Abstract

Background:

Child maltreatment has profound adverse effects on various aspects of an individual's development and can even influence the next generation.

Objective:

This study investigated risk factors for child abuse potential within a population of highly burdened families and explored how these factors interact in the intergenerational transmission of maltreatment.

Methods:

Families receiving support from the German early prevention program “Frühe Hilfen” completed clinical interviews and standardized questionnaires at one time point. All subjects included met at least one risk factor for child abuse (inclusion criteria: low socioeconomic status, parental mental illness or teenage parenthood). Participants came from diverse cultural backgrounds. Child abuse potential was assessed using the CAPI, a screening instrument that measures parental stress and associated risk factors rather than actual abuse outcomes.

Results:

A multiple regression analysis identified parental psychopathology (b = .52, p = .024), attachment quality (b = 1.07, p = .001) and low socioeconomic state (b = −.83, p = .018) as key risk factors for child abuse potential. A mediation analysis indicated an indirect pathway whereby parental child maltreatment experiences were associated with poorer mental health (b = .32, p = .002), which in turn was linked to lower attachment quality (b = 1.71, p = .006) and higher child abuse potential (b = .12, p = .002). Affective disorders emerged as particularly significant risk factors among all mental disorders (b = 6.39, p = .015).

Conclusion:

These findings underscore the need for accessible psychiatric and psychotherapeutic support for parents with histories of childhood maltreatment to support healthy parent–child relationships.

Clinical Trial registration:

https://www.drks.de/drks_web/, German Clinical Trials Register DRKS00022075.

Introduction

Child maltreatment1 (CM) or child abuse1 is internationally recognized as a serious public health problem. According to a recent meta-analysis, which included a total of 25 countries (from the following regions: North America, Europe, Developed Asia Pacific, East Asia, South Asia and America, West Asta and Africa), the prevalence of childhood exposure to physical violence is estimated between 16 and 17 percent (1). In 2023, the number of reports of child endangerment through neglect, psychological, physical, or sexual abuse reached a new high in Germany (73,700 cases reported), representing an increase of about 1,400 cases (+2%) compared to 2022 (with estimates suggesting a higher increase due to missing reports) (2). The COVID-19 pandemic has exacerbated the risk of child abuse due to lockdown measures, which have led to social isolation, unstable financial situations for families, and increased parental stress. Vulnerable families, in particular, have been disproportionately affected by these challenges (3).

CM often has profound negative long-term consequences, resulting in adverse physical, cognitive, psychological, and social outcomes for affected children (4). Especially chronic maltreatment (5) and exposure to various types of maltreatment have been found to contribute significantly to these negative outcomes (6). Early prevention programs are designed to support children and families to address risks and promote positive development before problems become severe. Despite existing prevention and intervention programs, their effectiveness remains limited. A meta-analysis indicated that early prevention programs yield only small and only short-term effects on parenting behavior, parental mental health and child's developmental outcome (7, 8). Given these challenges, the heterogeneity of current research findings and the variability in samples, methodologies, and measurement tools, there is a pressing need for further investigation into the risk factors and mechanisms underlying child maltreatment to develop more effective strategies for prevention and intervention (9, 10).

Risk factors for child abuse

Most identified risk factors for CM have been found at the parental level (11). For instance, low family income (r = .1662) is a known factor for child abuse and neglect. There is also evidence that socially isolated parents (r = .0371) may be more prone to maltreat their children (12). Parenting stress is another significant risk factor for CM. A study by Crum and colleagues examined the relationship between parental stress, abuse potential, and children's social and behavioral outcomes, finding a consistent association between increased parental stress and higher abuse potential (13). Parents with mental health disorders (r = .2591) are potentially at an elevated risk of maltreating their children; for example, parents with depression are significantly more likely to have children for whom maltreatment was reported compared to parents without depressive symptoms (14, 15). Furthermore, parents who have experienced maltreatment during their own childhood [here referred to as early-life maltreatment; (ELM, r = .1821)] are at increased risk of maltreating their children, potentially contributing to the intergenerational transmission of abuse and neglect (16–18). A meta-analysis of 84 studies found that the risk of child maltreatment in families with parents who experienced ELM was nearly three times higher than in families without such histories. A recent cross-sectional study corroborates these findings by examining maltreatment transmission across three generations (grandparents, parents, and young adults), demonstrating continuity of maltreatment across all three generations with a medium effect size (19). Several theoretical mechanisms have been proposed to explain this transmission. From a social learning perspective (20), children who are exposed to harsh or abusive parenting may internalize these behaviors as normative strategies for discipline or conflict resolution, later reproducing them in adulthood. In parallel, the stress sensitization hypothesis suggests that early exposure to chronic or severe stress can dysregulate biological and psychological stress response systems, rendering individuals more reactive to later stressors such as parenting challenges or socioeconomic adversity (21, 22). This heightened reactivity may increase the risk of maladaptive coping or aggressive responses under pressure. However, contemporary empirical research indicates that while such mechanisms exist, they do not determine outcomes. Therefore, intergenerational models are best understood as probabilistic rather than deterministic frameworks (23).

