Leveraging machine learning for effective child maltreatment prevention: A case study of home visiting service assessments.

BACKGROUND: Early identification of children and families who may benefit from support is crucial for implementing strategies that can prevent the onset of child maltreatment. Predictive risk modeling (PRM) may offer valuable and efficient enhancements to existing risk assessment techniques. OBJECTIVE: To evaluate the PRM's effectiveness against the existing assessment tool in identifying children and families needing home visiting services. PARTICIPANTS AND SETTING: Children born in hospitals affiliated with the Bridges Maternal Child Health Network in Orange County, California, from 2011 to 2016 (N = 132,216). METHODS: We developed a PRM tool by integrating a machine learning algorithm with a linked dataset of birth records and child protection system (CPS) records. To align with the existing assessment tool (baseline model), we limited the predicting features to the information used by the existing tool. The need for home visiting services was measured by substantiated maltreatment allegation reported during the first three years of the child's life. RESULTS: Of the children born in Bridges Network hospitals between 2011 and 2016, 2.7 % experienced substantiated maltreatment allegations by the age of three. Within the top 30 % of children with high-risk scores, the PRM tool outperformed the baseline model, accurately identifying 75.3 %-84.1 % of all children who would experience maltreatment substantiation, surpassing the baseline model's performance of 46.2 %. CONCLUSIONS: Our study underscores the potential of PRM in enhancing the risk assessment tool used by a prevention program in a child welfare center in California. The findings provide valuable insights to practitioners interested in utilizing data for PRM development, highlighting the potential of machine learning algorithms to generate accurate predictions and inform targeted preventive services.

Early Childhood Income Instability and Mental Health in Adolescence: Parenting Stress and Child Maltreatment as Mediators.

Ample research has examined how point-in-time or static measures of economic deprivation are associated with children's mental health outcomes. Less is known about the relationship between early childhood unstable income and mental health outcomes. Using the Future of Families and Child Wellbeing Study, this study examined (1) the latent patterns of early childhood economic well-being, predicted by income level and instability (i.e., direction and frequency of income change); (2) the association of income deprivation patterns with subsequent anxiety and depression symptoms, paying particular attention to the mediating roles of parenting stress and child maltreatment risk. The latent class analysis results suggested four distinct groups representing different combinations of income level and instability. Structural equation modeling results indicated indirect links between income deprivation patterns and mental health outcomes, through parenting stress and physical and psychological abuse. Findings indicated the importance of policies and programs promoting economic stability over the long run.

Risk and Protective Factors for Parental Involvement and Early Indicators of School Achievement in Alaska.

OBJECTIVES: Parental involvement can affect child school readiness, which in turn influences subsequent child learning outcomes. While social support, stress, caregiver psychological distress, and drinking could affect parental involvement, it is unknown whether and how these factors influence downstream child learning outcomes through parental involvement and child school readiness. This study tests those associations. METHODS: Using de-identified data provided by the Alaska Longitudinal Child Abuse and Neglect Linkage project (N = 683), we use Structural Equation Modeling to assess direct and indirect effects of paths embedded in the proposed model. RESULTS: This study found statistically significant indirect effects: (1) path linking stress faced by caregivers to child reading proficiency through caregiver psychological distress, parental involvement, and child school readiness, (2) path linking stress faced by caregivers to child reading proficiency through caregiver drinking, parental involvement, and child school readiness, and (3) path linking social support for caregivers to child reading proficiency through caregiver psychological distress, parental involvement, and child school readiness. Post-estimation showed that the sum of the magnitude of total effects of stress and the magnitude of total effects of support is significantly larger than either alone. CONCLUSIONS FOR PRACTICE: Findings suggest that reducing caregiver stress and offering social support could not only benefit caregivers but learning outcomes of their children as well. For child learning outcomes, simultaneously reducing stress and offering social support for caregivers, rather than just one of them alone, is suggested. These results are important for children, particularly for those raised by caregivers experiencing psychological distress or drinking issues.

Does a child maltreatment report source predict differences in immediate and subsequent report outcomes?

