Hostname: page-component-857557d7f7-nfgnx Total loading time: 0.001 Render date: 2025-11-21T14:12:09.294Z Has data issue: false hasContentIssue false
  • English
  • Français

Differential effects of childhood maltreatment types and timing on psychopathology in formerly out-of-home placed young adults

Published online by Cambridge University Press:  23 October 2025

Maria Meier Open the ORCID record for Maria Meier [Opens in a new window] ,
Inga Schalinski Open the ORCID record for Inga Schalinski [Opens in a new window] ,
Cyril Boonmann Open the ORCID record for Cyril Boonmann [Opens in a new window] ,
Nils Jenkel Open the ORCID record for Nils Jenkel [Opens in a new window] ,
Süheyla Seker Open the ORCID record for Süheyla Seker [Opens in a new window] ,
Delfine d’Huart Open the ORCID record for Delfine d’Huart [Opens in a new window] ,
Jörg M. Fegert ,
Vera Clemens ,
Marc Schmid Open the ORCID record for Marc Schmid [Opens in a new window]  and
David Bürgin Open the ORCID record for David Bürgin [Opens in a new window]
Maria Meier*
Affiliation:
Child and Adolescent Psychiatric Research Department, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland Department of Psychology, University of Konstanz, Konstanz, Germany
Inga Schalinski
Affiliation:
Faculty of Human Sciences, Institute of Psychology, Universität der Bundeswehr München, Neubiberg, Germany
Cyril Boonmann
Affiliation:
Child and Adolescent Psychiatric Research Department, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland Department of Child and Adolescent Psychiatry (LUMC Curium), Leiden University Medical Center, Leiden, The Netherlands Department of Forensic Child and Adolescent Psychiatry, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland
Nils Jenkel
Affiliation:
Child and Adolescent Psychiatric Research Department, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland
Süheyla Seker
Affiliation:
Department of Social Work, Stockholm University, Stockholm, Sweden
Delfine d’Huart
Affiliation:
Child and Adolescent Psychiatric Research Department, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland Department of Forensic Child and Adolescent Psychiatry, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland
Jörg M. Fegert
Affiliation:
Department of Child and Adolescent Psychiatry/Psychotherapy, University of Ulm, Ulm, Germany
Vera Clemens
Affiliation:
Department of Child and Adolescent Psychiatry/Psychotherapy, University of Ulm, Ulm, Germany
Marc Schmid
Affiliation:
Child and Adolescent Psychiatric Research Department, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland
David Bürgin
Affiliation:
Child and Adolescent Psychiatric Research Department, University Psychiatric Clinics Basel (UPK), University of Basel, Basel, Switzerland Jacobs Center for Productive Youth Development, University of Zurich, Zurich, Switzerland
*
Corresponding author: Maria Meier; Email: maria.meier@unibas.ch

Abstract

Background

Childhood maltreatment (CM) increases the risk for psychopathology and CM type, severity and timing are considered important modulating factors in this relationship. However, reported associations are heterogeneous and hardly considered vulnerable groups broadly exposed to CM.

Methods

We investigated the association between CM types and timing and psychopathology in formerly out-of-home placed young adults (N = 185; 32% women, age mean = 26.38 years, SD = 3.49). CM was assessed using the Maltreatment and Abuse Chronology of Exposure Scale. Conditional random forest regression was used to estimate the importance of CM types (abuse, neglect, peer victimization, and sexual abuse), timing (ages 3–18), and global measures (severity, multiplicity, and duration) on adult general, internalizing, and externalizing problems (Achenbach System of Empirically Based Assessment). We validated the results using diagnoses of mental disorders clustered with the Hierarchical Taxonomy of Psychopathology model.

Results

Global CM measures were stronger predictors of internalizing problems than CM type and timing. Abuse in early childhood was a stronger predictor of externalizing problems compared to global CM measures.

Conclusions

Considering CM type and timing might be valuable to guide maltreatment-informed interventions in therapeutic settings.

Keywords

childhood maltreatmentconditional random forest regressionHiTOPmachine learningpsychopathology

Information

Type
Research Article
Information
European Psychiatry , Volume 68 , Issue 1 , 2025 , e160
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press on behalf of European Psychiatric Association

Introduction

Childhood maltreatment (CM), including interpersonal victimization, abuse, and neglect, is a major public health concern and a strong risk factor for mental and physical diseases [Reference Copeland, Shanahan, Hinesley, Chan, Aberg and Fairbank1–Reference McLaughlin, Colich, Rodman and Weissman4]. CM is highly prevalent and associated with severe health issues and financial costs [Reference Hughes, Ford, Bellis, Glendinning, Harrison and Passmore5–Reference Hughes, Bellis, Hardcastle, Sethi, Butchart and Mikton7], making it a key target for prevention and intervention efforts. Individuals exposed to CM are at increased risk of a broad range of mental disorders, especially internalizing (e.g., mood and anxiety disorders or stress-related disorders) and externalizing disorders (e.g., substance use or disruptive behavior disorder). [Reference Green, McLaughlin, Berglund, Gruber, Sampson and Zaslavsky6, Reference Gardner, Thomas and Erskine8–Reference Zeanah and Humphreys11], likely mediated through transdiagnostic mechanisms [Reference McCrory, Gerin and Viding3, Reference Humphreys and Zeanah12]. They tend to exhibit earlier symptom onset, higher symptom severity and comorbidity, increased suicidal ideation, and poorer treatment response compared to individuals with the same diagnosis but without CM exposure [Reference Bürgin, Witt, Seker, d’Huart, Meier and Jenkel13–Reference Widom, DuMont and Czaja16].

Research on the sequelae of CM often focused on global CM scores that aggregate the severity of different CM types (e.g., sexual abuse) [Reference Koppold, Ruge, Hecker and Lonsdorf17–Reference Smith and Pollak19]. This approach revealed dose–response and type-specific effects [Reference Anda, Felitti, Bremner, Walker, Whitfield and Perry20–Reference Felitti, Anda, Nordenberg, Williamson, Spitz and Edwards22]. However, the frequent co-occurrence of exposure types constrains the inferences. Moreover, global CM scores conceal temporal information, limiting the exploration of questions such as whether CM has a stronger impact on health during sensitive developmental periods [Reference Dunn, Busso, Davis, Smith, Mitchell and Tiemeier23, Reference Dunn, Nishimi, Gomez, Powers and Bradley24].

For this reason, analyses need to acknowledge the type, timing, and duration of CM exposure. While commonly used CM measures rarely capture timing [Reference Koppold, Ruge, Hecker and Lonsdorf17, Reference Berman, McLaughlin, Tottenham, Godfrey, Seeman and Loucks25], some newer measures do (e.g., the Maltreatment and Abuse Chronology of Exposure scale [MACE], the International Society for the Prevention of Child Abuse and Neglect [ISPCAN] Child Abuse Screening Tool [ICAST-R], [Reference Clemens, Jarczok, Fegert, Brähler and Jud26] and the Stress and Adversity Inventory [STRAIN]) [Reference Jarczok, Lange, Meinck, Witt, Clemens and Fegert27–Reference Sturmbauer, Shields, Hetzel, Rohleder and Slavich29]. Research using these measures revealed that earlier exposure is often predictive of later exposure [Reference Hales, Saribaz, Debowska and Rowe30, Reference Russotti, Warmingham, Duprey, Handley, Manly and Rogosch31]. Making use of them may further our understanding of developmental risks, and help to target prevention and intervention efforts.

Nevertheless, analyses combining CM types and timing are hitherto scarce [Reference Schaefer, Cheng and Dunn32], likely because the complex dependencies of the data commonly violate the assumptions of widely used statistical tests. To remedy this, recent reports have utilized machine learning algorithms, such as conditional random forest regressions [Reference Schalinski, Teicher, Nischk, Hinderer, Müller and Rockstroh33–Reference Schalinski, Breinlinger, Hirt, Teicher, Odenwald and Rockstroh36]. While these analyses were based on data from patient populations, e.g., transdiagnostic inpatients [Reference Schalinski, Teicher, Nischk, Hinderer, Müller and Rockstroh33] or individuals suffering from PTSD [Reference Sicorello, Thome, Herzog and Schmahl34], no study has investigated the differential impact of types and timing of CM on psychopathology in individuals formerly placed in youth residential out-of-home care. This highly vulnerable group is underrepresented in research despite high rates of CM exposure [Reference Briggs, Greeson, Layne, Fairbank, Knoverek and Pynoos37, Reference Zelechoski, Sharma, Beserra, Miguel, DeMarco and Spinazzola38].