There is a slight tendency in the literature showing that emotional maltreatment over physical or sexual abuse might be even more harmful in the intergenerational transmission of maltreatment (9, 24, 25). Whereby all types of maltreatment are considered to be a risk factor for the development of depression in adulthood, emotional maltreatment was additionally more strongly related to depression severity according to a larger meta-analysis (26). A recent systematic review by Younas and Gutman highlights several key risk factors at different levels. On a family level, the number of children in the family (r = .186), level of social support (r = .037) and partnership status (r = .285) emerged as most important risk factors. On a parental level low socio-economic status (SES, r = .166), young parental age (r = .140), substance abuse (r = .127), parental health issues (r = .259), parental stress (r = .184) and parental history of childhood maltreatment (r = .182) were identified as the most important risk factors (10). As these risk factors (level of social support, partnership status, SES, parental age, substance abuse, parental health issues, parental stress, parental history of child maltreatment) were most frequently associated with abuse potential in different studies, they were selected to be included in a joint analysis for the current study.

Influence of interrelated risk factors

Research examining the interplay of various risk factors has shown that ELM experiences significantly increase the risk of clinically relevant depressive symptoms in parents (27–29). It is well established that ELM and depression often co-occur (30), suggesting that parental psychopathology may be a mediator in the intergenerational transmission of CM (24, 31). In addition to affective disorders, other psychiatric disorders associated with ELM experiences, such as borderline personality disorder (31, 32), substance abuse (33), post-traumatic stress disorder (34), and anxiety disorders (35, 36) could also increase the risk of child abuse. This link between parent's psychopathology and child abuse potential may be further mediated by the quality of the parenting behavior, difficulties in emotion regulation or parental competencies in reflective functioning (37, 38). Attachment, in particular, is often considered an additional mediator linked to parental mental health issues (39–43). Early in life, sensitive parents lay the foundation for a secure attachment by promptly and appropriately serving their infant's needs (44). However, children of parents with mental health issues, especially those exhibiting depressive symptoms, face an elevated risk of adverse developmental outcomes. A substantial body of evidence indicates that parents with depression are less responsive to their infants' cues, leading to poorer dyadic interactions and, consequently, weaker attachment quality (45). Attachment theory also suggests that parents who have ELM are at risk of developing poor attachment relationships, both with their parents and later with their own children (46). This is because their own caregiving experiences were marked by inconsistency and a lack of sensitivity, which led to the development of internal working models that view others as unreliable. These models then influence how these parents interact with their own children, impairing their ability to be emotionally present and apply positive discipline strategies during stressful or conflictual situations (47). An integrated perspective assumes that child abuse potential emerges from the interaction of parents’ early experiences, current psychological functioning, and parent–child relationship quality, rather than from isolated risk factors. This conceptual model directly informed the study hypotheses and guided our mediation analyses examining whether parental psychopathology and attachment quality statistically account for the association between parents' own experiences of childhood maltreatment and child abuse potential. Importantly, this model is probabilistic rather than causal and is tested here using cross-sectional mediation analyses to examine whether observed associations are consistent with the proposed theoretical ordering. Against this conceptual background, the present study further extends existing work by focusing on a highly burdened sample of families from Germany—a group that has received limited empirical attention despite the existence of broader international datasets. Only one other recent study examined multiple risk factors for CM within a German sample; however, this was a different sample from the one used in the present study. The focus here was primarily on early life maltreatment experiences among parents (necessary inclusion criterion) as a risk factor, which did not necessarily have to be present in this study. Moreover, while the previous study assessed the cumulative occurrence of risk factors, it did not explore their interrelations or underlying mechanisms (39). In the international context, there are more studies, and ours was able to successfully replicate some of the results. Notably, we employed structured clinical interviews to rigorously assess a wide spectrum of psychiatric disorders. To our knowledge, no prior study has integrated this breadth of clinically diagnosed mental health variables into a single theoretically-guided analysis of child abuse potential. This comprehensive approach adds a valuable perspective to the global literature by offering detailed insights from a local context with robust diagnostic precision.