BACKGROUND: Mandated reporting policies, a core response to the identification of child maltreatment, are widely debated. Currently, there are calls to abolish or scale back these policies to include only certain professionals. These calls warrant evaluation of whether there are any differences in child welfare outcomes based on report source. OBJECTIVE: To determine if the initial report source predicts immediate and long-term risk of re-referral, substantiation, and placement. PARTICIPANTS AND SETTING: We used yearly National Child Abuse and Neglect Data System (NCANDS) hotline report and placement data. Children (0-14y) with a first ever hotline report in 2012-2014 were followed for three years. The final sample included 2,101,397 children from 32 states. METHODS: We use descriptive and bivariate statistics to show initial report outcomes by reporter source type and logistic regression models to evaluate the effect of report source on immediate and subsequent report outcomes. RESULTS: Professional sources varied in levels of substantiation and placement, with law enforcement, medical, and social service sources showing much higher rates. Reports from professional sources have higher odds of initial report substantiation and foster care entry, and slightly lower odds of later re-report than nonprofessional sources. We found no differences between professional and nonprofessional sources in subsequent foster care entry. CONCLUSIONS: Reports from professional, nonprofessional, and unclassified sources have varying levels of risk in some of their short- and long-term outcomes. To the degree that child protective services embrace a long-term preventative role, reports by nonprofessional report sources may provide opportunities for prevention.

Child Protective Services Reports After Reunification: An Examination of the Risk of Being Rereported to Child Protective Services After Returning Home From Foster Care.

While a number of studies have examined the risk of reentry after exiting foster care, few studies have examined the risk of rereport to Child Protective Services (CPS) in the United States. Understanding more about rereports may help identify targets for supportive services that prevent reentry and promote safety. This study is the first to use nation-wide linked data to examine the risk of rereport for reunified children. The sample included children reunified after experiencing their first episode in foster care with at least one CPS report prior to care. With a 2000-day observation period, flexible parametric survival models with time varying hazard ratios were used to model the hazard or rereport conditional on prior CPS and foster care characteristics. Over 50% of the sample experienced a rereport after reunification. Results indicate that children entering foster care following a history of multiple CPS reports prior to placement were at substantially increased risk of rereport after returning home. A group of children with shorter stays in care had a high risk of recurrence within the first month, but this was not true over time. Implications for future research as well as permanency planning and addressing the needs of families with chronic reports are discussed.

Reducing Abuse and Neglect Recurrence Among Young Foster Children Reunified With Their Families.

BACKGROUND AND OBJECTIVES: Child maltreatment (CM) is a recurrent adverse life event known to cause enduring psychiatric impairment throughout life. For young children in protective custody for a first episode of CM, specialized court-coordinated intervention to optimize reunification has shown promise for preventing CM recidivism, with case series documenting short-term successes. METHODS: We tracked 10-year (Nov 2011-March 2022) court re-entry outcomes in a cohort of 272 young children, birth to six years, reunited with their families following placement in protective custody and court referral to the SYNCHRONY Project, a voluntary clinical service providing Incredible YearsTM parenting education, parental psychiatric care, and serial dyadic clinical evaluation to inform medical recommendations on safety for visitation and reunification. Re-entry was operationalized as rereferral to any Missouri Court and proportions compared with contemporaneous State and national data. RESULTS: SYNCHRONY-enrolled/reunified children experienced frequencies of guardianship (22%) and reunification (46%), in keeping with Missouri averages. In these categories, 3.4% and 7.1% respectively were re-referred to the Court over the 10-year follow-up. In care as usual nationally for this age group, the re-referral proportions are 18% (OR 7.5, P < .0001) and 35% (OR 6.1, P < .0001) respectively. In care as usual in Missouri across all ages, the re-referral proportion is 16% (odds ratio [OR] 3.09, P < .0001). CONCLUSIONS: Judicious implementation of evidence-based parenting education, 2-generation psychiatric care, and clinical consultation were associated with marked reduction in court re-entry versus care-as-usual and warrant consideration in intervention standards for young children in foster care.

Risk factors and neglect subtypes: Findings from a nationally representative data set.