We thus aimed to investigate how type, timing, duration, multiplicity, and severity of CM predict general, internalizing, and externalizing psychopathology in a sample of young adults formerly placed in youth residential care [Reference Green, McLaughlin, Berglund, Gruber, Sampson and Zaslavsky6, Reference Gardner, Thomas and Erskine8–Reference Zeanah and Humphreys11]. We used conditioned random forest regression to consider all predictors simultaneously and test which of the CM characteristics were most important to predict psychopathology in young adulthood. Furthermore, we used competitive hypothesis testing and contrasted whether the importance of type and timing exceeded the importance of global indicators of CM.

Methods

Sample and procedure

The sample stemmed from the Swiss Study for Clarification and Goal-Attainment in Child Welfare and Juvenile Justice Institutions (German: Modellversuch Abklärung und Zielerreichung in stationären Massnahmen, MAZ.) assessed from 2007 to 2012. It comprised N = 592 participants aged 5 to 27 (mean age = 15.86 years, SD = 2.99 years; 32.1% female) from 64 residential care institutions in Switzerland [Reference Schmid, Kölch, Fegert and Schmeck39]. Between 2018 and 2020, the follow-up study Youth Welfare Trajectories: Learning from Experience (German: Jugendhilfeverläufe: Aus Erfahrung Lernen, JAEL) was conducted on a subset of participants, who completed questionnaires via an online platform (www.weaskyou.ch) and were invited to (semi-structured clinical) interviews in person [Reference Schmid, Fegert, Clemens, Seker, d’Huart and Binder40]. The cross-sectional analyses presented here are based on data of n = 185 participants (mean age = 26.38 years, SD = 3.49 years, range = 16–38 years; 34% women) who took part in the face-to-face assessments. This sample size is comparable to previous studies in clinical and high-risk samples that used conditional random forest regression (reported sample sizes ranging from 43 to 341, with approximately half of them reporting samples of up to 200 participants) [Reference Schalinski, Teicher, Nischk, Hinderer, Müller and Rockstroh33–Reference Schalinski, Breinlinger, Hirt, Teicher, Odenwald and Rockstroh36, Reference Juen, Hecker, Hermenau, Teicher, Mikinga and Nkuba41–Reference Schalinski, Teicher, Carolus and Rockstroh47]. Included participants neither differed from non-included participants in psychosocial characteristics and mental health burden assessed at baseline (MAZ., see Supplemental Table S1) nor in regard to follow-up data of criminal records from the Swiss Federal Bureau of Statistics (BFS) up until the year 2017 (see Supplemental Table S2).

Participants provided written informed consent and received vouchers of up to CHF500. The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Commission of Northwestern Switzerland (EKNZ, Ref. 2017–00718). Findings on the links between CM and personality functioning [Reference d’Huart, Hutsebaut, Seker, Schmid, Schmeck and Bürgin48], neurobiological markers [Reference Bürgin, Clemens, Varghese, Eckert, Huber and Bruttin49–Reference Meier, Kantelhardt, Gurri, Stadler, Schmid and Clemens53], and the emergence and stability of psychopathology [Reference Bürgin, Witt, Seker, d’Huart, Meier and Jenkel13] are published elsewhere.

Measures

Childhood maltreatment

We used the Maltreatment and Abuse Chronology of Exposure (MACE-X) scale to assess the severity and timing of CM [Reference Teicher and Parigger28]. For each of the 75 items, participants reported whether and at what age (from 1 to 18) they experienced it (e.g., parents or caregivers living in the household did hurtful things like insulting you). Based on this, exposure to 10 forms of CM was assessed, which were summarized to four CM types to ease interpretability: parental abuse (ABUSE, summarizing parental physical and verbal abuse, parental non-verbal emotional abuse, witnessed physical violence toward parents, and witnessing violence toward siblings), parental neglect (NEGLECT summarizing emotional and physical neglect), violence by peers (PEER summarizing peer emotional and peer physical violence), and sexual abuse (SEXA, indexed by the subscale familial and non-familial sexual abuse). The severity of CM type per life year was indexed as weighted score (ranging from zero to 10) to ensure comparability across the four CM types. Additionally, we computed three global measures of CM. For doing so, we used previously suggested thresholds [Reference Thekkumthala, Schalinski, Parigger, Ruf-Leuschner, Elbert and Schauer54, Reference Isele, Teicher, Ruf-Leuschner, Elbert, Kolassa and Schury55] from validation studies. The three global measures comprised: duration (DURATION, range: 0–18 years), indicating how many years participants were exposed to at least one of the ten original CM types above threshold [Reference Thekkumthala, Schalinski, Parigger, Ruf-Leuschner, Elbert and Schauer54, Reference Isele, Teicher, Ruf-Leuschner, Elbert, Kolassa and Schury55]; severity across all ten types of CM (SEVERITY, range: 0–100); and multiplicity (MULTI, range: 0–10), indicating the number of ten CM types experienced above threshold [Reference Thekkumthala, Schalinski, Parigger, Ruf-Leuschner, Elbert and Schauer54, Reference Isele, Teicher, Ruf-Leuschner, Elbert, Kolassa and Schury55]. CM types (ABUSE, NEGLECT, PEER, SEXA) per year (3–18) and global measures (DURATION, SEVERITY, MULTI) were used as predictors in the conditional random forest regression. Similar to previous studies, [Reference Sicorello, Thome, Herzog and Schmahl34, Reference Herzog, Thome, Demirakca, Koppe, Ende and Lis35] we excluded reports of years one and two due to their low reliability (infantile amnesia). [Reference Teicher and Parigger28, Reference Sicorello, Thome, Herzog and Schmahl34] Including these years did not change the interpretation of results.

Mental health problems

Participants filled in the Achenbach System of Empirically Based Assessment (ASEBA; the Youth Self-Report [YSR; < 18 years; N = 5] and the Young Adult Self-Report [YASR; > 18 years; N = 180]) [Reference Agrez, Winkler Metzke and Steinhausen56, Reference Achenbach, Rescorla, McConaughey, Pecora, Wetherbee and Ruffle57], consisting of 118 or 124 items respectively, which were answered on a 3-point Likert scale (0 = not true, 1 = somewhat or sometimes true, 2 = very or often true). We used the total score, indicating overall problems, and the subscales internalizing and externalizing problems. Raw values were transformed to T-values [Reference Agrez, Winkler Metzke and Steinhausen56, Reference Achenbach, Rescorla, McConaughey, Pecora, Wetherbee and Ruffle57], with values above 60 indicating clinically relevant symptomatology. T-values were used as outcomes in the conditional random forest regression.

Diagnoses of mental disorders

To complement the self-report data, we further employed the Structured Clinical Interview for the Diagnostic and Statistical Manual (DSM) of Mental Disorders 5th revision DSM-5 Disorders – Clinician Version (SCID-5-CV). The SCID-5-CV is a semi-structured clinical interview assessing the adult disorder dimensions outlined in the DSM-5. Responses were recorded (1 = presence, 0 = absence), and identified diagnoses were categorized according to the F-codes of the International Statistical Classification of Diseases and Related Health Problems 10th edition (ICD-10). Following the Hierarchical Taxonomy of Psychopathology (HiTOP) model, we scored whether the criteria for a current mental disorder were fulfilled, and grouped ICD-10 codes into internalizing (ICD-10 F32–F34, F38–F38, F41, F43, F40–F42, F50.1, F50.3, F51, F60.31, F93.3, F93.8, F93) and externalizing diagnoses (F10–F25, F60.2, F60.3, F60.4, F60.81, F60.0, F63.2, F90–F92, F94). Unlike traditional diagnostic systems, HiTOP focuses on psychopathological syndromes and broader spectra based on the observed co-occurrence of symptoms. For the analysis, we used the variables any diagnosis (1 = presence, 0 = absence), at least one internalizing disorder (1 = presence, 0 = absence), at least one and externalizing disorder (1 = presence, 0 = absence).

