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.

Left behind? Educational disadvantage, child protection, and foster care.

BACKGROUND: Current educational policies for children involved with the Child Protective Services (CPS) system focus largely on the minority of children currently in or aging out of foster care, and target school stability and college access. OBJECTIVE: The present study investigates the nature of secondary (high school) education performance and attainment and post-secondary (college) enrollment among youth with prior or current CPS contact and their low-income, but not CPS-contacted, peers. METHOD: Following a cohort of over 63,000 high school students in Wisconsin, we use CPS investigation and placement records, and public school records to evaluate associations between CPS involvement and educational attainment. RESULTS: CPS-contacted youth have lower educational performance and greater academic challenges than their low-income peers. Youth aging out of care are uniquely disadvantaged with regard to on-time high school completion but complete high school and enroll in college at equal or higher rates than reunified youth. Across all groups, 55-75 % of those who graduated on time with "basic" or above English and math skills enrolled in college. Foster care experiences, such as time in care and placement instability, were not consistently associated with educational outcomes. CONCLUSION: Efforts to improve secondary education experiences are needed to bolster college and career pathways for disadvantaged youth.

What Does Child Protective Services Investigate as Neglect? A Population-Based Study.

Most child protective services (CPS) investigations involve allegations of neglect. Broad and vague definitions have led to concerns that CPS-investigated neglect is driven by poverty-based material hardship. In a representative sample of 295 neglect investigations in California in 2017, structured data and narrative text fields were used to characterize the types of neglect and concurrent parental risk factors investigated by CPS and to assess the rate and nature of investigated physical neglect, defined as inadequate food, housing, or hygiene. The most common types of neglect were inadequate supervision (44%) and failure to protect (29%), followed by physical neglect (14%). Common risk factors identified in neglect investigations were parental substance use (41%), domestic violence (21%), mental illness (18%), and co-reported physical or sexual abuse (29%). Nearly all investigations of physical neglect (99%) included concerns related to substance use, domestic violence, mental illness, co-reported abuse or an additional neglect allegation (i.e., abandonment). Given concerns identified in neglect investigations, economic supports are likely insufficient without an array of behavioral-health supports.

Rates of Emotional Disturbance Among Children in Foster Care: Comparing Federal Child Welfare Data and Medicaid Records in Two States.

The purpose of this study is to estimate the rate of emotional disturbance (ED) among children in foster care and assess the validity of the national foster care census data (AFCARS) measure of ED. This study used linked child protection and Medicaid records from 2014 and 2015, for the states of California and Wisconsin, as well as data from AFCARS, a federal population census of children in foster care which states are mandated to contribute to. ED is defined by AFCARS and includes an array of mental and behavioral health diagnoses. According to AFCARS, 13% of CA children in foster care and 15% of WI children in foster care had an ED, whereas Medicaid claims produce rates of 45% and 48%, respectively. Rates of ED among children in congregate care were underestimated by 43-46 percentage points, with substantial proportions having diagnoses of disruptive behavioral disorders. Despite the AFCARS ED measure being cited in congressional testimonies and its wide use in research, results from this study suggest that the AFCARS ED estimates are an unreliable metric for use in research, policy, or practice.

Caseworker documentation of mothers in foster care: A case study in California.

BACKGROUND: Young people in foster care frequently experience early parenting, and U.S. law requires the documentation of parents in care. The extent to which the child protection system (CPS) has accurately collected information on parents in care remains largely unknown and has ramifications for the planning and delivery of services. OBJECTIVE: Using California as a case study, this study documented young mothers in care using birth records, determined the proportion of mothers correctly identified in CPS records, and compared characteristics and placements of mothers who were or were not identified. POPULATION: All mothers aged 15-21 years documented by vital birth records (2014 and 2019) were probabilistically linked to CPS records, yielding 3199 mothers in care. METHOD: We compared mothers who were or were not identified in CPS records using bivariate statistics. A chi-square test for trends examined differences in documentation over time. RESULTS: A third of mothers were not identified as such in CPS records (n = 1136). The proportion of mothers identified by caseworkers rose from 50 % in 2014 to 71.5 % in 2019 (p < .001). Nonminor dependent mothers and those placed in supervised independent living programs were more likely to be misclassified in CPS data. CONCLUSIONS: Motherhood may not be documented in CPS records; however, significant improvements were evident. Results suggest that all nonminor dependents should be informed of available services for expectant and parenting individuals, especially if they are living in less restrictive placements, given they are more likely to give birth and less likely to be identified in CPS records than minor dependents.