In this study, we aimed to identify key factors associated with child abuse potential measured with the Child Abuse Potential Inventory (CAPI) which primarily reflects parental stress as a key indicator of child abuse potential (not abuse outcomes). We were particularly interested in interconnections between key risk factors, focusing on the intergenerational transmission of CM, within a high-risk German sample of parent-infant dyads, collected during and shortly after the Corona pandemic between 2020 and 2024. All participants were enrolled in the early prevention program (EPP) “Frühe Hilfen,” which offers low-threshold support to families in challenging circumstances with children up to 3 years (see “Materials and Methods” for more details) (41). Through a comprehensive assessment, including a full diagnostic interview for various psychiatric disorders and data on caregivers' own experiences of ELM, we conducted an in-depth characterization of the sample. First, we aimed to identify the most relevant risk factors which increase the risk for CM within our sample of highly burdened families (Research question 1). Based on the literature (10) and our integrated theoretical framework we hypothesized that the most significant risk factors for child abuse include parental risk factors (e.g., psychopathology, ELM experiences, socioeconomic status and attachment difficulties). We also expected to find interrelations between these main risk factors leading to our second research question, namely, to explore the relationships between the identified risk factors. A systematic review from 2022 identified parental depression, emotional availability of the caregiver, attachment and ELM in particular as possible mediators for harsh parenting style and problem behaviour within the offspring (47). Therefore, we specifically tested whether the association between parent's own ELM and child abuse potential was mediated by parental psychopathology and by quality of attachment to the child (How are the relevant risk factors related to each other?—Research question 2). We hypothesized that the association between parental risk factors and child abuse potential would be mediated by psychopathology and attachment quality. We tested whether different types of ELM and different mental health issues in parents affect the risk of child abuse more or less strongly (Research question 3). Regarding parental mental health, we hypothesized to find associations especially between the risk of child abuse and the presence of personality disorders, substance abuse, and affective disorders in the parent (17, 48, 49). We further anticipated that ELM is a risk factor regardless of the type of abuse, with a possible tendency for emotional maltreatment to emerge as a more severe risk factor, showing more interrelations with considered risk factors (25).

Materials and methodsStudy design

Data were collected in Aachen and surroundings. Aachen is a city with about 260,000 inhabitants and is known by a large technical university https://www.aachen.de—Demografiemonitorings der Stadt Aachen (accessed on 21 January 2026). Ninety-three parent-child dyads were included in the present study. The study was part of a larger research consortium entitled “Understanding and Breaking the intergenerational cycle of abuse” (UBICA-II), funded by the Federal Ministry of Research and Education (BMBF). Parents of children aged ≤ 24 months or women in the third trimester of pregnancy who were supported by the early prevention program “Frühe Hilfen” were included in the study. Families who agreed to participate were visited at home to assess primary and secondary outcome measures at three measurement points. However, only data from the baseline survey (T0) was included in the analyses for this study. At T0, early intervention support has just begun (no more than 5 appointments).

The study was approved by the local Research Ethics Committee of the University Hospital RWTH Aachen (reference number EK 221/19) and was carried out in accordance with the Declaration of Helsinki. The study was registered at the German Clinical Trials Register (DRKS00022075 on 8 July 2020). For more information on the UBICA project see https://www.ubica.site/tp_fruehehilfen.html (accessed on 21 January 2026). The study protocol can be found under https://www.mdpi.com/2227-9067/11/3/267 (accessed on 21 January 2024). All participating parents provided informed consent for their own and their child's participation.