Statistical analysis

Analyses were performed in R version 4.3.0 and RStudio version 2023.3.0.386 using the packages haven [Reference Wickham, Miller and Smith58], lm.beta [Reference Behrendt59], party [Reference Hothorn, Hornik and Zeileis60–Reference Strobl, Boulesteix, Zeileis and Hothorn62], and caret [Reference Kuhn63]. Figures were created using ggplot2 [Reference Wickham64], corrplot [Reference Wei and Simko65], and patchwork [Reference Pedersen66]. Pearson’s correlation coefficients were calculated for descriptive purposes. Three conditional random forest regressions were modeled to predict overall, internalizing, and externalizing problems using the global (DURATION, SEVERITY, MULTI) and type- and timing-specific CM measures (ABUSE, NEGLECT, PEER, SEXA per life year between three and 18) while controlling for sex and age. We competitively tested the importance of global CM measures against timing- and type-specific measures. In linear regression models, a normal distribution of residuals is assumed, and a high collinearity critically limits the interpretability of findings; conditional random forest regressions do not rely on specific distributions and can handle a high number of collinear predictors [Reference Strobl, Boulesteix, Kneib, Augustin and Zeileis61, Reference Strobl, Boulesteix, Zeileis and Hothorn62]. Hence, they have repeatedly been used to analyze the type- and timing-specific effects of CM on psychopathology and other outcomes [Reference Schalinski, Teicher, Nischk, Hinderer, Müller and Rockstroh33, Reference Schalinski, Breinlinger, Hirt, Teicher, Odenwald and Rockstroh36]. Random forest regressions are a machine learning approach that estimates a specified number of decision trees (“forest”) by randomly drawing a subset of the data and a subset of the predictor variables for consideration at the nodes. To avoid overfitting, the model is trained on 75% of the data and evaluated on the remaining 25%. We modeled the conditional random forest regression as described earlier [Reference Schalinski, Teicher, Nischk, Hinderer, Müller and Rockstroh33] and estimated each predictor’s variable importance (VI) by sequentially permuting each predictor in the model, refitting the random forest, and determining the extent to which each permutation increased the mean square error (MSE), an index of the goodness of fit of the model. While permuting important predictors results in larger increases in MSE, permuting unimportant predictors should have a negligible effect. We used a total of 100 permutations with randomly drawn training and test data and subsequently averaged the VI estimates. To obtain the statistical significance of each VI, the random forest analysis was run 5,000 times using reshuffled outcome measures to calculate random chance VIs and a SD for each predictor. We used Z-tests to derive the probability of observing a high VI by chance.

To validate the results, we used the CM measures that were significant predictors in the conditional random forest regression and tested whether the strongest predictor per CM type could predict mental disorder status in adulthood using binomial generalized linear models.

Results

Descriptive analyses

Sample characteristics are listed in Table 1. From N = 185 participants, 87.6% experienced at least one CM type (ABUSE, NEGLECT, PEER, and SEXA) above-threshold [Reference Thekkumthala, Schalinski, Parigger, Ruf-Leuschner, Elbert and Schauer54, Reference Isele, Teicher, Ruf-Leuschner, Elbert, Kolassa and Schury55]. Mean severity of each type per year can be derived from Figure 1 and prevalence rates can be derived from Supplemental Table S3. The mean prevalence of peer violence was less than 10% for ages three to five, and the mean prevalence of sexual abuse was less than 10% for ages three to 18, which may comprise the informativeness of the models in this regard.

Table 1. Description of the sample

Note: *These placement-related characteristics were self-reported retrospectively in young adulthood as part of the JAEL study.

Figure 1. Weighted severity of exposure scores (theoretical range 0–10) per life year and category of childhood maltreatment, i.e., parental abuse (ABUSE) and neglect (NEGLECT), peer violence (PEER), and sexual abuse (SEXA). Error bars represent SE.

Participants with any diagnosis reported significantly more overall mental health problems (ASEBA total mean = 57.46, SD = 10.31) compared to participants without a diagnosis (mean = 48.97, SD = 9.45), t(160.94) = −5.73, p < .001, d = −0.85. Participants suffering from an internalizing disorder reported significantly more internalizing problems (ASEBA internalizing mean = 60.13, SD = 10.23) compared with those without an internalizing diagnosis (mean = 49.63, SD = 10.03), t(114.77) = −6.57, p < .001, d = −1.04, and participants suffering from an externalizing disorder reported more externalizing problems (ASEBA externalizing mean = 56.17, SD = 10.68) compared with their counterparts (mean = 50.00, SD = 9.02), t(181.96) = −4.25, p < .001, d = −0.62.

Overall, 61% fulfilled the criteria for at least one mental disorder, with 25.95% being diagnosed with both an internalizing and externalizing disorder, 6.49% with an internalizing, and 27.57% with an externalizing disorder, and 39.46% with neither. Global and type- and timing-specific CM measures were significantly interrelated and associated with self-reported mental health problems (cf. correlation matrix in Supplemental Figure S1).

Predicting overall mental health problems from CM type and timing

Parental abuse during ages three, five, six, and seven (early to middle childhood) significantly predicted overall mental health problems, with the strongest predictor being parental abuse at age six (VI = 1.63, p = .002). During adolescence, sexual abuse at age 18 (VI = 0.06, p = .020), and peer violence throughout ages 14 to 18 significantly predicted overall mental health problems (strongest predictor: peer violence at age 14, VI = 1.21, p = .009). Global CM severity (VI = 2.72, p < .001) and multiplicity (VI = 1.18, p = .009) significantly predicted overall mental health problems. Parental neglect and CM duration were no significant predictors. Estimates of the VI of all predictors are depicted in Figure 2A and in Supplemental Table S4.

Figure 2. Variable importance of the timing-specific effects of different childhood maltreatment types (solid lines) and the global childhood maltreatment severity, multiplicity, and duration measures (dashed lines) for overall mental health problems (A), internalizing mental health problems (B), and externalizing mental health problems (C). High variable importance implies that the variable is a strong predictor of the respective subscale of self-reported mental health problems. MACE SEVERITY = severity of exposure. MACE MULTI = number of individual scales above cut-off. MACE Duration = duration of above-threshold exposure.

Parental abuse at age six significantly predicted disorder status (HiTOP any disorder; OR = 1.30, beta = 0.69, SE = .13, p = .034) when controlling for sex and age. Neither sexual abuse at age 18 (OR = 1.00, beta = 0.01, p = .988), peer violence at age 14 (OR = 1.79, beta = 0.58, SE = 0.074, p = .068), severity (OR = 1.74, beta = 0.55, SE = 0.01, p = .097), nor multiplicity (OR = 1.65, beta = 0.50, SE = 0.07, p = .133) predicted disorder status.

Predicting internalizing problems from CM type and timing

Parental abuse at age six (VI = 0.78, p = .027) and sexual abuse at ages 17 and 18 significantly predicted internalizing problems, with severity at age 18 being the stronger predictor (VI = 0.09, p = .008). Peer violence during adolescence, at ages 14, 15, and 18 significantly predicted internalizing problems (strongest predictor was PEER at age 14, VI = 1.2, p = .007). CM severity (VI = 3.02, p < .001), multiplicity (VI = 2.07, p < .001), and duration (VI = 1.78, p = .006) as well as sex (VI = 1.78, p = .006) significantly predicted internalizing problems. Parental neglect was no significant predictor. Global CM measures were stronger predictors of internalizing problems as compared with the strongest type- and timing-specific effect (PEER at age 14). Estimates of the VI of all predictors are depicted in Figure 2B and Supplemental Table S5.

Parental abuse at age six (OR = 3.31, beta = 1.20, SE = .12, p < .001) and peer violence at age 14 (OR = 3.49, beta = 1.25, SE = .07, p < .001) significantly predicted internalizing disorder status (HiTOP internalizing disorder) when controlling for sex and age. Sexual abuse at age 18 was no significant predictor of internalizing disorder status (p > .05), but all global measures were significant predictors (SEVERITY: OR = 3.31, beta = 1.20, SE = .01, p < .001; MULTI: OR = 2.70, beta = 0.99, SE = .07, p = .006; DURATION: OR = 2.84, beta = 1.04, SE = .02, p = .003).

Predicting externalizing problems from CM type and timing

Externalizing problems were significantly predicted by parental abuse during early to middle childhood, i.e., at ages three, four, five, six, and seven, with the strongest predictor being parental abuse at age three (VI = 3.57, p < .001). In addition, sexual abuse at age 18 significantly predicted externalizing problems (VI = 0.03, p = .043). Moreover, experiencing peer violence at ages 14–18 (i.e., during adolescence) significantly predicted externalizing problems, with the strongest predictor being peer violence at age 17 (VI = 1.55, p = .001). CM severity (VI = 2.22, p = .001), and multiplicity (VI = 0.84, p = .025) significantly predicted externalizing problems. Parental neglect and CM duration did not significantly predict externalizing problems. The strongest type- and timing-specific effect (ABUSE at age three) was a stronger predictor of externalizing problems as compared with the global measures. Estimates of the VI of all predictors are depicted in Figure 2C and listed in Supplemental Table S6.

