation determined at least one maltreatment allegation was substantiated or indicated) and the number and timing of allegations. Gaining a better understanding of prior CPS experience can help caseworkers and practitioners working with children and youth effectively target services and support to prevent maltreatment of the next generation. By sharing this innovative methodology with researchers, practitioners, and policymakers, we hope to increase opportunities for stakeholders to understand and prevent intergenerational maltreatment.

Background

To understand intergenerational patterns of maltreatment, it is necessary to examine the multiple risk factors, such as early pregnancy and childbearing (Pears & Capaldi, 2001) and childhood trauma (Dixon et al., 2009), which may mediate the relationship between childhood maltreatment experiences and abusive parenting practices.

Maltreatment as a Risk Factor for Pregnancy During Adolescence

Adolescent girls with a history of maltreatment were found to have a higher risk for early pregnancy (Garwood et al., 2015; Hillis et al., 2004; Kugler et al., 2018; Noll & Shenk, 2013; Putnam-Hornstein et al., 2013). Moreover, in a nationally representative sample of alleged victims, the National Survey of Child and Adolescent Well-Being (NSCAW) found that almost half of girls aged 18–20 had a pregnancy within 3 years after the index report of maltreatment (Casanueva et al., 2014). Adolescent pregnancy is associated with the cumulative effect of multiple adverse childhood experiences (ACEs), including child abuse which typically occurs within a broader social context of family dysfunction (Hillis et al., 2004). While child maltreatment and pregnancy during adolescence may plausibly overlap with other ACEs, Garwood et al. (2015) found that teens with victim allegations had a significantly higher rate of pregnancy even after controlling for poverty and other caregiver risk factors. Noll and Shenk (2013) found that even when controlling for demographic risk factors, adolescent females with substantiated maltreatment were more than twice as likely to become mothers during adolescence compared with their peers who were never alleged victims. Similarly, Putnam-Hornstein et al. (2013) found that among a population-based sample of adolescents who gave birth in California during a single year, 45% had been referred to CPS as alleged victims.

Early Motherhood as a Risk for Maltreatment

Additional research has provided support for the increased risk for maltreatment among children born to teen mothers. Buek et al. (2019), in their study of risk and protective factors that mediate the relationship between paternity establishment at birth and early maltreatment, found that the odds of maltreatment significantly increased for children of mothers younger than 18 years. Connelly and Straus (1992) collected self-reports of violence toward their children from a nationally representative sample of parents and found that the rate of physical abuse was higher the younger the mother was at the time of the child’s birth, even when controlling for family income, race, number of children in the home, age of abused child, mother's education, and marital status.

Intergenerational Maltreatment

Research has examined the relation between mothers’ childhood trauma, including experiences of maltreatment, with outcomes for their children. For example, Widom et al. (2015) found that children of parents with histories of childhood abuse and neglect were more likely to report sexual abuse and neglect and to have experienced CPS intervention at some point in their lives, compared with a matched comparison group of children. McDonnell and Valentino (2016) found that substantiated maltreatment of mothers predicted higher levels of both prenatal and postnatal depressive symptoms and higher levels of maladaptive socioemotional symptoms among their infants at 6 months of age. Putnam-Hornstein et al. (2015) found a link between an adolescent mother’s history of either unsubstantiated or substantiated maltreatment and her CPS involvement for maltreating her own children. The timing of a child’s maltreatment experiences may also affect their impact. For example, ACEs clearly affect certain aspects of development, yet research has demonstrated that proximal adversity may have a greater impact on social-behavioral outcomes (Font et al., 2020; Najman et al., 2010). Thornberry and Henry (2013) found that maltreatment that occurs in adolescence, or that begins in childhood and persists into adolescence, significantly increases the odds of maltreatment perpetration but childhood-limited maltreatment does not.

Based on a systematic review of the literature testing the intergenerational cycle of maltreatment hypothesis, Thornberry et al. (2012) concluded that improved methodological rigor was necessary to fully answer this question. They identified specific criteria necessary for stronger research, including sufficient sample size, use of control groups, clear and consistent definitions of maltreatment, and independent reporting of maltreatment for each generation. Similarly, the Child Welfare Information Gateway (2016) reviewed the evidence base for the persistence of child abuse and neglect across generations and found that the wide variation in sample sizes, measurement tools, and definition of key terms continued to limit the conclusions that can be drawn from this research. Madigan et al. (2019) conducted an extensive meta-analysis of the intergenerational maltreatment literature to assess the impact of methodological limitations and biases such as poor operational definitions of maltreatment, use of single informants to assess maltreatment in different generations, reliance on retrospective recall, and few studies with prospective design.

