Abstract
BACKGROUND AND OBJECTIVES: Children with disabilities are at an increased risk for maltreatment. However, the risk of maltreatment is unknown for children with specific types of birth defects. This study was conducted to determine whether the risk and predictors of maltreatment differ between children with and without 3 birth defects: Down syndrome, cleft lip with/without cleft palate, and spina bifida.
METHODS: This population-based study of substantiated childhood maltreatment was conducted in Texas from 2002 to 2011. Linked data were used to describe the risk and types of maltreatment that occurred before age 2 years in children with and without specific birth defects. Poisson regression was used to identify predictors of maltreatment and assess differences in those predictors between children with and without these specific birth defects.
RESULTS: The risk of maltreatment (any type) in children with cleft lip with/without cleft palate and spina bifida was increased by 40% and 58%, respectively, compared with children with no birth defects. The risk of any maltreatment was similar between children with Down syndrome and unaffected children. Across birth defect groups, the risk of medical neglect was 3 to 6 times higher than in the unaffected group. Child-, family-, and neighborhood-level factors predicted maltreatment in children with and without birth defects.
CONCLUSIONS: The overall risk of substantiated maltreatment was significantly higher for some but not all birth defect groups. The factors associated with increased risk were similar across groups. Enhancement of existing maltreatment prevention and early intervention programs may be effective mechanisms to provide at-risk families additional support.
- aRR —
- adjusted rate ratio
- BPA —
- British Pediatric Association
- CI —
- confidence interval
- CL ± P —
- cleft lip with or without cleft palate
- TBDR —
- Texas Birth Defects Registry
- TDFPS —
- Texas Department of Family and Protective Services
What’s Known on This Subject:
Children with disabilities are at an increased risk for maltreatment. As a group, birth defects have been shown to increase the risk of maltreatment.
What This Study Adds:
Few studies have examined maltreatment in children with specific birth defects. This study found many factors predictive of maltreatment in young children, including type of birth defect as well as child-, family-, and neighborhood-level factors.
Identifying and preventing child maltreatment is an established public health priority.1 Maltreatment of children with disabilities is of particular concern because this vulnerable group is overrepresented in the maltreated population.2 Although 9% to 18% of children aged <18 years in the United States have a disability, up to 54% of the children served by Child Protective Services reportedly have some form of disability.3–5 Furthermore, in population-based studies, the prevalence of maltreatment in children with disabilities is higher than in children without a known disability, and the relative proportions of maltreatment types (eg, physical abuse) vary across disability category (eg, language delay, physical disability).6–11
Birth defects, which occur in ∼1 in 33 births in the United States, include a range of abnormalities in function, structure, and metabolism12 and are a leading cause of disability.13,14 Compared with children without birth defects, children with birth defects are at increased risk for maltreatment and out-of-home placement.15,16 However, given the heterogeneous nature of birth defects, the medical needs and disabilities of affected children vary. Consequently, the risk and type of maltreatment may differ across birth defect categories.
The present study was undertaken to determine whether the overall risk, type, and predictors of maltreatment in children aged <2 years with specific birth defects differ from unaffected children. We evaluated 3 common birth defects that are characterized by different types of disabilities: Down syndrome (intellectual impairment), cleft lip with or without cleft palate (CL ± P; facial malformation and speech impairment), and spina bifida (physical disability).
Methods
This population-based cohort study included ∼3 million children born in Texas between 2002 and 2009. We examined substantiated child maltreatment, as defined by the State of Texas Family Law Code 261, and predictors of substantiated maltreatment among children 4 days to 2 years of age. State administrative and US census data were linked to identify the study population and covariates of interest. This study was approved by the institutional review boards at the Texas Department of State Health Services and the University of Texas Health Science Center at Houston.