Parental abuse at age three (OR = 2.05, beta = 0.71, SE = .22, p = .025) and peer violence at age 17 (OR = 1.90, beta = 0.64, SE = .11, p = .041) were significant predictors of externalizing disorder status (HiTOP externalizing disorder) when controlling for sex and age. However, sexual abuse at age 18 and global CM severity, multiplicity, and duration did not significantly predict externalizing disorder status (all p > .05).

Discussion

We investigated the impact of types and timing of CM on general, internalizing, and externalizing psychopathology among young adults with previous experience of residential out-of-home care. Using conditional random forest regressions, we found that CM severity and multiplicity were important predictors of general psychopathology, next to type- and timing-specific effects of peer abuse in mid-to-late adolescence and sexual abuse in late adolescence. CM severity and multiplicity were the most important predictors of internalizing problems, followed by peer victimization in mid-to-late adolescence, with minor effects of early parental and late adolescent sexual abuse. Regarding externalizing problems, CM severity, parental abuse in early childhood, and effects of peer violence in mid-to-late adolescence were most important, with minor effects of sexual abuse in late adolescence. Most effects were replicated using data from clinical interviews. Our findings add a nuanced perspective on the association between CM and psychopathology in young adults with a history of youth residential care in Switzerland.

The high severity and co-occurrence of reported CM (87.6%) align with those of other samples of young adults in and after out-of-home care [Reference Briggs, Greeson, Layne, Fairbank, Knoverek and Pynoos37, Reference Rebbe, Nurius, Ahrens and Courtney67–Reference Fosse, Skjelstad, Schalinski, Thekkumthala, Elbert and Aanondsen69]. Studies using the same CM measure found comparable rates in psychiatric inpatients (90%) [Reference Schalinski, Teicher, Nischk, Hinderer, Müller and Rockstroh33], slightly higher rates in Norwegian adults with previous youth residential care placements (98%) [Reference Åsen, Schalinski, Lehmann, Lydersen, Von Oertzen and Greger70], and substantially lower rates in the healthy community (51.4%) [Reference Schalinski, Breinlinger, Hirt, Teicher, Odenwald and Rockstroh36]. Rates of mental disorders (61%) were higher compared to those of the general population (18%) and slightly higher compared to adults previously placed in out-of-home care (31–45% depending on the type of out-of-home care) [Reference Seker, Boonmann, Gerger, Jäggi, d’Huart and Schmeck71]. In our study, 32% fulfilled the criteria for an internalizing diagnosis and 54% for an externalizing diagnosis, with 26% fulfilling the criteria for both. This lines up with results indicating that externalizing disorders are more common than internalizing disorders in out-of-home-placed individuals, particularly when placed within juvenile justice institutions [Reference Seker, Boonmann, Gerger, Jäggi, d’Huart and Schmeck71]. Our findings underline the need for intensified prevention efforts, suitable psychosocial and therapeutic support, and targeted interventions in this vulnerable population.

Confirming prior evidence of a dose–response relationship between CM and psychopathology [Reference Green, McLaughlin, Berglund, Gruber, Sampson and Zaslavsky6, Reference Gardner, Thomas and Erskine8–Reference Zeanah and Humphreys11], global CM measures were strong predictors of mental health problems, with larger effects for internalizing than externalizing problems, likely due to its impact on emotional functioning (e.g., shame, guilt, insecurity) [Reference O’Loghlen, Galligan and Grant72, Reference Ion, Bîlc, Pițur, Pop, Szentágotai-Tătar and Miu73], self-concept (e.g., negative self-perception, impaired self-esteem) [Reference Hu, Yang, He, Wang, Xu and Zhu74, Reference Melamed, Botting, Lofthouse, Pass and Meiser-Stedman75], and the neurobiological systems involved in emotion and stress regulation (e.g., limbic system) [Reference Teicher and Samson76, Reference Ioannidis, Askelund, Kievit and Van Harmelen77]. Similar results have also been reported for other youth residential care cohorts [Reference Briggs, Greeson, Layne, Fairbank, Knoverek and Pynoos37, Reference Greger, Myhre, Lydersen and Jozefiak68, Reference Fischer, Dölitzsch, Schmeck, Fegert and Schmid78].

Beyond the global effects of CM, we found timing-specific effects, particularly regarding abuse in early childhood and peer victimization in late adolescence, with stronger effects on externalizing than internalizing symptoms. While considerable variability in reported sensitive periods for internalizing and externalizing disorders exists, there is little consistency in the periods identified for specific CM types or categories (e.g., abuse vs. neglect) [Reference Schaefer, Cheng and Dunn32]. For depression, cumulative effects of CM seem to be most pronounced in mid-childhood [Reference Li, Gao, Su, Zhang, Yang and D’Arcy79]. In contrast, we found that type- and timing-specific effects were strongest in early childhood and mid-to-late adolescence, which might be related to our sample of youths in residential care, who were oftentimes placed to prevent maltreatment within the family context. As such, residential care placements might exert their most protective effects against CM in mid-childhood and early adolescence but might be prone to further adverse peer victimization.

Of note, sexual abuse was rarely reported in our sample, critically limiting the conclusions that could be drawn on this CM type. Additionally, experiences of neglect were reported by approximately 90% of participants throughout ages 3 to 18 (see Supplementary Table S3), contributing substantially to the global CM measures and limiting the probability of detecting type- and timing-specific effects in this regard due to possible ceiling effects. As such, our sample is characterized by high prevalence rates of parental neglect throughout childhood and adolescence and overall low prevalence rates of sexual abuse. Such ceiling and floor effects may substantially impact the results of random forest regressions, as they artificially restrict the variance of a predictor, diminishing the informative value for meaningful sample splits, and thereby distorting the estimation of variable importance [Reference Strobl, Boulesteix, Zeileis and Hothorn62]. It must therefore be stressed that the finding of parental neglect not being a significant predictor of mental health outcomes is likely due to the given exposure distribution in our sample and the associated bias introduced to the random forest regression model.

While not all findings concerning the externalizing dimension could be replicated, this might be related to the poorer overlap between self-reported externalizing problems assessed by the ASBEA and the clinically assessed externalizing diagnoses grouped by the HiTOP model. While the latter subsumes substance use disorders (a common diagnosis in our sample), the ASEBA’s externalizing domain only contains a few items on substance misuse. In sum, while the literature on timing effects is steadily growing, sensitive periods for psychiatric disorders and related endpoints do not converge on single periods of vulnerability [Reference Schaefer, Cheng and Dunn32]. Our work adds novel type- and timing-specific insights in a highly vulnerable subgroup of young adults with previous youth residential care placements.

Our results showcase the complex CM trajectories of institutionalized individuals and the significance of diverse maltreatment experiences that may derive from parents in early childhood (i.e., before placement) as well as from peers in adolescence (i.e., during placement). Even though individuals in our sample were first placed in an institution at an average age of 11 and stayed in out-of-home placements for an average of 7 years thereafter, the majority of the sample was still in contact with their families of origin throughout their stays. This may explain the high chronicity of CM exposure despite the intervention of youth welfare institutions. This moreover, underlines the need to routinely and systematically screen for CM in youth residential care. Doing so might help to integrate safeguarding measures [Reference Li, Gao, Su, Zhang, Yang and D’Arcy79], timely enroll individuals in therapeutic and pedagogical interventions, and foster feelings of safety and connectedness [Reference Lorenc, Lester, Sutcliffe, Stansfield and Thomas80, Reference Zhang, Conner, Lim and Lefmann81]. Our results suggest that prevention and intervention strategies might specifically tackle the potential intergenerational transmission of externalizing behaviors, such as aggression and violence within families [Reference Widom, DuMont and Czaja16, Reference Widom82, Reference Dodge, Bates and Pettit83], and cycles of violence and revictimization in peer groups. This is a complex endeavor, as strategies may need to target various contexts and groups of perpetrators simultaneously. By doing so, one needs to critically appraise the potential victim–offender overlap [Reference Eriksson and Mazerolle84], but targeted, personalized approaches might prove helpful in breaking vicious interdependencies [Reference Averdijk, Ribeaud and Eisner85, Reference Reijntjes, Kamphuis, Prinzie, Boelen, Van Der Schoot and Telch86]. It should thereby not be overlooked that neglect was the predominant CM type experienced in this population and targeting neglect, starting in early childhood, may therefore have a strong and lasting reducing impact on overall CM severity.