A Longitudinal Analysis of Concerning Psychotropic Medication Regimens Among Adolescents in Foster Care.

PURPOSE: To provide a population-based examination of psychotropic medication use before and after entry into foster care (FC), with special attention on the use of concerning medication regimens: polypharmacy, stimulants, and antipsychotics. METHODS: Using linked administrative Medicaid and child protective service data from Wisconsin, we follow a cohort of early adolescents ages 10-13 years who entered FC between June 2009 and December 2016 (N = 2,998). Descriptive statistics and Kaplan Meyer survival curves illustrate the timing of medication. Cox proportional hazard models identify hazard of outcomes (new medication, polypharmacy, antipsychotic, and stimulant medication) during FC. Separate models were run for adolescents with and without a psychotropic medication claim in the six months before FC. RESULTS: Overall 34% of the cohort entered with a pre-existing psychotropic medication, accounting for 69% of adolescents with any psychotropic medication claim during FC. Similarly, the majority of adolescents with polypharmacy, antipsychotics or stimulants during FC entered with those prescriptions. Among youth with pre-entry medication, rates of polypharmacy (56%), antipsychotic (50%) and stimulants (64%) were high. Among adolescents who entered FC with no prior medication, placement disruptions (30 days before or after) predicted new medication. DISCUSSION: Although a great deal of attention - and policies - have focused on youth in care, there is high reliance on psychotropic medications within the broader population of maltreated adolescents, indicating a need for timely and accurate re-assessment of current and past medications upon entry. Adolescents should also be actively involved in their own health care.

Prevalence of Mental Health Diagnoses Among Early Adolescents Before and During Foster Care.

OBJECTIVE: This study aimed to characterize the rates and types of diagnosed mental health (MH) disorders among children and adolescents before and during foster care (FC) overall and by race and ethnicity. METHODS: We used population-based linked administrative data of medical assistance (public insurance) claims records and child protective services data from a cohort of early adolescents who entered FC at 10 to 14 years old. MH diagnoses were coded according to the International Classification of Diseases, Ninth and Tenth Revisions, Clinical Modification (ICD-9 and ICD-10) and included adjustment disorders, disruptive disorders, attention-deficit/hyperactivity disorder (ADHD), anxiety disorders, mood disorders, attachment disorders, autism, and other disorders. RESULTS: Before FC entry, 41% of children and adolescents had at least 1 MH diagnosis. Attention-deficit/hyperactivity disorder (25%), mood disorders (18%), and disruptive disorders (15%) were the most common pre-entry diagnoses. Among early adolescents entering FC with no previous diagnosis, 52% were later diagnosed with adjustment disorder (accounting for 73% of all youth with a new diagnosis during FC). White early adolescents had higher rates of diagnosed MH disorders before FC, whereas racial/ethnic minority early adolescents were more likely to receive a MH diagnosis during FC. Black early adolescents were more likely than White and Hispanic early adolescents to be diagnosed with disruptive disorders and less likely to be diagnosed with anxiety or adjustment disorders during FC. CONCLUSION: Results highlight the high rates of MH needs among early adolescents before entry into FC, whether detected before or during FC. Results also illustrate disparities in pre-entry MH care between racial/ethnic minority and White early adolescents, with minority youth less likely to be receiving services before entry.

Three-Year Custody Outcomes Among Infants Investigated by Child Protection Systems for Prenatal Substance Exposure in California.