Early prevention program (EPP)

The EPP service “Frühe Hilfen” is available for all families with children up to the age of three, from pregnancy onwards. They offer psychosocial support and home visiting services to families in need. Almost half of the families usually contact the EPP themselves. The other half is referred to the EPP via pediatricians, clinics, midwives and the youth welfare offices. The aim is to enable every child to develop healthily and grow up without violence (for more information in the EPP see https://www.fruehehilfen.de/grundlagen-und-fachthemen/grundlagen-der-fruehen-hilfen/was-sind-fruehe-hilfen/, accessed 26 January 2026). We included two EPP teams in Aachen and the surrounding areas with comparable team structures and professional backgrounds (i.e., family midwives, social workers, or pediatric nurses). Staff members had to provide informed consent to participate in the trial.

Sample

We included all families (mothers or fathers, depending on who preferred to participate and their children) seeking support in the EPP who met the inclusion criteria (see Table 1) and were willing to participate. The risk factors were selected according to the descriptive data from clients in-need for longer term support of the EEP based on the annual reports (84).

CaregiverChild-

Must be supported by the EPP

-

Score >4 on at least one of the five subscales of the Adult Adolescent Parenting Inventory (AAPI version 2.1)

-

Or the presence of socioeconomic disadvantage (low socioeconomic status)

-

Or the presence of parental risk characteristics (mental illness)

-

Or teenage parenthood

-

And written informed consent provided by the caregiver

<24 months of age

EPP, early prevention program; AAPI, adult and adolescent parenting inventor.

MeasurementsChildhood experience of care and abuse questionnaire (CECA.Q)

Parental childhood maltreatment experiences were assessed at T0 using the German version of the CECA.Q, which is a retrospective self-report questionnaire on parent's behaviors towards children up to an age of 17 years. The German versions have been authorized by the author (A. Bifulco) of the original versions. The internal consistency of the CECA.Q in the German sample was.92 for Cronbach's α for both parents on the antipathy scale. For the neglect scale, Cronbach's α was .88 for the mother and .91 for the father (50). The questionnaire distinguishes between different types of maltreatment and assesses emotional maltreatment, physical and sexual abuse (51). Emotional maltreatment is assessed by 16 items related to parental care (8 items for antipathy and 8 for neglect) experienced by the mother and father, so that a total of four scales for emotional maltreatment can be formed. The items are rated on a 5-point Likert scale from “1 = yes definitely” to “5 = not at all” and can be added to a sum score (total score 16–80). There are defined cut-offs for each subscale. The presence of physical and sexual abuse is assessed by a “yes” or “no” question each. For the emotional maltreatment scale, Cronbach's alpha in the current sample was acceptable (α = .646). In order to compare the effects of mental vs. physical/sexual abuse, the four parental care scales of the CECA.Q (mother's and father's antipathy and neglect), were averaged and considered as a measure of emotional maltreatment in the current analyses [for a similar approach see Lehrl (52) and Hezhan et al. (53)]. The variable “presence of physical and/or sexual abuse” was dichotomized (1 = if physical and/or sexual abuse was experienced by the mother and/or father, or, in the case of sexual abuse, by others and 0 = if participants have not experienced physical or sexual abuse at all).