Child neglect is a multidimensional concept encompassing various forms. Prior studies suggest that risk factors differ by neglect subtypes such as physical or supervisory neglect, but few studies address how risk factors vary between other neglect subtypes. This study aimed to examine how risk factors were related to neglect subtypes such as physical neglect, lack of supervision, exposure to domestic violence, substance-abusing parent, and mixed neglect. This study used secondary data from a nationally representative sample of children (National Survey of Child and Adolescent Well-Being-II, N = 5,872), and 786 children with a first-time child protective services investigation for neglect allegations alone were selected. Multinomial logistic regression analysis was used to explore how individual, family, and community risk factors may be associated with specific neglect subtypes. Five risk factors were able to discriminate between subtypes of neglect. For example, being a young child is associated with a greater risk of experiencing multiple forms of neglect. Caregiver's mental health problem is associated with a higher likelihood of being referred for multiple forms of neglect, particularly as compared with the risk of being referred for lack of supervision. Having poor social support is associated with a higher risk of physical neglect, and caregiver high stress is related to a higher risk of domestic violence. While most intervention programs target risk factors for overall child neglect, not specific neglect subtypes cases, our findings suggest that the intervention approach based on an understanding of the heterogeneity in risk factors between neglect subtypes is advisable. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

The Application of Family Stress Model to Investigating Adolescent Problematic Behaviors: The Moderating Role of Assets.

The Family Stress Model framework proposes that household income can influence child and youth development through caregiver psychological distress. While prior studies have observed stronger associations among households with lower income, the role of assets has been ignored. This is unfortunate, as many existing policies and practices that intend to improve child and family well-being are focused on assets. The purpose of this study is to clarify whether asset poverty moderates the direct and indirect effects of paths linking household income, caregiver psychological distress, and adolescent problematic behaviors. Using the 2017 and 2019 Panel Study of Income Dynamic Main Study and 2019 and 2020 Child Development Supplements, we find that the family stress processes consisting of household income, caregiver psychological distress, and adolescent problematic behaviors are less intensive for families with more assets. These findings not only add our knowledge of FSM by taking account the moderating role of assets but also advance our understanding that assets can benefit child and family well-being through alleviating family stress processes.

Racial/Ethnic Differences in Child Protective Services Reporting, Substantiation and Placement, With Comparison to Non-CPS Risks and Outcomes: 2005-2019.

We used National Child Abuse and Neglect Data System and Census data to examine Black-White and Hispanic-White disparities in reporting, substantiation, and out-of-home placement both descriptively from 2005-2019 and in multivariate models from 2007-2017. We also tracked contemporaneous social risk (e.g., child poverty) and child harm (e.g., infant mortality) disparities using non-child protective services (CPS) sources and compared them to CPS reporting rate disparities. Black-White CPS reporting disparities were lower than found in non-CPS risk and harm benchmarks. Consistent with the Hispanic paradox, Hispanic-White CPS reporting disparities were lower than risk disparities but similar to harm disparities. Descriptive and multivariate analyses of data from the past several years indicated that Black children were less likely to be substantiated or placed into out-of-home care following a report than White children. Hispanic children were slightly more likely to be substantiated or placed in out-of-home care than White children overall, but this difference disappeared in multivariate models. Available data provide no evidence that Black children were overreported relative to observed risks and harms reflected in non-CPS data. Reducing reporting rates among Black children will require addressing broader conditions associated with maltreatment.

Testing a QuantCrit-Informed Approach to the Empirical Study of Race/Ethnicity and Child Maltreatment.

The past several years have seen calls from QuantCrit scholars to "disaggregate" samples into same-race groups. To date, however, there has been no attempt to empirically evaluate the benefits of disaggregation within a child welfare sample. Using a child maltreatment dataset derived from the National Child Abuse and Neglect Data System and Census data, we empirically evaluate the utility of employing sample disaggregation (in which separate records are created for White, Black and Latino populations in each county) as well as variable creation disaggregation (in which we avoid using "full county" economic measures, but instead employ "same race/ethnicity" measures). Using model fit and convergence with findings from individual-level studies as evaluation metrics, we find that both kinds of disaggregation are demonstrably beneficial. We recommend that sample and variable disaggregation be considered by any future researchers using national geographically structured child maltreatment data.

Understanding the interconnected relationships between intimate partner violence and child maltreatment: The role of maternal material hardship.