Beyond screening and trauma-informed interventions within the care system, it is important to prevent CM from happening in the first place and to implement universal, indicated, and selective prevention programs. Universal programs such as Triple P (Positive Parenting Program) and family-centered approaches like (Early) Head Start have been shown to significantly reduce aversive parenting practices and the risk of abuse at the population level [Reference Li, Peng and Li87, Reference Green, Ayoub, Bartlett, Von Ende, Furrer and Chazan-Cohen88]. For example, Triple P consistently reduced harsh parenting and improved positive parenting [Reference Li, Peng and Li87], and Head Start reduced rates of abuse in young children [Reference Green, Ayoub, Bartlett, Von Ende, Furrer and Chazan-Cohen88]. In cases in which CM is already evident, Multisystemic Therapy for Child Abuse and Neglect (MST-CAN) serves as a robust intervention, with strong evidence showing that it reduces abusive and neglectful parenting, parental distress, and the need for out-of-home placements [Reference Swenson, Schaeffer, Henggeler, Faldowski and Mayhew89, Reference Buderer, Hefti, Fux, Pérez, Swenson and Fürstenau90]. Together, these evidence-based approaches demonstrate that for the effective reduction of CM, a combination of early universal prevention, indicated, selective prevention, as well as targeted intervention is urgently needed.

There are some limitations to our study. First, the retrospective self-report CM measure may be subject to biases [Reference Bürgin, Boonmann, Schmid, Tripp and O’Donovan91–Reference Danese and Widom93]. Indeed, there is poor overlap between prospective versus retrospective CM measures [Reference Baldwin, Reuben, Newbury and Danese94]. While subjective reports might weigh more heavily in relation to psychopathology [Reference Smith and Pollak19, Reference Danese and Widom93], the validity might be limited in our sample. Importantly, while the MACE-X separates different perpetrators (e.g., peers, siblings, parents), “parent” is introduced as “an adult with whom one lived at that time.” Thus, it was unclear whom the participants referred to specifically and forms of abuse experienced in the family of origin (e.g., on weekend leaves) and beyond (e.g., inside or outside the institution) might be mixed. Second, the sample was heterogeneous regarding individual residential care paths and received interventions. Due to restrictions in sample size, we refrained from addressing this heterogeneity (e.g., by analyzing subsamples). While the sample is fairly representative of out-of-home care institutions in Switzerland [Reference Jäggi, Schmid, Bürgin, Saladin, Grob and Boonmann95], and there appears to be no systematic bias between included and non-included participants (see Supplemental Tables S1 and S2), we cannot rule out the possibility that participants in non-included institutions, all those who did not participate at baseline, or participants who dropped out after baseline were more burdened than included participants, which might bias our results.

Despite these shortcomings, our study has several strengths: We investigated a hard-to-reach and vulnerable group of young adults with a history of youth residential care using a comprehensive measure of CM and transdiagnostic measures of mental health problems and psychopathology. We employed a state-of-the-art machine learning algorithm to handle a high number of multicollinear predictors indexing CM type and timing on general, internalizing, and externalizing problems. We validated our findings using clinical diagnoses and applied competitive hypothesis testing to contrast the importance of global CM measures to type- and timing-specific predictors [Reference Schaefer, Cheng and Dunn32].

Our findings may convey important implications for universal prevention, targeted interventions and decision-making in health care systems and child welfare policies. Complementing the prevention of transdiagnostic risk factors, such as CM, the promotion of accessible and suitable interventions for vulnerable populations in various care settings might counteract some maltreatment-associated risks [Reference Bunting, Montgomery, Mooney, MacDonald, Coulter and Hayes96], especially if type- and timing-specific CM reports are acknowledged. Preventing CM and reducing its overall severity, especially regarding early parental and domestic abuse and later peer revictimization, might be an important target to improve mental health among out-of-home placed children and adolescents.

Supplementary material

The supplementary material for this article can be found at http://doi.org/10.1192/j.eurpsy.2025.10127.

Data availability statement

Data are available upon request.

Author contribution

Maria Meier: Formal analysis, Visualization, Writing – Original Draft, Writing – Review & Editing. Inga Schalinski: Formal analysis, Visualization, Writing – Review & Editing. Cyril Boonmann, Nils Jenkel, Süheyla Seker, Delfine d’Huart, Jörg M. Fegert, Vera Clemens: Writing – Review & Editing. Marc Schmid: Resources, Writing – Review & Editing, Funding Acquisition. David Bürgin: Investigation, Data Curation, Writing – Original Draft, Writing – Review & Editing.

Financial support

MAZ. and JAEL were funded by the Swiss Ministry of Justice (PI: Marc Schmid). MM was employed via a DFG/SNSF Lead Agency Grant (DFG: CL 878/2–1; SNSF: 204033). DB was supported by the Research Fund for Excellent Junior Researcher of the University of Basel for the study “Childhood Trauma – Questioning, Understanding, Exploring, Seeing, Testing (CT-Quest).”

Competing interests

The authors declare none.