MINI DIPS

The MINI DIPS is a brief, structured diagnostic interview and offers a quick yet reliable diagnostic assessment of mental disorders according to DSM-5 and ICD-10 and was assessed at T0 only. The MINI DIPS Open Access provides an overview of the general burden of the persons examined and enables the documentation of important anamnestic information for practice and research. The DIPS interview has been regularly updated since the 1990s and its reliability, validity, and acceptance were repeatedly tested in large samples of outpatient, inpatient, and research populations and has demonstrated high interrater (κ ≈ .88–.98) and test-retest reliability (κ ≈ .80–.91) and good criterion validity (e.g., high agreement with SCID-5 and sensitivity/specificity >.85) (54). The interview contains dichotomous filter questions (yes/no) on various psychiatric disorders and distinguishes between current symptoms or past symptoms. If these filter questions are answered in the affirmative, these are followed by specific questions on the symptom criteria, predominantly in the “applies” and “does not apply” format. The ICD-10 or DSM-5 checklists can then be used to check whether the diagnostic criteria are met. If filter questions are answered in the negative, the specific questions are skipped, and the next disorder is addressed. Finally, a 9-point Likert scale is used to assess the level of function and the extent of impairment caused by the diagnosis(es) in question. In case of comorbid disorders, this item is answered only once considering the general level of function and extent of impairment across all diagnoses. The range from 0 to 3 is considered subclinical (slightly disruptive/not really). The range from 4 to 8 is considered clinically relevant (4: clearly disturbing, 6: strongly disturbing, 8: very strongly disturbing). In this paper, the functional level was included in the mediation analysis as a metric variable. In the present study, the internal consistency, as assessed by Cronbach's Alpha, was acceptable to good (α = .745). Interviews were conducted by an experienced psychologist which was supervised by a professor.

International personality disorder examination (IPDE)

The International Personality Disorder Examination/ICD-10 Module (German edition) is a semi-structured clinical interview to diagnose personality disorders as defined by the ICD-10 and DSM-IV classification systems. The interview is the most widely used screening tool for personality disorders based on worldwide field trials, which allows international application (55). It distinguishes between symptoms occurring before or after the age of 25. During implementation the clinician records the scores for each response on the IPDE Answer Sheet. After conducting the interview, the clinician can assign a definite, probable or negative diagnosis for each personality disorder. For this, precise scoring guidelines can be used. In the present study, only the areas of dissocial personality disorder and emotionally unstable personality disorder (borderline) were assessed. The agreement between different assessors is high. Kappa values are mostly between .65 and .72 for certain or probable diagnoses (56). The internal consistency, as assessed by Cronbach's Alpha, was acceptable for this study (α = .637).

Parenting stress inventory (PSI)

The PSI distinguishes between two main sources of parental stress: (1) child characteristics and behaviors that result in specific demands on parents and (2) limitations in parental functioning that impair the resources available to parents to cope with the demands of raising and caring for their child. The PSI contains 48 items corresponding to five subscales that record sources of stress that originate from the child's behavior and are associated with particular demands for the parents (distractibility/hyperactivity of the child, acceptability, demands, adaptability and mood) as well as seven subscales that record impairments of parental functions (attachment, social isolation, doubts about parental competence, depression, health, personal limitations, partner relationship). The PSI is a self-report (e.g., caregiver-report) questionnaire and can be used as a screening procedure to identify current threats to parent-child interaction at an early stage, including threats to the child's well-being due to high parental stress. T-norms are available for the main scales and for the overall scale of the PSI. Stanine norms are available for the subscales. The validity of the instrument has been examined in several studies that have reported substantial associations between PSI scores and various indicators of parental stress (57). In the present study, only the Stanine norms of the attachment subscale were included in the mediation analysis. Higher Stanine scores thereby indicate difficulties in bonding with the child. The attachment subscale measures a distant relationship with the child, which is expressed in an inability to empathize with the child and reliably assess his or her needs. A German version of the PSI showed high internal consistency reported in previous studies (α = .95) (57). Cronbach's alpha in our sample was good for the attachment subscale (α = .784). T-scores of the main parent scale were not included in the regression and mediation analysis because some of the measured constructs which are factored into the total score, e.g., depression, partner relationship and isolation, overlap with other variables in the models.

Child abuse potential inventory (CAPI)