Madigan et al. (2019) identified methodologies in which CPS records were used to assess maltreatment of both the parent and the child as meeting a “high standard criterion.” Only one such study was identified by Thornberry et al. (2012), but Madigan et al. (2019) identified nine studies that relied on CPS records to measure maltreatment in both generations, including Bartlett et al. (2017) and Putnam-Hornstein et al. (2015), suggesting an improvement over time in methodology in this area. CPS records rely on consistent, policy-driven definitions of maltreatment and reflect input from multiple reporters and investigators over time, rather than the recollections of a single individual, which is necessary to maintain independence of independent and dependent variables (Thornberry et al., 2012). In many prior studies, maltreatment for one generation was measured from administrative data, but maltreatment for the other generation comes from self-report or other subjective assessments. While self-report of maltreatment may capture the whole range of maltreatment experiences beyond what is reported to CPS, parents and children may underreport experienced maltreatment. Buisman et al. (2020) found that parents may underreport perpetrated maltreatment because of social stigma or fear of legal consequences and children may underreport experienced maltreatment out of loyalty to their parents or fear of punishment; self-reported maltreatment by both generations may be biased due to distorted memories and subjective interpretations of experiences.

Current Study

In this retrospective, single cohort study, we first examined what proportion of adolescent mothers identified as maltreatment perpetrators have prior experiences with CPS as alleged victims of maltreatment. We then explored which characteristics of adolescent mother perpetrators were associated with their having had prior experiences with CPS. Finally, we considered the types of maltreatment that were reported for these adolescent mother perpetrators and whether there is a relationship between the type of maltreatment experienced as a child and the type of maltreatment perpetrated.

While we report on whether past CPS experience included any substantiated reports in which the adolescent mother perpetrator had been previously identified as a victim, the focus is on any experience with maltreatment allegations accepted for CPS investigation. Hussey et al. (2005) found that substantiation status was not associated with child outcomes, and other researchers have found that a history of either unsubstantiated or substantiated maltreatment similarly predicted future CPS involvement among young mothers (Bartlett et al., 2017; Putnam-Hornstein et al., 2015).

While this study focuses on a sample of adolescent mothers substantiated as perpetrators and the characteristics of this sample as they relate to prior childhood CPS engagement, we do not specifically address prior CPS experience or maltreatment as risk factors for adolescent pregnancy nor do we examine early motherhood as a risk for maltreatment. The purpose of the current study is to describe adolescent mother perpetrators and contribute to a better understanding of their prior CPS involvement.

Method

Perpetrators in the NCANDS Child File

In the current research, we examine administrative data reported to NCANDS.Footnote 1 States construct an electronic file containing child-specific records for each report of alleged maltreatment for which a determination about the CPS response was reached during the reporting year. The data include characteristics of the children and perpetrators, the types of maltreatment alleged, the CPS response findings, the risk factors of the child and the caregivers, the services that were provided, and the perpetrators (U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau, 2021). NCANDS defines a perpetrator as a person who was determined to have caused or knowingly allowed the maltreatment of a child. Perpetrators are only included in the data if a maltreatment allegation is substantiated. For FFY 2019, NCANDS received data on 525,319 unique perpetrators; less than 2% of these were younger than age 18. Sixty-eight percent of victims were maltreated by their mothers, either acting alone, with their [the victim’s] father or another person (U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau, 2021).

Sample

Data from 27 states were used to create the dataset for this research. States were chosen based on the completeness and validity of their data on perpetrators and their relationship to the victims and on maintaining the same unique identifiers over the 14-year span of this study. We also confirmed each of these states’ practices of maintaining the same unique identifiers for individuals whether they were entered into the data system as children who were alleged maltreatment victims or as perpetrators. The sample included 8205 adolescent mothers who were identified as maltreatment perpetrators of their children.

Dataset Construction

When state child welfare information systems maintain the same unique identifier for individuals whether they are entered as an alleged victim or as a perpetrator, we can link records for individuals with CPS experience as both perpetrators and as alleged victims. Our study relied on this approach to discover past CPS experience as alleged victims over a 14-year period among a sample of adolescent mother perpetrators. A flowchart of the steps of dataset construction is presented in Appendix 1.