References

Copeland, WE, Shanahan, L, Hinesley, J, Chan, RF, Aberg, KA, Fairbank, JA, et al. Association of childhood trauma exposure with adult psychiatric disorders and functional outcomes. JAMA Netw Open. 2018;1:e184493. https://doi.org/10.1001/jamanetworkopen.2018.4493.Google Scholar
Bellis, MA, Hughes, K, Ford, K, Ramos Rodriguez, G, Sethi, D, Passmore, J. Life course health consequences and associated annual costs of adverse childhood experiences across Europe and North America: a systematic review and meta-analysis. Lancet Public Health. 2019;4:e517–28. https://doi.org/10.1016/S2468-2667(19)30145-8.Google Scholar
McCrory, EJ, Gerin, MI, Viding, E. Annual research review: childhood maltreatment, latent vulnerability and the shift to preventative psychiatry – the contribution of functional brain imaging. J Child Psychol Psychiatry. 2017;58:338–57. https://doi.org/10.1111/jcpp.12713.Google Scholar
McLaughlin, KA, Colich, NL, Rodman, AM, Weissman, DG. Mechanisms linking childhood trauma exposure and psychopathology: a transdiagnostic model of risk and resilience. BMC Med. 2020;18:96. https://doi.org/10.1186/s12916-020-01561-6.Google Scholar
Hughes, K, Ford, K, Bellis, MA, Glendinning, F, Harrison, E, Passmore, J. Health and financial costs of adverse childhood experiences in 28 European countries: a systematic review and meta-analysis. Lancet Pub Health. 2021;6:e848–57. https://doi.org/10.1016/S2468-2667(21)00232-2.Google Scholar
Green, JG, McLaughlin, KA, Berglund, PA, Gruber, MJ, Sampson, NA, Zaslavsky, AM, et al. Childhood adversities and adult psychiatric disorders in the national comorbidity survey replication I: associations with first onset of DSM-IV disorders. Arch Gen Psychiatry. 2010;67:113. https://doi.org/10.1001/archgenpsychiatry.2009.186.Google Scholar
Hughes, K, Bellis, MA, Hardcastle, KA, Sethi, D, Butchart, A, Mikton, C, et al. The effect of multiple adverse childhood experiences on health: a systematic review and meta-analysis. Lancet Publ Health. 2017;2:e356–66. https://doi.org/10.1016/S2468-2667(17)30118-4.Google Scholar
Gardner, MJ, Thomas, HJ, Erskine, HE. The association between five forms of child maltreatment and depressive and anxiety disorders: a systematic review and meta-analysis. Child Abuse Negl. 2019;96:104082. https://doi.org/10.1016/j.chiabu.2019.104082.Google Scholar
Kessler, RC, McLaughlin, KA, Green, JG, Gruber, MJ, Sampson, NA, Zaslavsky, AM, et al. Childhood adversities and adult psychopathology in the WHO World Mental Health surveys. Br J Psychiatry. 2010;197:378–85. https://doi.org/10.1192/bjp.bp.110.080499.Google Scholar
Lewis, SJ, Danese, A. Trauma-related psychopathology in children and adolescents: recent developments and future directions. In: Shaping the future of child adolescent mental health. Elsevier; 2023, p. 145–74. https://doi.org/10.1016/B978-0-323-91709-4.00011-1.Google Scholar
Zeanah, CH, Humphreys, KL. Child abuse and neglect. J Am Acad Child Adolesc Psychiatry. 2018;57:637–44. https://doi.org/10.1016/j.jaac.2018.06.007.Google Scholar
Humphreys, KL, Zeanah, CH. Deviations from the expectable environment in early childhood and emerging psychopathology. Neuropsychopharmacology. 2015;40:154–70. https://doi.org/10.1038/npp.2014.165.Google Scholar
Bürgin, D, Witt, A, Seker, S, d’Huart, D, Meier, M, Jenkel, N, et al. Childhood maltreatment and mental health problems in a 10-year follow-up study of adolescents in youth residential care: a latent transition analysis. Dev Psychopathol. 2023:1–16. https://doi.org/10.1017/S0954579423001426.Google Scholar
Heim, C, Binder, EB. Current research trends in early life stress and depression: review of human studies on sensitive periods, gene–environment interactions, and epigenetics. Exp Neurol. 2012;233:102–11. https://doi.org/10.1016/j.expneurol.2011.10.032.Google Scholar
Teicher, MH, Samson, JA. Childhood maltreatment and psychopathology: a case for ecophenotypic variants as clinically and neurobiologically distinct subtypes. Am J Psychiatry. 2013;170:1114–33. https://doi.org/10.1176/appi.ajp.2013.12070957.Google Scholar
Widom, CS, DuMont, K, Czaja, SJ. A prospective investigation of major depressive disorder and comorbidity in abused and neglected children grown up. Arch Gen Psychiatry 2007;64:49. https://doi.org/10.1001/archpsyc.64.1.49.Google Scholar
Koppold, A, Ruge, J, Hecker, T, Lonsdorf, TB. The many faces of early life adversity – content overlap in validated assessment instruments as well as in fear and reward learning research. PsyArXiv; 2023. https://doi.org/10.31234/osf.io/m9abu.Google Scholar
Lacey, RE, Minnis, H. Practitioner review: twenty years of research with adverse childhood experience scores – advantages, disadvantages and applications to practice. J Child Psychol Psychiatry. 2020;61:116–30. https://doi.org/10.1111/jcpp.13135.Google Scholar
Smith, KE, Pollak, SD. Rethinking concepts and categories for understanding the neurodevelopmental effects of childhood adversity. Perspect Psychol Sci. 2021;16:67–93. https://doi.org/10.1177/1745691620920725.Google Scholar
Anda, RF, Felitti, VJ, Bremner, JD, Walker, JD, Whitfield, C, Perry, BD, et al. The enduring effects of abuse and related adverse experiences in childhood: a convergence of evidence from neurobiology and epidemiology. Eur Arch Psychiatry Clin Neurosci 2006;256:174–86. https://doi.org/10.1007/s00406-005-0624-4.Google Scholar
Brown, DW, Anda, RF, Tiemeier, H, Felitti, VJ, Edwards, VJ, Croft, JB, et al. Adverse childhood experiences and the risk of premature mortality. Am J Prev Med. 2009;37:389–96. https://doi.org/10.1016/j.amepre.2009.06.021.Google Scholar
Felitti, VJ, Anda, RF, Nordenberg, D, Williamson, DF, Spitz, AM, Edwards, V, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. Am J Prev Med. 1998;14:245–58. https://doi.org/10.1016/S0749-3797(98)00017-8.Google Scholar
Dunn, E, Busso, DS, Davis, KA, Smith, ADAC, Mitchell, C, Tiemeier, H, et al. Sensitive periods for the effect of child maltreatment on psychopathology symptoms in adolescence. Comp Psychiatry. 2023;9:145–53. https://doi.org/10.1159/000530120.Google Scholar
Dunn, EC, Nishimi, K, Gomez, SH, Powers, A, Bradley, B. Developmental timing of trauma exposure and emotion dysregulation in adulthood: are there sensitive periods when trauma is most harmful? J Affect Disord. 2018;227:869–77. https://doi.org/10.1016/j.jad.2017.10.045.Google Scholar
Berman, IS, McLaughlin, KA, Tottenham, N, Godfrey, K, Seeman, T, Loucks, E, et al. Measuring early life adversity: a dimensional approach. Dev Psychopathol. 2022;34:499–511. https://doi.org/10.1017/S0954579421001826.Google Scholar
Clemens, V, Jarczok, M, Fegert, JM, Brähler, E, Jud, A. Timing and chronicity of child maltreatment in Germany: results from a representative sample. Publ Health. 2024;235:173–9. https://doi.org/10.1016/j.puhe.2024.06.019.Google Scholar
Jarczok, M, Lange, S, Meinck, F, Witt, A, Clemens, V, Fegert, JM, et al. Widening epidemiological data on the prevalence of child maltreatment: validation of the German ICAST-R in a student sample and national household survey. Child Abuse Negl. 2023;137:106038. https://doi.org/10.1016/j.chiabu.2023.106038.Google Scholar
Teicher, MH, Parigger, A. The ‘maltreatment and abuse chronology of exposure’ (MACE) scale for the retrospective assessment of abuse and neglect during development. PLOS One. 2015;10:e0117423. https://doi.org/10.1371/journal.pone.0117423.Google Scholar
Sturmbauer, SC, Shields, GS, Hetzel, E-L, Rohleder, N, Slavich, GM. The stress and adversity inventory for adults (adult STRAIN) in German: an overview and initial validation. PLOS One 2019;14:e0216419. https://doi.org/10.1371/journal.pone.0216419.Google Scholar
Hales, GK, Saribaz, ZE, Debowska, A, Rowe, R. Links of adversity in childhood with mental and physical health outcomes: a systematic review of longitudinal mediating and moderating mechanisms. Traum Viol Abuse 2023;24:1465–82. https://doi.org/10.1177/15248380221075087.Google Scholar
Russotti, J, Warmingham, JM, Duprey, EB, Handley, ED, Manly, JT, Rogosch, FA, et al. Child maltreatment and the development of psychopathology: the role of developmental timing and chronicity. Child Abuse Negl 2021;120:105215. https://doi.org/10.1016/j.chiabu.2021.105215.Google Scholar
Schaefer, JD, Cheng, TW, Dunn, EC. Sensitive periods in development and risk for psychiatric disorders and related endpoints: a systematic review of child maltreatment findings. Lancet Psychiatry 2022;9:978–91. https://doi.org/10.1016/S2215-0366(22)00362-5.Google Scholar