OBJECTIVE: Infants affected by prenatal alcohol and drug use are more likely to be removed from parental custody than those in the general population, although it is unclear whether their custody outcomes differ from infants investigated by child protection systems (CPS) for other reasons. This analysis seeks to compare trajectories of involvement and custody outcomes among infants investigated by CPS with and without documentation of prenatal substance exposure (PSE). METHOD: We used vital birth records linked to administrative CPS records to examine the timing of system involvement and 3-year custodial outcomes among investigated infants with and without identified PSE. We defined PSE according to documentation on the state's standardized hotline screening form, which CPS completes upon referral for alleged maltreatment. We estimated the likelihood a child was in nonparental custody at age 3 by specifying multivariable generalized linear models, adjusted for covariates available in the birth record. RESULTS: In our sample of 22,855 infants investigated by CPS in 2017 in California, more than 26% had documentation of PSE. These infants experienced an accelerated timeline of system penetration and were 2.2 times as likely to be in nonparental placement at age 3. DISCUSSION: PSE confers an independent risk of custody interruption among infants investigated by CPS. The younger age of these infants, complexity of parental substance use, and potential misalignment of administrative permanency timelines with parental recovery all suggest the need for increased research, policy, and programmatic interventions to serve this vulnerable population.

Children with PSE face environmental risks in the early developmental period. Often in the United States, CPS is relied on to assess and mitigate these risks. Amid calls for a public health response to PSE, it is essential to understand how children with PSE interact with CPS. We describe the incidence and timing of custody interruptions in a large U.S. state, comparing infants with PSE to those investigated by CPS for other reasons. This study extends current understanding by demonstrating the independent risk of custody interruption conferred by PSE status.

Lifetime rates and types of subsequent child protection system contact following a first report of neglect: An age-stratified analysis.

An estimated 1 in 3 U.S. children will be the subject of a child protective services (CPS) investigation during their lifetime, typically for allegations of neglect. Whether and how an initial report of neglect is addressed may place children on divergent trajectories for safety and stability throughout childhood. The purpose of this study is to track subsequent CPS contact among children born in California in 2000 who were first investigated by CPS for neglect allegations alone (no co-occurring abuse) and not permanently separated from their families of origin (i.e., not removed or reunified if removed). We estimated the rates of subsequent CPS referrals, substantiated maltreatment, placement in foster care, and allegations of physical and sexual abuse by age 18. We assessed how rates of subsequent contact varied by initial CPS response and age at first investigation. Supplemental analyses disaggregated data by race and ethnicity. Results indicate that 64% of children initially investigated for neglect alone were re-referred to CPS by age 18 and 16% experienced a subsequent removal; however, these estimates varied greatly by age. Four out of five (79% to 83%) of children initially investigated as infants had one or more subsequent CPS referrals during childhood. Children were not only re-referred for allegations of neglect; more than half of children re-referred were reported for allegations of physical or sexual abuse, indicating that abuse risk was either missed during the initial CPS investigation or escalated afterward. The failure to address maltreatment risks when children first present to the system is a complex problem with no easy solution. Our findings document that a majority of children initially referred for neglect experience future CPS involvement, often for allegations of physical or sexual abuse.

A Population-Based Study of Mental Health Diagnoses and Child Protection System Involvement Among Medicaid-Insured Children.

OBJECTIVE: To determine the population prevalence of diagnosed mental health disorders among Medicaid-insured children <18 years old in California based on levels of current and past child protection system (CPS) involvement. STUDY DESIGN: In this retrospective, population-based study, we examined the full population of children enrolled in California's Medicaid program for at least 1 month between 2014 and 2015 and who had at least 1 claim during that period (n = 3 352 886). Records for Medicaid-insured children were probabilistically linked to statewide CPS records of maltreatment and foster care placements since 1998. A primary or secondary mental health diagnosis was classified using International Classification of Diseases codes. RESULTS: Overall, 14% (n = 470 513) of all children insured through Medicaid in 2014-2015 had a documented mental health diagnosis. Among children with a diagnosis, the percentage with CPS involvement (ie, any report for maltreatment) was nearly twice that of the Medicaid population overall (50.4% vs 26.9%). This finding held across all diagnostic groups but with notable variations in magnitude. A graded relationship emerged between the level of CPS involvement and the likelihood of a mental health diagnosis. Diagnoses among children reported for maltreatment were common, regardless of placement in foster care. CONCLUSIONS: Findings document high rates of both mental health diagnoses and past child protection involvement in a population of Medicaid-insured children. Most children reported for maltreatment will never be placed in foster care, underscoring the importance of ensuring that the children who remain at home receive the proper array and coordination of services.