Unique perpetrator dataset

Based on methodology developed by Shusterman et al. (2005), we constructed a study sample of unique perpetrators included in the NCANDS data for FFYs 2016–2018 meeting the following criteria: female, parent of the victims, and age 12–19 at the time of the report. Any perpetrator who was coded in the dataset as having a prior perpetrator record based on the prior perpetrator field in NCANDS was excluded, as well as any perpetrators reported as unknown sex or unknown age in other records with the same ID. Our focus for this analysis was to identify those who were perpetrators, so we included only the first report for each perpetrator included. Our dataset of 8205 unique perpetrators included for each adolescent mother perpetrator demographics (i.e., age, race) and circumstances of her current perpetrated maltreatment, including the number and age of child victims, types of maltreatment, and other perpetrators who were associated with her in the same report.

Identifying perpetrators with prior CPS experience

We constructed a multiyear combined file for 2005–2018 for the 27 states, and then restructured it to identify all unique females in the age range of our perpetrator sample who appeared in the file as both alleged victims and perpetrators. We matched the unique perpetrator dataset created during the prior step with this dataset to identify which perpetrators in our sample had prior experience with CPS and were reported to NCANDS since 2005. Among the 8205 unique perpetrators, 5080 had prior CPS experience.

Historical File

We then matched our unique perpetrator dataset again to the multiyear data file (2005–2018), including all records associated with each unique child ID, to identify whether they had prior CPS experience and to capture the full range of their experience with CPS since 2005. We conducted data quality checks to confirm that all records showed the same child birthdate, age, and sex. The 5080 unique perpetrators with prior CPS experience were reflected in 20,783 records. We combined data from all records for each child and created aggregated descriptors of the child’s experience with CPS during this 14-year time frame including whether the child was ever reported to CPS for each type of maltreatment and whether any of these maltreatments were substantiated, the overall number of reports, and whether reports were experienced at each age and during each year.

Combined perpetrator and child analysis file

We combined the above files into one that included both the characteristics of the adolescent mother as a perpetrator during 2016–2018 and any reports back to 2005 in which she was an alleged victim of maltreatment.

Analyses

We examined characteristics of our sample of adolescent mother perpetrators and their experiences with CPS as subjects of substantiated or unsubstantiated CPS investigations during the prior 14-year period. For our retrospective analysis, we calculated odds ratios (Tenny & Hoffman, 2020) to estimate the likelihood that an adolescent mother perpetrator had prior CPS involvement based on her demographic characteristics and current experience as a perpetrator as well as to examine whether there was an association between current maltreatments by the adolescent mother perpetrators and their prior experiences as victims of specific maltreatment types.

Results

Among the 8205 adolescent mother perpetrators in our sample, 6257 (76%) were age 18 or 19, 1693 (21%) were 16 or 17, and 255 (3%) were age 12 to 15. The mean age was 18.1 years. The largest proportion of the adolescent mother perpetrators (39%) were White, with 29% Hispanic and 19% African American. Adolescent mother perpetrators were associated with between one and four victims (M = 1.2 children, s.d. = 0.47). The victims were mainly very young (M = 9.6 months, s.d. = 12.1) with 91% of mothers having maltreated a child aged one or younger. Most of the adolescent mother perpetrators (88%) had neglected their children, while 13% physically abused their children, and fewer than 1% sexually abused their children (see Table 1). Forty-six of the adolescent mother perpetrators (fewer than 1%) were found responsible for a fatal maltreatment.

Table 1 Young mother perpetrator descriptive statistics and characteristics
Full size table

Proportion of Adolescent Mother Perpetrators with Prior CPS Experience

Figure 1 displays the proportion of adolescent mother perpetrators in our sample who had prior experience with CPS since 2005 and whether this experience included a substantiated maltreatment allegation, an unsubstantiated maltreatment allegation, or both. Nearly two-thirds (62%) had CPS experience with as alleged victims during this time. Nearly one-third (32%) had at least one report with a substantiated maltreatment, 27% experienced both substantiated and unsubstantiated reports, and 30% only experienced CPS reports that were unsubstantiated.