Population and Sources
Data from birth and death records, obtained from the Texas Department of State Health Services Vital Statistics Unit, were used to identify the cohort. All live-born children delivered January 1, 2002, through December 31, 2009, to women who were Texas residents at the time of delivery were eligible to be included in the study (N = 3 130 324). Children who died on or before the third day of life (n = 6206) were excluded.
Children with birth defects were identified by using surveillance data from the Texas Birth Defects Registry (TBDR). TBDR is a population-based registry that collects data on birth defects throughout Texas by using an active surveillance approach.17 TBDR codes diagnoses by using the Centers for Disease Control and Prevention’s modification of the British Pediatric Association Classification of Diseases (BPA codes).
The birth defects included in the present study were: Down syndrome (BPA codes 758.000–758.090), CL ± P (BPA codes 749.100–749.220), and spina bifida (BPA codes 741.000–741.990). Potential cases with BPA codes for both spina bifida and anencephaly (BPA codes 740.000–740.100) were excluded. Children with multiple defects were included unless they had ≥2 of the conditions under study (eg, children with both CL ± P and spina bifida were excluded). Children unaffected by any type of birth defect, as determined by TBDR, comprised the “unaffected” comparison group.
Maltreatment information was obtained from the Child Protective Service records from the Texas Department of Family and Protective Services (TDFPS). Texas defines 8 types of child maltreatment: neglectful supervision, physical abuse, physical neglect, medical neglect, sexual abuse, abandonment, emotional abuse, and refusal to assume parental responsibility (definitions can be found in the Supplemental Appendix).18 The main outcome variable for these analyses was any type of substantiated maltreatment reported between days 4 and 730 of life (first 2 years of life). Substantiated maltreatment is defined as an overall case disposition of “reason to believe,” meaning that after a complete investigation by TDFPS, a court ruled that there was a preponderance of evidence indicating that maltreatment had occurred. Child-level data on all substantiated child maltreatment cases over a 10-year period (January 1, 2002–December 31, 2011) for children born between January 1, 2002, and December 31, 2009, were obtained from TDFPS. For children with >1 maltreatment case before their second birthday, only the information for the first substantiated case was used. Date of the reported allegation(s), maltreatment type(s), perpetrator type(s) (eg, parent), and allegation disposition(s) (eg, reason to believe) were provided for each substantiated case. Information on unsubstantiated cases and any caregiver changes that may have occurred between birth and the time of maltreatment were not provided by TDFPS. Children identified by TDFPS who did not have a substantiated type of maltreatment (n = 552) or whose date of report was >1 week after their date of death (n = 56) were excluded because the circumstances surrounding the maltreatment case were unclear.
Covariates
All covariates were selected a priori based on the maltreatment literature. Covariate data for children with and without birth defects were ascertained from birth certificates and included information about the child (gestational age at birth, birth weight, gender, and plurality), mother (race/ethnicity, age, education, parity, marital status, type of payment at birth, and census tract of residence at the time of birth), and father (race/ethnicity, age at birth, and education). Because paternal information was missing for >10% of children, a dichotomous variable indicating whether information on the father was present or absent was created.
For the birth defect groups, children were categorized as having an isolated birth defect (1 major birth defect) or multiple major birth defects. Major defects are defined as conditions causing structural changes to the body that have a significant impact on health, development, or ability.19 This variable was included in all analyses that were restricted to include data only from children with a specific birth defect (ie, no unaffected children) because the presence of additional major birth defects may be associated with the risk of maltreatment.
Neighborhood-level data, based on the mother’s residence at the time of birth, were obtained from the 5-year 2005 to 2009 American Community Survey and included percentage of poverty, vacant housing, unemployment, and single-headed households with children aged <5 years according to census tract. These variables were standardized against the distribution for the state of Texas and were used to create a composite measure for neighborhood economic impoverishment as described by Korbin et al.20 A neighborhood was considered to have high economic impoverishment if it was >1 SD above the mean level of impoverishment for the state.