Schalinski, I, Teicher, MH, Nischk, D, Hinderer, E, Müller, O, Rockstroh, B. Type and timing of adverse childhood experiences differentially affect severity of PTSD, dissociative and depressive symptoms in adult inpatients. BMC Psychiatry 2016;16:295. https://doi.org/10.1186/s12888-016-1004-5.Google Scholar
Sicorello, M, Thome, J, Herzog, J, Schmahl, C. Differential effects of early adversity and posttraumatic stress disorder on amygdala reactivity: the role of developmental timing. Biol Psychiatry Cogn Neurosci Neuroimaging 2021;6:1044–51. https://doi.org/10.1016/j.bpsc.2020.10.009.Google Scholar
Herzog, JI, Thome, J, Demirakca, T, Koppe, G, Ende, G, Lis, S, et al. Influence of severity of type and timing of retrospectively reported childhood maltreatment on female amygdala and hippocampal volume. Sci Rep 2020;10:1903. https://doi.org/10.1038/s41598-020-57490-0.Google Scholar
Schalinski, I, Breinlinger, S, Hirt, V, Teicher, MH, Odenwald, M, Rockstroh, B. Environmental adversities and psychotic symptoms: the impact of timing of trauma, abuse, and neglect. Schizophr Res. 2019;205:4–9. https://doi.org/10.1016/j.schres.2017.10.034.Google Scholar
Briggs, EC, Greeson, JKP, Layne, CM, Fairbank, JA, Knoverek, AM, Pynoos, RS. Trauma exposure, psychosocial functioning, and treatment needs of youth in residential care: preliminary findings from the NCTSN core data set. J Child Adolesc Trauma. 2012;5:1–15. https://doi.org/10.1080/19361521.2012.646413.Google Scholar
Zelechoski, AD, Sharma, R, Beserra, K, Miguel, JL, DeMarco, M, Spinazzola, J. Traumatized youth in residential treatment settings: prevalence, clinical presentation, treatment, and policy implications. J Fam Violence. 2013;28:639–52. https://doi.org/10.1007/s10896-013-9534-9.Google Scholar
Schmid, M, Kölch, M, Fegert, JM, Schmeck, K. Abschlussbericht Modellversuch Abklärung und Zielerreichung (MAZ) in stationären Massnahmen; 2013.Google Scholar
Schmid, M, Fegert, JM, Clemens, V, Seker, S, d’Huart, D, Binder, M, et al. Misshandlungs- und Vernachlässigungserfahrungen in der Kindheit: Ein Risikofaktor für die soziale Teilhabe ehemals außerfamiliär platzierter junger Erwachsener: Ergebnisse der schweizweiten Kohortenstudie „Jugendhilfeverläufe: Aus Erfahrung Lernen (JAEL)“. Kindh Entwickl. 2022;31:22–39. https://doi.org/10.1026/0942-5403/a000366.Google Scholar
Juen, F, Hecker, T, Hermenau, K, Teicher, MH, Mikinga, G, Nkuba, M, et al. Child maltreatment in a high adversity context: associations of age, type and timing of exposure with psychopathology in middle childhood. Child Abuse Negl. 2024;157:107060. https://doi.org/10.1016/j.chiabu.2024.107060.Google Scholar
Schalinski, I, Teicher, MH. Type and timing of childhood maltreatment and severity of shutdown dissociation in patients with schizophrenia spectrum disorder. PLOS One. 2015;10:e0127151. https://doi.org/10.1371/journal.pone.0127151.Google Scholar
Schalinski, I, Jozefiak, T, Stenzel, N, Breinlinger, S, Teicher, MH, Rockstroh, B. Childhood maltreatment in patients with mental illness as a risk factor for obesity. Psychiatry Res. 2024;342:116264. https://doi.org/10.1016/j.psychres.2024.116264.Google Scholar
Gerke, J, Koenig, AM, Conrad, D, Doyen-Waldecker, C, Pauly, M, Gündel, H, et al. Childhood maltreatment as risk factor for lifetime depression: the role of different types of experiences and sensitive periods. Ment Health Prev. 2018;10:56–65. https://doi.org/10.1016/j.mhp.2018.03.002.Google Scholar
Grauduszus, Y, Sicorello, M, Demirakca, T, Von Schröder, C, Schmahl, C, Ende, G. New insights into the effects of type and timing of childhood maltreatment on brain morphometry. Sci Rep. 2024;14:11394. https://doi.org/10.1038/s41598-024-62051-w.Google Scholar
Hutson, LM, Dan, R, Brown, AR, Esfand, SM, Ruberto, V, Johns, E, et al. Childhood sexual abuse and lifetime depressive symptoms: the importance of type and timing of childhood emotional maltreatment. Psychol Med. 2024;54:4385–95. https://doi.org/10.1017/S003329172400268X.Google Scholar
Schalinski, I, Teicher, MH, Carolus, AM, Rockstroh, B. Defining the impact of childhood adversities on cognitive deficits in psychosis: an exploratory analysis. Schizophr Res. 2018;192:351–6. https://doi.org/10.1016/j.schres.2017.05.014.Google Scholar
d’Huart, D, Hutsebaut, J, Seker, S, Schmid, M, Schmeck, K, Bürgin, D, et al. Personality functioning and the pathogenic effect of childhood maltreatment in a high-risk sample. Child Adolesc Psychiatry Ment Health. 2022;16:95. https://doi.org/10.1186/s13034-022-00527-1.Google Scholar
Bürgin, D, Clemens, V, Varghese, N, Eckert, A, Huber, M, Bruttin, E, et al. Adverse and traumatic exposures, posttraumatic stress disorder, telomere length, and hair cortisol – exploring associations in a high-risk sample of young adult residential care leavers. Brain Behav Immun Health. 2022;26:100524. https://doi.org/10.1016/j.bbih.2022.100524.Google Scholar
Bürgin, D, Varghese, N, Eckert, A, Clemens, V, Unternährer, E, Boonmann, C, et al. Higher hair cortisol concentrations associated with shorter leukocyte telomere length in high-risk young adults. Sci Rep. 2022;12:11730. https://doi.org/10.1038/s41598-022-14905-4.Google Scholar
d’Huart, D, Steppan, M, Seker, S, Bürgin, D, Boonmann, C, Birkhölzer, M, et al. Prevalence and 10-year stability of personality disorders from adolescence to young adulthood in a high-risk sample. Front Psychiatry. 2022;13:840678. https://doi.org/10.3389/fpsyt.2022.840678.Google Scholar
Seker, S, Boonmann, C, d’Huart, D, Bürgin, D, Schmeck, K, Jenkel, N, et al. Mental disorders into adulthood among adolescents placed in residential care: a prospective 10-year follow-up study. Eur Psychiatry. 2022;65:e40. https://doi.org/10.1192/j.eurpsy.2022.30.Google Scholar
Meier, M, Kantelhardt, S, Gurri, L, Stadler, C, Schmid, M, Clemens, V, et al. Childhood trauma is linked to epigenetic age deceleration in young adults with previous youth residential care placements. Eur J Psychotraumatology. 2024;15:2379144. https://doi.org/10.1080/20008066.2024.2379144.Google Scholar
Thekkumthala, D, Schalinski, I, Parigger, A, Ruf-Leuschner, M, Elbert, T, Schauer, M. KERF-I. Belastende Kindheitserfahrungen. 2023. https://doi.org/10.23668/PSYCHARCHIVES.12495.Google Scholar
Isele, D, Teicher, MH, Ruf-Leuschner, M, Elbert, T, Kolassa, I-T, Schury, K, et al. KERF–Ein Instrument zur umfassenden Ermittlung belastender Kindheitserfahrungen: Erstellung und psychometrische Beurteilung der deutschsprachigen MACE (Maltreatment and Abuse Chronology of Exposure) Scale. Z Für Klin Psychol Psychother. 2014;43:121–30. https://doi.org/10.1026/1616-3443/a000257.Google Scholar
Agrez, U, Winkler Metzke, C, Steinhausen, H-C. Psychometrische Charakteristika der deutschen Version des Young Adult Self-Report (YASR). Z Für Klin Psychol Psychother. 2011;40:75–84. https://doi.org/10.1026/1616-3443/a000073.Google Scholar
Achenbach, TM, Rescorla, L, McConaughey, S, Pecora, P, Wetherbee, K, Ruffle, T. Achenbach system of empirically based assessment. In: Handbook of psychological and education assessment of children: personality, behavior, and context. New York, NY: The Guilford Press, Vol. 2. 2003, p. 406–32. https://psycnet.apa.org/record/2004-00018-000.Google Scholar
Wickham, H, Miller, E, Smith, D. Haven: import and export “SPSS”, “Stata” and “SAS” files. 2023.Google Scholar
Behrendt, S. lm.beta: add standardized regression coefficients to linear-model-objects. 2023.Google Scholar
Hothorn, T, Hornik, K, Zeileis, A. Unbiased recursive partitioning: a conditional inference framework. J Comput Graph Stat 2006;15:651–74. https://doi.org/10.1198/106186006X133933.Google Scholar
Strobl, C, Boulesteix, A-L, Kneib, T, Augustin, T, Zeileis, A. Conditional variable importance for random forests. BMC Bioinformat 2008;9:307. https://doi.org/10.1186/1471-2105-9-307.Google Scholar
Strobl, C, Boulesteix, A-L, Zeileis, A, Hothorn, T. Bias in random forest variable importance measures: illustrations, sources and a solution. BMC Bioinformat 2007;8:25. https://doi.org/10.1186/1471-2105-8-25.Google Scholar
Kuhn, M. Building predictive models in R using the caret package. J Stat Softw 2008;28. https://doi.org/10.18637/jss.v028.i05.Google Scholar
Wickham, H. ggplot2: elegant graphics for data analysis. New York: Springer; 2016.Google Scholar
Wei, T, Simko, V. R package “corrplot”: visualization of a correlation matrix. 2021.Google Scholar
Pedersen, TL. patchwork: the composer of plots. 2019.Google Scholar
Rebbe, R, Nurius, PS, Ahrens, KR, Courtney, ME. Adverse childhood experiences among youth aging out of foster care: a latent class analysis. Child Youth Serv Rev. 2017;74:108–16. https://doi.org/10.1016/j.childyouth.2017.02.004.Google Scholar