Fig. 1
figure 1

Adolescent mother perpetrator prior CPS experience (N = 8205)

Full size image

Table 2 shows that most adolescent mother perpetrators with prior CPS experience had multiple victim allegation reports during childhood; approximately 75% experienced more than one report, and nearly one quarter (24%) experienced six or more reports. We also found that these adolescent mother perpetrators had CPS experiences over multiple years; 31% experienced maltreatment reports during 4 or more different years. Forty-four percent had experience with CPS during both adolescence and childhood, and 45% during only adolescence; the smallest proportion (11%) had CPS experience only during childhood. Adolescent mother perpetrators may have experienced reports of more than one type of maltreatment, either substantiated or unsubstantiated, during the 14-year period of this study. Adolescent mother perpetrators with prior CPS experience in our sample experienced more reports of neglect relative to other maltreatment types. Forty-three percent experienced a substantiated report of neglect, and 78% experienced an unsubstantiated report of neglect. Twelve percent experienced a substantiated report of physical abuse, and 45% experienced an unsubstantiated report of physical abuse. Two-thirds of the adolescent mother perpetrators with prior CPS experience in our sample (66%) experienced reports of more than one maltreatment type, either substantiated or unsubstantiated, during this 14-year period.

Table 2 Perpetrator prior CPS history characteristics (n = 5080)
Full size table

Characteristics Associated with Prior CPS Experience

To explore characteristics of adolescent mother perpetrators that were associated with prior experiences with CPS, Table 3 presents odds ratios (ORs) and 95% confidence intervals for the estimated likelihood that an adolescent mother perpetrator has had prior CPS involvement as a function of her demographic characteristics and current experience as a perpetrator. Odds ratios above 1.00 indicate that an adolescent mother perpetrator is more likely and odds ratios below 1.00 indicate that she is less likely to have had prior CPS involvement based on specific characteristics. Table 3 shows that, compared with adolescent mother perpetrators ages 18–19, the odds of having prior CPS experience are significantly higher for both age groups 12–15 (OR = 4.84, CI [3.38, 6.94]) and 16–17 (OR = 2.78, CI [2.45, 3.16]). In terms of race, the odds of having prior CPS contact are significantly higher for African American (OR=1.39, CI [1.22, 1.57]) and Hispanic (OR =1.25, CI [1.12-1.4]) adolescent mother perpetrators and significantly lower for adolescent mother perpetrators identified as Asian or Pacific Islander (OR = 0.62, CI [0.39, 0.99]) and multiple race (OR = 0.78, CI [0.66, 0.91]) compared with those identified as White. There was no significant association between having prior CPS experience and being identified as American Indian/Alaska Native compared with being identified as White. Finally, the odds of having prior CPS experience were significantly higher for adolescent mother perpetrators when acting with a relative or nonrelative (i.e., not a romantic partner or the child’s father) (OR = 1.32, CI [1.07, 1.63]) than for those acting alone, and were significantly lower for those acting with a male partner (OR = 0.88, CI [0.8-0.96]).

Table 3 Perpetrator characteristics associated with prior CPS experience (n = 8205)
Full size table

Maltreatment Type Across Generations

Finally, we examined the relationship between the type of maltreatment experienced as a child and the type of maltreatment perpetrated. Table 4 shows the associations between the type of maltreatment perpetrated and the type of maltreatment experienced as a child among adolescent mother perpetrators with prior CPS experience. While perpetrators in our sample may be associated with multiple maltreatment types, the majority (90%) were associated with only one maltreatment type. For this analysis, we divided the sample of perpetrators with prior CPS experience into two mutually exclusive groups: those who physically abused their children (n = 511) and those who were perpetrators of other types of maltreatment did not physically abuse their children (n = 4565). This analysis excluded four perpetrators of sexual abuse. We calculated the ORs and 95% confidence intervals for the estimated likelihood that perpetrators in each of these two groups experienced each of the different maltreatment types as children. For adolescent mother perpetrators who were physically abusive, the odds are significantly higher for having been physically abused (OR = 1.4, CI [1.1, 1.8]) or sexually abused (OR = 1.73, CI [1.4, 2.2]) as a child compared with adolescent mother perpetrators who were not physically abusive. For adolescent mother perpetrators who were physically abusive, the odds are significantly lower for having been neglected (OR = 0.56, CI [0.5, 0.7]), having experienced psychological or other maltreatment (OR = 0.55, CI [0.4, 0.8]), or having experienced an unsubstantiated maltreatment allegation (OR = 0.72, CI [0.6, 0.9]) as a child compared with adolescent mother perpetrators who were not physically abusive.