Individual and Social Risk and Protective Factors as Predictors of Trajectories of Post-traumatic Stress Symptoms in Adolescents.

The present study elucidates heterogeneity in post-traumatic stress symptoms (PTSS) across adolescence in a sample of youth who have experienced myriad types and combinations of potentially traumatic events (PTEs), including substantiated physical abuse, sexual abuse, neglect and/or at least one other self-reported PTE. A machine learning technique was used to assess a multivariate set of variables (e.g., PTEs, individual risk and protective factors, social risk and protective factors, and racial and ethnic minority status) as predictors of PTSS trajectory group membership. The sample included 498 maltreated (n = 275) and comparison (n = 223) 14-19-year-old female adolescents (M = 15.27, SD = 1.06 at Time 1) assessed annually until age 19. 45.7% of participants were White, 45.3% Black, 0.4% Native American, 0.8% Hispanic, and 7.7% other. Growth mixture modeling identified three distinct trajectories of PTSS: (1) recovery (56%); (2) moderate, chronic (25%); and (3) high, chronic (19%). An elastic net model was used to test predictors of membership in the recovery versus the high, chronic PTSS trajectory groups. Results demonstrated that the recovery trajectory was characterized by the absence of sexual abuse, physical abuse, and other traumas, higher self-esteem, less affective dysregulation, less risky peers, lower levels of parent depression, and being of racial/ethnic minority status. Findings help to characterize individual variation in trajectories of PTSS following PTEs by underscoring the unique trauma responses of racial and ethnic minority youth and offering possible targets of interventions to promote recovery from PTSS.

A Population-Based Examination of Suicide and Child Protection System Involvement.

PURPOSE: The purpose of this study was to provide a population-based analysis of child protection system (CPS) involvement among children and adolescents who died by suicide. METHODS: We performed a case-control study of child and adolescent suicide and CPS involvement. Using linked birth, death, and CPS records, we longitudinally followed all children born in California in 1999 and 2000 (N = 1,052,333) in CPS and death records through 2017. Cases were defined as children who died in California and had a manner of death coded as suicide using the International Classification of Diseases, 10th revision (n = 170). Each suicide case was matched to four living controls, and children were classified based on CPS exposure: no history, reported for alleged child maltreatment, substantiated for child maltreatment, and placed in foster care. Crude suicide rates were documented, and conditional logistic regression models were used to estimate the adjusted odds of suicide. RESULTS: Among children and adolescents who died by suicide, 56.5% had a history of past allegations of abuse or neglect. Children with any CPS history had three times the odds of suicide compared to children with no history. No additional risk was found for children substantiated or placed in foster care compared to children with only an allegation. CONCLUSIONS: Suicide risk is not isolated to the relatively small group of children and youth placed in foster care. Findings reinforce the importance of increased attention to the experiences of the larger universe of children who remain at home after alleged or substantiated maltreatment.

A population-based case control study of suicide among youth reported for abuse and neglect.