BACKGROUND: Evidence indicates an association between intimate partner violence (IPV) and child maltreatment (CM). Although research shows mothers' IPV victimization is a risk factor for CM, the specific elements within IPV and CM are not well understood. Some studies suggest material hardship plays a role in the relationships between IPV and CM. However, the relationship between different typologies of IPV and CM considering material hardship has rarely been examined. OBJECTIVE: This study fills this gap by assessing the direct effects of IPV victimization on material hardship and CM, direct effects of material hardship on CM, and indirect effects of IPV victimization on CM via material hardship with attention to typologies of IPV and CM. METHODS: Data included responses by 3086 mothers in the Fragile Families Child Wellbeing Study. RESULTS: Structural equation modeling showed that material hardship predicted higher risk of CM regardless of the presence of IPV or what type (child physical assault: ß = 0.11, p < 0.001; child psychological aggression: ß = 0.12, p < 0.001; child neglect: ß = 0.06, p < 0.01). However, mothers who experienced controlling and emotional IPV had a higher risk of material hardship, and this increased their children's risk of CM (child physical assault: ß = 0.02, p < 0.01; child psychological aggression: ß = 0.02, p < 0.01; child neglect: ß = 0.01, p < 0.05). This was the only type of IPV associated with CM through its association with material hardship. CONCLUSIONS: The varied findings suggest that intervention efforts in families with co-occurring IPV and CM should mobilize economic support to IPV survivors with consideration of the varied effects of different types of IPV.

Documenting Opportunity for Systematic Identification and Mitigation of Risk for Child Maltreatment.

This is a communication of preliminary data as a matter of priority in relation to Clinical Trials protocol ID 2018110118; NCT04438161. This protocol represents, to our knowledge, a first-ever attempt to convert an epidemiologic discovery on risk for child maltreatment (CM) into a readily deployable modification of obstetrical practice designed to offset risk for CM and its psychiatric sequelae. Before1 and during the coronavirus disease 2019 (COVID-19 pandemic),2,3 CM has incurred a burden of epidemic proportions to U.S. children, with confirmed incidents occurring on the order of 12% of the population. Wu et al.4 and Putnam-Hornstein and Needell5 previously established that profiles of risk ascertained exclusively from birth records identified specific groups of newborns at highly elevated risk for official-report CM. For example, infants with the joint characteristics of low birth weight, more than 2 siblings, and maternal characteristics of being unmarried, on Medicaid, and smoking during pregnancy (ascertained separately) were found to have a 7-fold risk for maltreatment compared with the population average.4 Putnam-Hornstein and Needell showed that newborns with 3 or more risk factors ascertained from birth records (including any of the above, delayed prenatal care, less than high school maternal education, and maternal age less than 24 years) comprised 15% of an epidemiologic birth cohort but accounted for more than half of all the children in the cohort who experienced substantiated official-report maltreatment by the age of 5 years. This study explored whether prospective implementation of birth records screening in an urban obstetrical service recapitulated the association with CM observed in an epidemiologic context and whether families in higher echelons of risk (ascertained in this manner through birth records) could be prospectively engaged in supportive interventions of demonstrated effect in reducing the occurrence of CM. This work follows on promising efforts elsewhere to use birth records information to prioritize support services for young families,6 though such innovations have yet to be systematically incorporated into obstetrical or newborn medical services of U.S. health systems.

Parameterizing Toxic Stress in Early Childhood: Maternal Depression, Maltreatment, and HPA-Axis Variation in a Pilot Intervention Study.