Table 4 Prior maltreatment experience of perpetrators by maltreatment type (n = 5076)a
Full size table

Discussion

This research demonstrates that more than one-half of adolescent mother perpetrators, and more than three-quarters of those younger than 18, had a prior CPS experience as alleged victims. While our retrospective single cohort design did not allow for estimation of the prevalence of CPS experience among other adolescent mothers or young women in this age group, NCANDS estimates the prevalence of CPS experience in the general population; for FFY 2019, an estimated 4.8% of children in the population received either an investigation or assessment (U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau, 2021). Furthermore, the adolescent mother perpetrators in our sample experienced maltreatment circumstances that are linked with the most adverse outcomes. Our methodology of looking back 14 years does not allow us to confirm the full childhood maltreatment history for most of our sample, but we have enough information to know that many of these adolescent mother perpetrators experienced a pattern of maltreatment reports that could be described as chronic. Three-quarters experienced more than one report and nearly one-half experienced 3 or more reports; nearly one-third experienced maltreatment reports during more than 4 different years. Chronic maltreatment, especially when it begins during infancy and early childhood, has been linked with more maladaptive outcomes (Manly et al., 2001), and children who experienced chronic maltreatment of more than one type have also been shown to have higher levels of aggressive behavior, depression, anxiety, and emotion dysregulation than non-maltreated children (Warmingham et al., 2019). The majority of adolescent mother perpetrators with prior CPS experience in our sample (89%) experienced maltreatment reports during adolescence, either exclusively or in addition to during childhood. Thornberry et al. (2012) found that maltreatment during adolescence predicted subsequent maltreatment of one’s own children more so than maltreatment only during childhood. Ben-David et al. (2015) found that parents who experienced sustained exposure to maltreatment, particularly chronic neglect, were at increased risk for abusing their own children compared with parents who had experienced single incidents of abuse or neglect. Furthermore, Bartlett et al. (2017) found that mothers’ childhood history of multiple maltreatment types was associated with increased maltreatment of their children. These findings are consistent with the theory that intergenerational maltreatment results from disrupted emotional development due to chronic stress (Jonson-Reid et al., 2012).

Findings of prior victimization for perpetrators of different maltreatment types provided some additional insight into the nature of intergenerational maltreatment transmission. This issue is challenging to study because victims often experience more than one type of maltreatment in their lifetime. For this analysis, we examined the maltreatment histories of perpetrators of physical abuse only from those perpetrators of non-physical abuse. Notably, the maltreatment experienced during childhood may have included multiple types of substantiated maltreatment as well as multiple allegations of maltreatment, subsequently unsubstantiated. Widom et al. (2015) found that offspring of maltreated mothers are at risk for neglect, however, they did not find evidence for intergenerational transmission of physical abuse. Consistent with Widom et al. (2015), our findings support evidence for intergenerational transmission of neglect, but our findings also indicate that physical abuse may be intergenerationally transmitted. Perpetrators of physical abuse also have significantly higher odds of being a victim of sexual abuse. These findings point to the importance of recognizing specific elements of young mothers’ past trauma when identifying appropriate services to prevent future maltreatment.

Our single cohort design includes adolescent mother perpetrators, some of whom do and some who do not have prior CPS experience as alleged victims. The results provide some insight into characteristics and circumstances of adolescent mothers who maltreat their children but do not have prior experience as victims of alleged maltreatment themselves. For example, adolescent mother perpetrators connected with an additional perpetrator identified as the child’s father, stepfather, or the mother’s partner were less likely than adolescent mother perpetrators acting alone to have had prior experience with CPS. The presence of these men in the mothers’ lives may have contributed to the risk that resulted in their being identified as a perpetrator. Future research with a similar methodology could explore the prior CPS experience of and maltreatment prevention programs for young fathers, which was not included in this study. Furthermore, adolescent mothers who were connected to other perpetrators who were their relatives were significantly more likely than adolescent mother perpetrators acting alone to have prior experience with CPS; these young mothers are likely continuing to live with their abusive family of origin. The parents or relatives who previously (or concurrently) abused the adolescent mother perpetrators are presumably abusing the adolescent mothers’ children now as well.