BACKGROUND: Research has indicated an association between child abuse and adolescent suicide. Little population-based information exists, however, about the nature of maltreatment experiences or interactions with the child protection system (CPS). OBJECTIVE: To examine child maltreatment characteristics and system-level responses associated with risk of adolescent suicide. PARTICIPANTS AND SETTING: Linked vital death records and CPS records were used to identify the population of adolescents who died by suicide in California between 2010 and 2017 and who had a history of at least one report to CPS prior to death. METHOD: A case control design was used, with cases defined as a suicide of an adolescent with a history of CPS involvement. Using CPS records, living controls were then matched to cases based on year of birth, sex, race and ethnicity, and age of first child maltreatment allegation. A conditional logistic regression model was used to estimate the adjusted odds of adolescent suicide across various CPS and maltreatment characteristics. RESULTS: Recent CPS involvement, allegations of physical abuse, and allegations of sexual abuse emerged as significant risk factors for death by suicide. No differences in suicide risk were observed between youth with unsubstantiated or substantiated allegations. CONCLUSIONS: Suicide risk appears to be more closely tied to specific maltreatment experiences than to substantiation or placement into foster care. As adolescent suicide rates rise, better a understanding of risk factors among already vulnerable populations of youth is critical.

The Children's Data Network: Harnessing the scientific potential of linked administrative data to inform children's programs and policies.

The Children's Data Network (CDN) is a data and research collaborative focused on the linkage and analysis of administrative records. In partnership with public agencies, philanthropic funders, affiliated researchers, and community stakeholders, we seek to generate knowledge and advance evidence-rich policies that improve the health, safety, and well-being of the children of California. Given our experience negotiating access to and working with existing administrative data (and importantly, data stewards), the CDN has demonstrated its ability to perform cost-effective and rigorous record linkage, answer time-sensitive policy- and program-related questions, and build the public sector's capacity to do the same. Owing to steadfast and generous infrastructure and project support, close collaboration with public partners, and strategic analyses and engagements, the CDN has promoted a person-level and longitudinal understanding of children and families in California and in so doing, informed policy and program development nationwide. We sincerely hope that our experience-and lessons learned-can advance and inform work in other fields and jurisdictions.

A Descriptive Analysis of the Child Protection Histories of Youth and Young Adults Arrested in California.

In this study, we use population-based linked administrative data to document the full child protective service (CPS) histories of arrested youth and young adults. We extracted records for all individuals aged 24 and under who were arrested in California in 2014 and 2015. These records were probabilistically linked to statewide CPS records dating back to 1998. Overall, 43.4% of those arrested had a history of CPS involvement. Among individuals with uncensored CPS records (born in 1998 or later), 60.2% had past CPS involvement. Findings indicate that youth and young adults booked for a felony offense were more likely to have a history of CPS involvement than those booked for misdemeanors. A multinomial model served to confirm bivariate findings. This study provides further evidence that community concerns of childhood maltreatment were common among criminal justice-involved young adults.

A birth cohort study of Asian and Pacific Islander children reported for abuse or neglect by maternal nativity and ethnic origin.

Research identifying racial and ethnic disparities in child protective services (CPS) involvement in the U.S. has focused on the overrepresentation of Black children and the growing Latino child population. Little attention has been paid to children of Asian origin, the most underrepresented group of children in the U.S. CPS system. The objective of this analysis was to examine subgroup patterns of CPS involvement in California for Asian and Pacific Islander (API) children prospectively based on maternal nativity and ethnic origin. We extracted data for API children born in California in 2006 and 2007 (N=138,858) from population-based birth records and linked those records to CPS records spanning the first 5 years of life (through 2012). We assessed distributional differences in risk indicators for the full birth cohort of API children and calculated a summary risk variable representing the cumulative number of risks present at birth. Generalized linear models were used to estimate API children's adjusted relative risk of CPS report by subgroup. Overall, 12.2% of children born in California in the 2006-2007 birth cohort were API. The majority of API children had foreign-born mothers (80.9%). Children of U.S.-born Hawaiian, Guamanian, or Samoan mothers had the highest rate, with 20.4% being reported to CPS by their 5th birthday. The lowest rates of child abuse and neglect reporting were observed among children of foreign-born Asian Indian (2.5%), Korean (2.7%), and Chinese (2.8%) mothers, compared to 5.4% of all Asian and Pacific Islander children, and 14.8% of children in general population. Findings underscore the presence of disparities in CPS involvement among API children, which has implications for health and well-being across the life course and for targeted maltreatment prevention strategies.