Data Linkage
Data from TBDR and TDFPS were separately linked to vital statistics records by using probabilistic data strategies that included 12 matching algorithms using unique combinations of parent and child names, dates of birth, and Social Security numbers. The linked TBDR and TDFPS were then merged by using the birth certificate number as the unique identifier. Neighborhood-level data, calculated from the American Community Survey, were merged by using census tract numbers. Data linkage and de-identification was performed by TBDR data analysts using SAS version 9.1 (SAS Institute, Inc, Cary, NC).
Statistical Analysis
Counts and frequencies were used to describe the 4 study groups (unaffected, Down syndrome, CL ± P, and spina bifida). A χ2 test was used to compare characteristics of each birth defect group with the unaffected group. P values <.05 were considered statistically significant.
Poisson regression was used to estimate the unadjusted risk ratio and 95% confidence interval (CI) for the association between each birth defect group and substantiated maltreatment. In addition, among children who had a substantiated maltreatment report, unadjusted relative risks comparing maltreatment case characteristics (eg, maltreatment type) of each birth defect group versus the unaffected group were estimated.
Modified Poisson regression with a robust SE adjustment was used to identify predictors of overall maltreatment. Adjusted rate ratios (aRRs) were estimated, and 95% CIs that did not include 1.00 were considered statistically significant. Analyses were initially conducted separately for each birth defect group. Specifically, data from unaffected children and children with a specific birth defect (eg, CL ± P) were fitted to a model that included all of the covariates that were selected a priori. To determine whether the association of each covariate with maltreatment differed according to birth defect status (eg, the association between plurality and maltreatment depending on the presence or absence of CL ± P), interaction terms involving a birth defect indicator variable and each covariate were individually added to the models, and likelihood ratio tests were used to compare the fit of the models with and without interaction terms. Interaction terms were considered statistically significant if P values were <.05. Because the tests for interaction were all nonsignificant (data not presented), the data from all groups (Down syndrome, CL ± P, spina bifida, and unaffected) were combined and fitted to the final model including all covariates. Statistical analyses were performed by using Stata version 12.1 (StataCorp, College Station, TX).21
Results
The percentage of successful matches between TBDR and vital statistics records was >98% for all years. Overall, 85.8% of children with a substantiated TDFPS investigation during the study period could be matched to vital statistics records. The rate of matching increased over the study period from 81.7% in 2002 to 90.6% in 2009 (Table 1).
Results of Data Linkage Between Birth, Substantiated Child Maltreatment, and Birth Defect Record, Texas, 2002 to 2009
Population
A total of 2 977 758 children were included in these analyses (95% of live births in Texas, 2002–2009), including 2 970 101 unaffected children, 3743 with Down syndrome, 2943 with CL ± P, and 971 with spina bifida. The birth defect groups differed from the unaffected group on a range of demographic factors (Table 2). In general, these differences were consistent with known epidemiologic characteristics of the birth defect group. For example, mothers of children with Down syndrome tended to be older than mothers of unaffected children, which is consistent with the known association between Down syndrome and maternal age.22
Demographic Characteristics According to Study Group in Children Aged 4 Days to 2 Years, Texas, 2002 to 2011
Overall Maltreatment
Of the 2 970 101 unaffected children included in the study, 67 049 (23 per 1000 children) had a substantiated case of maltreatment before their second birthday (Table 3). In the Down syndrome group, the prevalence of maltreatment (20 per 1000) was not significantly different from the prevalence in the unaffected group (aRR: 1.08 [95% CI: 0.85–1.37]). In contrast, the prevalence and relative risk of maltreatment in the CL ± P group (prevalence: 36 per 1000; aRR: 1.40 [95% CI: 1.14–1.71]) and the spina bifida group (prevalence: 38 per 1000; aRR: 1.58 [95% CI: 1.12–2.24]) were significantly higher compared with the unaffected group.