The German version of the CAPI is a factor-analytically abbreviated and newly validated version of the Child Abuse Potential Inventory [CAPI, (58)]. It consists of 63 items and is a self-report (e.g., caregiver-report) screening tool for the detection of child abuse potential that is answered in a dichotomous agree/disagree format. Moreover, psychometric evaluation of the brief CAPI version (BCAPI) in German mothers and fathers supports its factorial validity (e.g., 59). The CAPI records current stress faced by parents as an indicator of possible risk to the child's well-being. This screening scale is primarily intended to identify and predict moderate to severe physical child abuse and neglect (It is not designed to assess the risk of sexual abuse or emotional neglect). The development of the CAPI was guided by the understanding that child abuse is influenced by a complex interplay of psychological and interpersonal factors. The scale includes the most predictive items to ensure it functions as a practical, reliable, and valid assessment tool for child abuse potential not abuse outcomes. The evaluation leads to a primary clinical scale for recording the parental stress level and three validity scales in order to record specific tendencies of response bias (social desirability, unreflected response behavior, inconsistent response behavior). In our sample, a total of 16 participants reached cut-off values on at least one of the three validity scales and therefore had to be excluded from the analyses. The raw value was converted into T-values (for risk samples) using norm tables and included in the current analyses. The internal consistency of the stress scale can be described as very good with a Cronbach's alpha of .91 (58). Cronbach's alpha for the current sample was good (α = .883) in the present sample.

Questionnaire for social support (FSozu)

The German F-Sozu self-report (e.g., caregiver report) questionnaire was conducted in the short form with 22 items. The F-Sozu operationalizes current social support as perceived or anticipated support from the social environment. The concept underlying the method is based on cognitive approaches and measures the subjective conviction of receiving support from others when needed as well as the assessment of being able to draw on resources from the social environment. The items are in statement form (e.g., “I have friends/relatives who are good listeners when I want to express myself”). The respondents indicate their level of agreement with these statements on a five-point Likert scale. Comparative norms for clinical and non-clinical groups are available for all questionnaire forms (60). Previous research has shown high internal consistency (α = .93) (60). Cronbach's alpha was acceptable to good for the social integration subscale (α = .693), which was used in the analysis.

Socioeconomic status (SES)

The socioeconomic status of the family was collected according to the GEDA study by the Robert Koch Institute (61). A total score representing the family SES is calculated by study stuff from the caregiver's highest educational qualification, the occupational status and the net household income. The scores range is between 3 and 21, whereby higher scores indicate a better SES.

Statistical analysis

Data analyses were conducted in IBM SPSS Statistics 29 (IMB corp., Armonk, USA). All statistical tests were two-tailed and p-values <0.05 were considered statistically significant. First, we calculated descriptive statistics for the variables of interest (see Table 2). Second, we calculated Pearson correlation coefficients among the main variables (see Table 7). To explore our first research question, a linear regression analysis was run to assess the association between selected risk factors and child abuse potential, as the dependent variable, in our high-risk group. Selected risk factors were: parental psychopathology, parental childhood maltreatment experiences, family socioeconomic status, parenting stress (attachment), origin, level of social support, parental age and partnership status of caregiver. Risk factors were assessed at T0 and chosen based on a recent systematic review on parental risk factors on child maltreatment (10). Therein the above-mentioned risk factors have shown the most frequent significant associations with child abuse across 68 included studies (10). The post-hoc power analysis given α = 0.05, total sample size = 75, number of predictors = 7 and effect size f2 = 0.15 revealed a satisfactory power of 95% for this regression analysis using G-Power 3 (62). To address our second research question, a serial mediation analyses was run using the PROCESS macro (63) to explore whether the link between parental childhood maltreatment experiences and child abuse potential were mediated by caregivers psychopathology as well as by the quality of attachment. The independent variables were the parental emotional maltreatment scales of the CECA.Q as well as the physical and sexual abuse scales. Mediator variables were the MINI DIPS functional level (Mediator 1, M1) and the PSI attachment subscale (Mediator 2, M2). Dependent variable was the child abuse potential (CAPI) (see Figure 1). In total, two models were calculated, one for the CECA.Q emotional maltreatment scale and one for the CECA.Q physical/sexual abuse scale. We conducted two separate mediation analyses to specifically examine the effects of emotional maltreatment, which has been underrepresented in prior research (64). In contrast, parental physical and sexual abuse frequently co-occurred in our sample (55% co-occurrence), making it statistically and conceptually appropriate to analyze them together in a single model. Running separate models for these highly overlapping variables could have reduced statistical power and obscured shared effects. Indirect effects were tested by bias-corrected bootstrapping of the 95% confidence interval (n = 5,000 runs). Statistical power for detecting indirect effects was evaluated based on the simulation results reported by Fritz and MacKinnon (65). For medium-sized effects of both the a and b paths, a sample size of approximately N = 71 is required to achieve .80 power when using bias-corrected bootstrap confidence intervals. Given the present sample size (N = 73), the study was sufficiently powered to detect indirect effects of at least medium magnitude.