Adverse experiences superseding a child's capacity to sustain regulation of emotion and adaptive function are theorized to constitute "toxic stressors" when they induce a deleterious biological response within an individual. We ascertained presumptive parameters of toxic stress among 164 low-income infants and toddlers (ages 4-48 months) from 132 families enrolled in Early Head Start (EHS). We randomized a subset of these families into a pilot intervention arm of parenting education (the Incredible Years, TIY), which supplemented the EHS curriculum. Official report child abuse and neglect (CAN) and child behavior were serially ascertained over the course of the study. We observed relatively low associations among maternal depression, CAN, caregiver-child relationship quality, hair cortisol, and adverse child behavioral outcomes. Moreover, despite poverty and the high prevalence (51%) of CAN in this sample, the frequency of clinical-level internalizing and externalizing behavior among the children did not exceed that of the general population, by their parents' report. The pilot supplementation of EHS with TIY improved attendance in group meetings but did not significantly reduce adverse behavioral outcomes or CAN. This study revealed marked independence of standard indices of toxic stress (child maltreatment, maternal depression, caregiver emotional unavailability) which have been presumed to be risk factors for the development of psychopathology. That they were weakly inter-correlated, and only modestly predictive of child behavioral outcomes in this EHS sample, caution against presumptions about the toxicity of individual stressors, highlight the importance of ascertaining risk (and compensatory influences) comprehensively, suggest buffering effects of programs like EHS, and demonstrate the need for greater understanding of what parameterizes resilience in early childhood.

"We're in this together": Perceived effects of breast cancer on African American survivors' marital relationships.

Objective: We examined married African American breast cancer survivors' perceptions of how cancer affected their marriage, social support from their spouses, and spouses' physical and mental health. Method: We conducted a thematic analysis of semi-structured interviews with 15 married African American breast cancer survivors who had participated in a larger randomized controlled trial. Interviews were professionally transcribed and then independently coded by two coders. Results: Themes emerged related to the challenges of maintaining mutually supportive relationships. There was variability in the perceived effects of cancer on relationships, as well as uncertainty about cancer's effects on their husbands' emotional and physical health and the adequacy of emotional and tangible support from their husbands. Participants described husbands' key role in promoting wives' positive body image, as well as the challenges of negotiating sexual activity. Participants were receptive to help from medical professionals in dealing with relationship issues. Conclusions: Findings show variability in couples' responses to cancer, with some patients and couples adapting well and others needing additional support. During treatment and at follow-up, oncology social workers can assess patients' and family members' social support needs, provide mental health services, and provide patient navigation to help patients and caregivers access health care and community resources.

Neglect subtypes in relation to rereport and foster care entry outcomes.

BACKGROUND: There have been mixed findings on whether neglect increases the risk of re-report and foster care entry compared with other child maltreatment types. Studies suggest that child neglect can be broken into subtypes which, in turn, may lead to different child welfare outcomes. OBJECTIVE: The purpose of this study was to examine whether the re-report and foster care entry outcomes varied by child maltreatment type, specifically the subtypes of neglect. PARTICIPANTS AND SETTING: Data were derived from a regional longitudinal study with linked CPS and state-level administrative data. Participants (n = 4867) included children with reports of physical abuse, sexual abuse, and six neglect subtypes. Cox regression models were used to assess the probability of recurrence and foster care entry outcomes. RESULTS: Most neglect subtypes were more likely to be re-reported than physical abuse and sexual abuse, and physical neglect was more likely to be re-reported than lack of supervision (HR = 1.20). Children reported for physical neglect was more likely to be placed in foster care than children with reports of physical abuse and other subtypes of neglect. CONCLUSIONS: This study provides new insight on whether families with child neglect allegations are at increased risk of re-report or foster care entry outcomes compared with other child maltreatment allegations. Given the significant individual, system, and societal costs of neglect, we hope this study facilitates rigorous investigation of child neglect and the potentially modifiable factors or targets for intervening to prevent the re-report of child maltreatment and foster care entry.

Profiles of Teen Dating Violence and Association With Depression Among Chinese Teens.

Teen dating violence (TDV) is a major global public health concern. Few studies, however, have examined profiles of TDV in Chinese societies and how these profiles might be associated with teens' mental health. The current study analyzed a sample of 891 middle and high school students with dating experience in Taiwan, Hong Kong, and Shanghai. Latent class analysis (LCA) and multinomial logit regression analysis were performed in an attempt to identify profiles of TDV and then investigate possible associations between class membership and self-reported depression. The results of LCA suggested that a four-class model was the best fit for the data: Severe/Multi-Type TDV (5.51%), Controlling Behavior (13.08%), Non/Low TDV (64.50%), and Physical Violence (16.91%). The best-fit model suggested bidirectionality, meaning among teen partners in an abusive relationship, both tended to participate in violent acts and controlling behaviors. The results of the multinomial regression showed that, compared with the Non/Low TDV class, teens in the Severe/Multi-Type TDV class or Controlling Behavior class had greater odds of screening positive for depression. There was no significant difference in the risk of depression between the Physical Violence class and the Non/Low TDV class. Implications for future research and practice are also discussed herein.