Overall Prevalence (per 1000 Children) and Risk of Maltreatment According to Study Group in Children Aged 4 Days to 2 Years, Texas, 2002 to 2011
Characteristics of Maltreatment
Among children with substantiated maltreatment, age at the time of the first substantiated maltreatment report was divided into 5 categories: <1 month, 1 to 6 months, 7 to 12 months, 13 to 18 months, and 19 to 23 months (Table 4). Among maltreated children, those with a birth defect were ∼2 to 3 times more likely than unaffected children to have an episode of maltreatment reported before 1 month of age; this difference was statistically significant only for the CL + P and spina bifida groups.
Characteristics of Maltreatment: Crude RR According to Age and Specific Maltreatment Type in Children Aged 4 Days to 2 Years, Texas, 2002 to 2011
Neglectful supervision was the most common form of maltreatment in all 4 groups (Table 4). Compared with unaffected maltreated children, the risk of neglectful supervision was similar (Down syndrome) or significantly lower in children with birth defects (CL ± P and spina bifida). In contrast, the risk of medical neglect was 3.6 to 6.2 times higher in the 3 birth defects groups compared with the unaffected group. The risk of multiple types of maltreatment was not significantly different in any of the 3 birth defects groups compared with their unaffected counterparts.
Parents were the most frequent perpetrators of substantiated maltreatment in each group (93%–100%). No significant differences in perpetrator type were found between the birth defects groups and the unaffected group (data not shown).
Predictors of Maltreatment
Analyses of covariates to identify predictors of maltreatment were initially conducted separately by using data from unaffected children and children with a specific birth defect (eg, CL ± P). The covariate indicating whether the child had an isolated birth defect or multiple major birth defects was included in group-specific regression models. This covariate was not significant in crude or adjusted models for any birth defect group (data not presented). A comparison of the fully adjusted models revealed that the predictors of maltreatment (ie, the association between maltreatment and each covariate) were similar across models, and interaction terms were not statistically significant. The adjusted model for the entire population (all groups combined) is presented in Table 5 and is described in the following sections.
aRRs for Maltreatment in Children Aged 4 Days to 2 Years
Child-level Predictors
After adjusting for all other factors, children with a substantiated report of maltreatment were more likely than children without a substantiated report to have been premature (aRR: 1.18 [95% CI: 1.15–1.21]) or very premature (aRR: 1.28 [95% CI: 1.21–1.35]); have a low birth weight (aRR: 1.39 [95% CI: 1.34–1.43]); and be male (aRR: 1.05 [95% CI: 1.04–1.07]). Twins and other multiples were less likely to have a substantiated maltreatment case than were singletons (aRR: 0.55 [95% CI: 0.52–0.58]).
Family-level Predictors
All of the family characteristics that were examined varied significantly between children with and without a substantiated maltreatment report. Compared with mothers of children without a substantiated maltreatment report, mothers of maltreated children were more likely to have had previous live births (aRRs ranging from 1.6–3.2 comparing mothers with 1 to ≥3 previous births vs primiparous mothers), a high school education or less (aRR<HS vs >HS: 2.43 [95% CI: 2.36–2.49]; aRRHS vs >HS: 1.79 [95% CI: 1.75–1.84]), and had the delivery costs covered by Medicaid (aRRMedicaid vs other: 2.71 [95% CI: 2.66–2.76]). The risk of maltreatment was also elevated for children whose mothers were young, unmarried, and who did not include paternity information on the birth certificate compared with children whose mothers were older, married, and who did include paternity information on the birth certificate (aRR: 1.5–1.8). Children of white non-Hispanic mothers were at significantly greater risk of maltreatment compared with children with mothers in all other race/ethnic groups (aRR: 1.6–2.4).
Neighborhood-level Predictors
Neighborhood-level economic impoverishment was also associated with an increased risk of maltreatment. Children living in a neighborhood with a high degree of economic impoverishment were 1.12 times (95% CI: 1.09–1.15) more likely to have a substantiated maltreatment report than those in less impoverished neighborhoods.