N at T092Sociodemographic characteristicsChild's age in months (M, SD)2.61 (3.65)Caregiver's age in years (M, SD)30.84 (6.94)Sex of caregiver (f/m)(85/7)Sex of infant (f/m)(50/42)SES (M, SD)8.54 (4.14)Origin: EU including Switzerland (n/%)57 (62.0%) Turkey (n/%)6 (6.6%) Syria (n/%)3 (3.3%) Africa (n/%)8 (8.8%) Other (n/%)17 (18.7%)Number of siblingsa (M, SD)0.84 (1.05)Is the caregiver in a partnership?b (yes/no)(74/18)

SES, socioeconomic status.

a

Biological siblings, half-siblings and/or stepsiblings.

b

The partner did not necessarily have to be the biological father of the index child.

Mediation model with parental psychopathology and quality of attachment as mediators and visualization of direct and indirect pathways. CECA.Q, childhood experience of care and abuse questionnaire; PSI, parenting stress inventory; CAPI, child abuse potential inventory.

To explore our third research question, a linear regression analysis was conducted to assess which psychiatric disorders were associated with an increased potential for child abuse. The predictors included in the analyses were disorders that have been linked to heightened child abuse potential in previous studies (16, 19, 40, 66, 67). Therefore, affective disorders (depression), anxiety disorders, personality disorders, trauma disorders and substance abuse were included as binary predictors. The post-hoc power analysis given α = 0.05; Total sample size = 75, number of predictors = 5 and effect size f2 = 0.15 revealed a satisfactory power of 91% using G-Power 3 (62).

ResultsDescriptive data on parental psychopathology and own experiences of childhood maltreatment

Main study analyses were run for either 75 (regression analyses) or 73 (mediation analysis) participants since 16 cases had to be excluded since the data in the CAPI could only be interpreted to a limited extent and furthermore only complete cases were included in the analyses, which led to the exclusion of 3 further participants for the regression analysis (n = 74) and of 4 further participants for the mediation analysis (n = 73). Missing data were minimal, with at least 90 of 93 participants completing each measure. Cases with missing values were retained and descriptive inspection revealed no indications of systematic demographic or risk differences between complete and incomplete cases.

Overall, 53% of all caregivers reported any kind of ELM. With respect to the PSI, 37% of caregivers reported clinically relevant attachment issues with their children and nearly 60% reported clinically elevated levels of parenting stress on the parental level. Four percent of participants had a CAPI T-score above 60 indicating an increased risk for child abuse. Descriptive statistics for the entire sample are presented in Tables 3–5.

N92Functional level MINI DIPS (M, SD)3.24 (2.40) Subclinical symptoms ≤3 (n/%)48 (52.2%) Clinical symptoms ≥4 (n/%)44 (47.8%)CECA.Q Emotional maltreatment (M, SD)17.88 (7.17) Mother antipathy ≥25 (n/%)30 (32.3%) Mother neglect ≥22 (n/%)23 (24.7%) Mother physical abuse (n/%)25 (27.2%) Father antipathy ≥25 (n/%)19 (20.9%) Father neglect ≥24 (n/%)25 (27.5%) Father physical abuse (n/%)20 (21.7%) Sexual abuse (n/%)20 (21.7%)PSI Attachment (M, SD)9.83 (4.03) Attachment SN ≥ 7 (n/%)34 (37.0%) Overall parental stress (M, SD)75.45 (22.35)N75 CAPI (M, SD)36.6 (10.11)  CAPI T ≥ 60 (n, %)3 (4.00%)

Caregiver characteristics.

CECA.Q, childhood experience of care and abuse questionnaire; PSI, parenting stress inventory; SN, stanine; CAPI, child abuse potential inventory.

Mini Dips diagnosesN = 93%Is there a disorder in the anxiety spectrum? (yesa/no)(42/51)(45.2/54.8) Panic disorder? (yesa/no)(18/24)(19.4/25.8