Greger, HK, Myhre, AK, Lydersen, S, Jozefiak, T. Child maltreatment and quality of life: a study of adolescents in residential care. Health Qual Life Outcomes. 2016;14:74. https://doi.org/10.1186/s12955-016-0479-6.Google Scholar
Fosse, R, Skjelstad, DV, Schalinski, I, Thekkumthala, D, Elbert, T, Aanondsen, CM, et al. Measuring childhood maltreatment: psychometric properties of the norwegian version of the maltreatment and abuse chronology of exposure (MACE) scale. PLOS One. 2020;15:e0229661. https://doi.org/10.1371/journal.pone.0229661.Google Scholar
Åsen, ME, Schalinski, I, Lehmann, S, Lydersen, S, Von Oertzen, T, Greger, HK. Child maltreatment in young adults with residential youth care background: prevalence and post-placement trends. Child Abuse Negl. 2024;157:107074. https://doi.org/10.1016/j.chiabu.2024.107074.Google Scholar
Seker, S, Boonmann, C, Gerger, H, Jäggi, L, d’Huart, D, Schmeck, K, et al. Mental disorders among adults formerly in out-of-home care: a systematic review and meta-analysis of longitudinal studies. Eur Child Adolesc Psychiatry. 2022;31:1963–82. https://doi.org/10.1007/s00787-021-01828-0.Google Scholar
O’Loghlen, E, Galligan, R, Grant, S. Childhood maltreatment, shame, psychological distress, and binge eating: testing a serial mediational model. J Eat Disord. 2023;11:96. https://doi.org/10.1186/s40337-023-00819-7.Google Scholar
Ion, A, Bîlc, MI, Pițur, S, Pop, CF, Szentágotai-Tătar, A, Miu, AC. Childhood maltreatment and emotion regulation in everyday life: an experience sampling study. Sci Rep. 2023;13:7214. https://doi.org/10.1038/s41598-023-34302-9.Google Scholar
Hu, Y, Yang, Y, He, Z, Wang, D, Xu, F, Zhu, X, et al. Self-concept mediates the relationships between childhood traumatic experiences and adolescent depression in both clinical and community samples. BMC Psychiatry. 2024;24:224. https://doi.org/10.1186/s12888-024-05671-w.Google Scholar
Melamed, DM, Botting, J, Lofthouse, K, Pass, L, Meiser-Stedman, R. The relationship between negative self-concept, trauma, and maltreatment in children and adolescents: a meta-analysis. Clin Child Fam Psychol Rev. 2024;27:220–34. https://doi.org/10.1007/s10567-024-00472-9.Google Scholar
Teicher, MH, Samson, JA. Annual research review: enduring neurobiological effects of childhood abuse and neglect. J Child Psychol Psychiatry. 2016;57:241–66. https://doi.org/10.1111/jcpp.12507.Google Scholar
Ioannidis, K, Askelund, AD, Kievit, RA, Van Harmelen, A-L. The complex neurobiology of resilient functioning after childhood maltreatment. BMC Med. 2020;18:32. https://doi.org/10.1186/s12916-020-1490-7.Google Scholar
Fischer, S, Dölitzsch, C, Schmeck, K, Fegert, JM, Schmid, M. Interpersonal trauma and associated psychopathology in girls and boys living in residential care. Child Youth Serv Rev. 2016;67:203–11. https://doi.org/10.1016/j.childyouth.2016.06.013.Google Scholar
Li, M, Gao, T, Su, Y, Zhang, Y, Yang, G, D’Arcy, C, et al. The timing effect of childhood maltreatment in depression: a systematic review and meta-analysis. Trauma Violence Abuse. 2023;24:2560–80. https://doi.org/10.1177/15248380221102558.Google Scholar
Lorenc, T, Lester, S, Sutcliffe, K, Stansfield, C, Thomas, J. Interventions to support people exposed to adverse childhood experiences: systematic review of systematic reviews. BMC Public Health. 2020;20:657. https://doi.org/10.1186/s12889-020-08789-0.Google Scholar
Zhang, S, Conner, A, Lim, Y, Lefmann, T. Trauma-informed care for children involved with the child welfare system: a meta-analysis. Child Abuse Negl. 2021;122:105296. https://doi.org/10.1016/j.chiabu.2021.105296.Google Scholar
Widom, CS. Does violence beget violence? A critical examination of the literature. Psychol Bull. 1989;106:3–28. https://doi.org/10.1037/0033-2909.106.1.3.Google Scholar
Dodge, KA, Bates, JE, Pettit, GS. Mechanisms in the cycle of violence. Science. 1990;250:1678–83. https://doi.org/10.1126/science.2270481.Google Scholar
Eriksson, L, Mazerolle, P. A cycle of violence? Examining family-of-origin violence, attitudes, and intimate partner violence perpetration. J Interpers Violence. 2015;30:945–64. https://doi.org/10.1177/0886260514539759.Google Scholar
Averdijk, M, Ribeaud, D, Eisner, M. Childhood predictors of violent victimization at age 17 years: the role of early social behavioral tendencies. J Pediatr. 2019;208:183–190.e1. https://doi.org/10.1016/j.jpeds.2018.12.056.Google Scholar
Reijntjes, A, Kamphuis, JH, Prinzie, P, Boelen, PA, Van Der Schoot, M, Telch, MJ. Prospective linkages between peer victimization and externalizing problems in children: a meta-analysis. Aggress Behav. 2011;37:215–22. https://doi.org/10.1002/ab.20374.Google Scholar
Li, N, Peng, J, Li, Y. Effects and moderators of triple P on the social, emotional, and behavioral problems of children: systematic review and meta-analysis. Front Psychol. 2021;12:709851. https://doi.org/10.3389/fpsyg.2021.709851.Google Scholar
Green, BL, Ayoub, C, Bartlett, JD, Von Ende, A, Furrer, C, Chazan-Cohen, R, et al. The effect of early head start on child welfare system involvement: a first look at longitudinal child maltreatment outcomes. Child Youth Serv Rev. 2014;42:127–35. https://doi.org/10.1016/j.childyouth.2014.03.044.Google Scholar
Swenson, CC, Schaeffer, CM, Henggeler, SW, Faldowski, R, Mayhew, AM. Multisystemic therapy for child abuse and neglect: a randomized effectiveness trial. J Fam Psychol. 2010;24:497–507. https://doi.org/10.1037/a0020324.Google Scholar
Buderer, C, Hefti, S, Fux, E, Pérez, T, Swenson, CC, Fürstenau, U, et al. Effects of multisystemic therapy for child abuse and neglect on severity of neglect, behavioral and emotional problems, and attachment disorder symptoms in children. Child Youth Serv Rev. 2020;119:105626. https://doi.org/10.1016/j.childyouth.2020.105626.Google Scholar
Bürgin, D, Boonmann, C, Schmid, M, Tripp, P, O’Donovan, A. Fact or artefact? Childhood adversity and adulthood trauma in the U.S. population-based health and retirement study. Eur J Psychotraumatology. 2020;11:1721146. https://doi.org/10.1080/20008198.2020.1721146.Google Scholar
Hardt, J, Rutter, M. Validity of adult retrospective reports of adverse childhood experiences: review of the evidence. J Child Psychol Psychiatry. 2004;45:260–73. https://doi.org/10.1111/j.1469-7610.2004.00218.x.Google Scholar
Danese, A, Widom, CS. Objective and subjective experiences of child maltreatment and their relationships with psychopathology. Nat Hum Behav. 2020;4:811–8. https://doi.org/10.1038/s41562-020-0880-3.Google Scholar
Baldwin, JR, Reuben, A, Newbury, JB, Danese, A. Agreement between prospective and retrospective measures of childhood maltreatment: a systematic review and meta-analysis. JAMA Psychiatry. 2019;76:584. https://doi.org/10.1001/jamapsychiatry.2019.0097.Google Scholar
Jäggi, L, Schmid, M, Bürgin, D, Saladin, N, Grob, A, Boonmann, C. Shared residential placement for child welfare and juvenile justice youth: current treatment needs and risk of adult criminal conviction. Child Adolesc Psychiatry Ment Health. 2021;15:2. https://doi.org/10.1186/s13034-020-00355-1.Google Scholar
Bunting, L, Montgomery, L, Mooney, S, MacDonald, M, Coulter, S, Hayes, D, et al. Trauma informed child welfare systems—a rapid evidence review. Int J Environ Res Public Health. 2019;16:2365. https://doi.org/10.3390/ijerph16132365.Google Scholar
Figure 0

Table 1. Description of the sample

Figure 1

Figure 1. Weighted severity of exposure scores (theoretical range 0–10) per life year and category of childhood maltreatment, i.e., parental abuse (ABUSE) and neglect (NEGLECT), peer violence (PEER), and sexual abuse (SEXA). Error bars represent SE.

Figure 2

Figure 2. Variable importance of the timing-specific effects of different childhood maltreatment types (solid lines) and the global childhood maltreatment severity, multiplicity, and duration measures (dashed lines) for overall mental health problems (A), internalizing mental health problems (B), and externalizing mental health problems (C). High variable importance implies that the variable is a strong predictor of the respective subscale of self-reported mental health problems. MACE SEVERITY = severity of exposure. MACE MULTI = number of individual scales above cut-off. MACE Duration = duration of above-threshold exposure.

Supplementary material: File

Meier et al. supplementary material

Meier et al. supplementary material
Download Meier et al. supplementary material(File)
File 158.7 KB
Submit a response

Comments

No Comments have been published for this article.