Study Strengths and Limitations

Our use of administrative data to investigate intergenerational maltreatment overcomes some of the methodological weaknesses that Child Welfare Information Gateway (2016) and Thornberry et al. (2012) identified in earlier studies. Rather than relying on individuals’ self-reported recollections, we used data collected in real time by child welfare professionals based on established policies and practice and according to national standards. However, we recognize that the 27 states included in our dataset may have different state-level policies for screening in, investigating, and substantiating maltreatments. Furthermore, while administrative CPS data may be more objective than parental or child self-report of child maltreatment, a disadvantage is that these records include only maltreatments that are observed, reported, and substantiated and therefore exclude maltreatments that go unnoticed by professionals or cases not severe enough to be substantiated by CPS (Buisman et al., 2020).

Thornberry et al. (2012) questioned the validity of longitudinal studies of intergenerational continuity that did not include sufficient exposure time for either or both generations. Our research included data over a 14-year period for the first generation of maltreatment. To include Child Files prior to 2005 would have resulted in excluding additional states due to data quality concerns. Even with this relatively long timeframe, we are still lacking information from infancy and early childhood for most perpetrators in our sample, as they were born between 1995 and 2005. States continue to improve their reporting each year, so the ability to include a longer time span in future research will continue to grow.

The pattern of intergenerational maltreatment demonstrated here may also be attributed to a surveillance effect. Widom et al. (2015) suggested that evidence of intergenerational maltreatment may simply reflect CPS systems’ disproportionate scrutiny of families with past histories of child maltreatment and dismissal of reports from families who are reported for the first time. However, Drake et al. (2017) found that unique reports by mental health or social service providers accounted for only a small percentage of total reports, suggesting that surveillance does not have a disproportionate effect on maltreatment reporting patterns.

We recognize that our study shares limitations noted with other studies of the intergenerational transmission of maltreatment that relied on retrospective or cross-sectional designs in which childhood maltreatment history is collected from a sample of parents (Madigan et al, 2019; Pears & Capaldi, 2001). Our single cohort design included only adolescent mothers who had already been identified as perpetrators and did not include adolescent mothers not identified as perpetrators, or adolescents who were neither mothers nor perpetrators. We do not have information about individuals who had experience with CPS as an alleged victim of maltreatment but did not subsequently become parents or who became parents but did not maltreat their children. In future studies, our methodology of linking child and perpetrator identifiers in NCANDS can be used with a prospective design to track the subsequent experience as perpetrators among children who have experience with CPS as subjects of CPS investigations. Such a design would involve millions of records and would require more complex dataset construction but would ultimately provide a more complete insight into the nature of intergenerational maltreatment.

Recommendations for Practice

These findings support the importance of assessing new adolescent mothers for unresolved trauma and adverse childhood experiences during prenatal and newborn pediatric health visits (Bartlett et al., 2017; Garner et al., 2011; Iyengar et al., 2014; Schickedanz et al., 2018). While Child Welfare Information Gateway (2016) cautioned against submitting already-traumatized parents and their children to unwarranted intervention, any specific parenting services for young parents who have experience with maltreatment allegations should be provided with a trauma-informed approach. The Family First Prevention Act (Family First Prevention Services Act, 2018) will be a helpful resource to this end. This legislation is aimed at preventing at-risk families from entering the child welfare system by supporting evidence-based mental health services, substance abuse treatment, and in-home parenting skills (Holbrook & Hudziak, 2020; Tung et al., 2019). Implementation of a trauma-informed approach would consider parents’ prior history of maltreatment as victims or alleged victims.

The National Child Traumatic Stress Network (NCTSN) has developed and implemented a range of trauma-informed approaches to guide optimal service provision to children and families affected by trauma. Many of these approaches overlap in their content and practices. Thus, clinicians should evaluate the intervention objectives and practice elements in terms of their fit with the specific needs and preferences of the individual. In a trauma-informed child welfare system, all parties involved recognize and respond to the impact of traumatic stress on the children, caregivers, and service providers who have contact with the system and infuse and sustain trauma awareness, knowledge, and skills into their programs, organizational cultures, practices, and policies. They collaborate to maximize physical and psychological safety, facilitate the recovery of the child and family, and support their ability to thrive (Chadwick Trauma-Informed Systems Project, 2013; The National Child Traumatic Stress Network, 2016). This enhanced understanding of the needs of children who have experienced trauma has resulted in increased integration of trauma-informed approaches (Child Welfare Information Gateway, 2015). While trauma-informed parenting programs that train parents to respond to children’s trauma are widespread, a need remains for programs that address parents’ own trauma experiences that may lead to maltreatment of their own children. Developing such programs may be inherently challenging, as those with a history of maltreatment may be inhibited from trusting and relying on others for support, while building social support networks and reducing social isolation can be a critical mechanism for breaking the cycle of maltreatment (Berlin et al., 2011).