Effects of Child Protective Custody Status and Health Risk Behaviors on Health Care Use Among Adolescents.

OBJECTIVE: To determine whether current protective custody status (ie, youth currently in the temporary or permanent custody of child protective services, eg, foster and kinship care) contributes to increased health care utilization compared to youth never in protective custody. Health characteristics (eg, mental health diagnoses) and behaviors (eg, substance use) were expected to account for differences in health care use among the two groups. METHODS: Retrospective child welfare administrative data and linked electronic health records data were collected from a county's child welfare system and affiliated freestanding children's hospital between 2012 and 2017. Youth currently in protective custody (n = 2787) were identified and demographically matched to peers never in custody (n = 2787) who received health care from the same children's hospital. Health care use, health risk behaviors, and social, demographic, and diagnostic data were extracted and compared for both cohorts. RESULTS: In baseline models, health care use was higher for youth in protective custody compared to peers. In adjusted models that included health risk behaviors and patient characteristics, protective custody status was associated with decreased primary and missed care, and no longer a significant predictor of other types of health care use. CONCLUSIONS: Youth had significantly higher utilization while in protective custody than their demographically similar peers; however, health risk behaviors appear to account for most group differences. Identification of current custody status in pediatric settings and addressing health risk behaviors in this population may be important for health care systems interested in altering health care use and/or cost for this population.

Neighborhood Contexts and Child Maltreatment Reports Among Families Receiving AFDC/TANF: A Longitudinal and Multilevel Study.

Attention to the relationship between neighborhood context and child maltreatment is growing. However, no study exists that considers families' residential moves and neighborhood changes longitudinally. This is particularly relevant to disadvantaged families who move frequently. Our sample includes children who experienced a child maltreatment report (CMR) or an AFDC case in early childhood. We followed up these children from 1995 to 2009 through various administrative databases. We used multilevel logistic growth curve models to estimate the CMR likelihood at each age from 1 to 16. Estimates were limited to ages on AFDC/TANF to trace families' residential addresses based on AFDC/TANF payee records. Our findings highlight the importance of tracing residential neighborhoods in a longitudinal study. While doing so, we identify some possible neighborhood contextual influences. These, however, are small in contribution to overall risk and are less observable among children that are more vulnerable.

Direct and Indirect Cost Savings From Sexually Transmitted Infection Testing, Treatment, and Counseling Among Foster Youth.

BACKGROUND: Sexually transmitted infection (STI) prevention programs can decrease the economic burden of STIs. Foster youth have higher rates of STIs compared with their peers; however, information on direct costs and indirect costs averted by STI testing, treatment, and counseling among foster youth is lacking. METHODS: This study used data from a comprehensive medical center for foster youth over a 3-year study period from July 2017 to June 2020. Direct and indirect costs averted by testing and treatment of chlamydia, gonorrhea, and syphilis, as well as HIV testing and counseling, were calculated based on formulas developed by the Centers for Disease Control and Prevention and adjusted for inflation. RESULTS: Among the 316 youth who received medical services during this time, 206 were sexually active and tested for STIs and/or HIV. Among 121 positive STI test results, 64.5% (n = 78) were positive for chlamydia, 30.6% (n = 37) were positive for gonorrhea, and 5.0% (n = 6) were positive for syphilis. Treatment was provided to all. Overall, $60,049.68 in direct medical costs and $73,956.36 in indirect costs were averted. CONCLUSIONS: Given the rates of STIs among this population and the economic benefit of STI treatment, it is imperative to continue to provide intensive and comprehensive, individualized sexual health care for foster youth. Traditional care management may miss the opportunity to prevent, identify, and treat STIs that comprehensive wraparound care can achieve. This study suggests that comprehensive wraparound care is a cost-effective way to identify, treat, and prevent STIs among foster youth.