For an adolescent mother with a history of maltreatment, a trauma-informed approach focuses on teaching skills to help process thoughts and emotions associated with past events as well as train safety and parenting skills. In turn, trauma-informed techniques may serve to intervene immediately from the birth of the child, rather than after maltreatment has been reported. The Combined Parent–Child Cognitive Behavioral Therapy (CPC-CBT) is an example of an effective trauma-informed therapeutic intervention that addresses the intergenerational transmission of abuse. CPC-CBT is a short-term and strengths-based therapy program for children ages 3 to 17 and their parents (or caregivers). Therapy involves helping the child heal from the trauma of maltreatment and empower parents to use effective noncoercive parenting strategies and strengthen the parent–child relationships while helping families stop the cycle of violence (Runyon & Deblinger, 2014). Studies among CPS-involved children have found evidence supporting the efficacy of the Attachment and Biobehavioral Catch-up (ABC) parenting program. ABC is a trauma-informed parenting program that has effectively increased rates of secure attachment among infants living with their CPS-involved birth parents (Caron et al., 2016).

This research also supports the need for focused prevention programs for adolescent girls with CPS experience who have not yet become parents. While all children who are maltreated require support and services, adolescent girls who live with abusive parents or caregivers are at additional risk (Garwood et al., 2015). Preventing pregnancy among adolescent girls with a history of maltreatment can ultimately prevent them from maltreating their infants and young children. In addition to education on pregnancy prevention and access to family planning, adolescent girls should be exposed to healthy family relationships and models of positive parenting. Therefore, a trauma-informed approach is critical for programs aiming to prevent pregnancy among youth with a history of maltreatment. Pregnancy prevention programs for adolescents must consider the possibility that participants have experienced childhood trauma and may continue to engage in risky sexual behavior despite having received accurate sexual education (Schladale, 2013).

The U.S. Department of Health and Human Services Office of Population Affairs teen pregnancy prevention program (TPP) funds organizations that implement evidence-based programs to prevent teen pregnancy across the USA. Some of these programs specifically focus on youth who are in foster care or who have experienced trauma. For example, Power Through Choices, a 10-session prevention curriculum specifically designed for adolescents (ages 13—18) involved in the child welfare or juvenile justice system, has been shown to be effective in reducing teen pregnancy and associated sexual risky behaviors (U.S. Department of Health & Human Services, Office of Adolescent Health, 2016). The Practice Self-Regulation for Sexual Health (PS-R) program, designed for youth who are affected by adverse childhood experiences, provides intervention programs that help youth understand the impact of trauma on their sexual decision-making and sexual health. In addition to supporting safe sex practices and reducing risky sexual behavior, the program supports healthy relationships by providing its participants tools to help regulate their emotions. PS-R is currently being evaluated in two federally funded randomized controlled studies. A wide range of other successful pregnancy prevention programs may be adapted to address the needs of specific subgroups of vulnerable individuals and can provide educational resources while also ameliorating adolescent trauma symptoms. However, adaptations to teen pregnancy prevention programs should only be made if they do not compromise the program’s core components and should only be undertaken with the support of an expert in curriculum development, pedagogy, or health behavior change theory (Rolleri et al., 2014).

Study Implications and Conclusions

The results presented here offer new directions for future research and bridge gaps in the literature regarding intergenerational maltreatment on a national scale. Furthermore, the findings of this study demonstrate that there is a need for key stakeholders and policy makers to develop and implement different types of prevention programs to reduce intergenerational maltreatment risk factors among adolescent mothers. As past findings show, prevention services are effective in reversing the negative effects of traumatic events.

Finally, this research unveils a promising methodology to study intergenerational maltreatment on a national scale. NCANDS is now in its 30th year of collecting state CPS data, and child-level data is available for nearly 20 years for many states. By linking child and perpetrator identifiers across multiple years, we have a rich opportunity to study patterns of maltreatment over time, both with retrospective analyses of perpetrators’ prior CPS experiences and ultimately examining characteristics of maltreated children who later become perpetrators.