Discussion
The risk of overall maltreatment before 2 years of age was significantly higher for children with CL ± P and spina bifida compared with unaffected children, even after adjusting for child-, family-, and neighborhood-level factors. In contrast, children with Down syndrome were not at increased risk of maltreatment before 2 years of age compared with unaffected children. These findings suggest that, in addition to the presence of a birth defect, the specific condition (ie, CL ± P and spina bifida) may be an important predictor of the risk of maltreatment before age 2 years.
The results of these analyses also indicate that children with birth defects are at risk for different types of maltreatment compared with their unaffected contemporaries. Among children with substantiated maltreatment, the risk of medical neglect was significantly higher among all 3 birth defect groups than in the unaffected group. Their medical complexity may account for the increased risk for medical neglect in these groups versus other forms of neglect. In addition, children with birth defects were significantly more likely than unaffected children to be maltreated during the first month of life, although this association was statistically significant for the CL ± P and spina bifida groups only. For both CL ± P and spina bifida, the medical care typically required during the first few months of life is complex, and it is recommended that these conditions be treated by specialized teams of providers.23–25
All of the family-level risk factors exhibited were associated with substantiated maltreatment (relative risk ≥1.5). Poverty-related factors, such as low maternal education, Medicaid use, and multiple previous births, were associated with 2- to 3-fold increases in risk for maltreatment in this study and have been consistently associated with maltreatment in earlier studies.15,26–28 In contrast, our finding that children whose mothers were white non-Hispanic were nearly twice as likely to be maltreated as any other race ethnic group, after adjusting for all factors, has not been consistently found by other investigators. However, similar trends were reported in a large linkage study from California, a state demographically comparable to Texas.15
It is possible that some substantiated cases of maltreatment included in these analyses were based on a family’s previous involvement with TDFPS and not maltreatment of the study child. To decrease this possibility, we excluded substantiated cases reported before day 4 of life. In addition, because children with birth defects are likely to interact with mandatory reporters more often than unaffected children, surveillance bias is a concern.
Family mobility (out of Texas) was not taken into account. Consequently, children with substantiated maltreatment reports in other states were not classified as maltreated in these analyses. Therefore, the prevalence of maltreatment in all groups is likely an underestimate.
Conclusions
Findings from this study suggest that the specific type of birth defect may be an important factor to consider when identifying children aged <2 years who are at increased risk for maltreatment. Further analysis should be conducted to determine whether the increased risk of maltreatment in children with specific birth defects persists throughout childhood. In addition, children with other birth defects should be examined to determine whether they are also at increased risk of maltreatment compared with unaffected children.
The results of this study, in combination with findings from other population-based studies of maltreatment in children with disabilities,6,7,15,28 may inform policies and services aimed at improving the well-being of families.V For example, the Maternal, Infant, and Early Childhood Home Visiting Program of the Patient Protection and Affordable Care Act has provided all states with funding to support home visitation programs aimed at improving the outcomes of at-risk families with young children. Our findings could be used to enhance the targeting of populations with the highest risk for maltreatment. In addition, young children with developmental delays, including children with the birth defects examined in the present study, qualify for early childhood intervention services (Part C) provided by the Individuals with Disabilities Education Improvement Act. These services are provided in the home and are intended to support both the family and the child. Despite these services being free or low-cost, many qualifying children do not receive them.29 Increasing enrollment in existing services, as well as enhancing these services to include child maltreatment prevention strategies,30–34 may prove to be an efficient and cost-effective way of preventing maltreatment in children with birth defects.
Acknowledgments
The authors thank Dr Brock Boudreau and the TDFPS for their support and entrusting the study authors with their data. They also thank Dr Margaret Caughy for her guidance and support.
Footnotes
- Accepted September 15, 2015.
- Address correspondence to Bethanie S. Van Horne, DrPH, University of Texas Health Science Center at Houston, Developmental Pediatrics, 7000 Fannin St, Suite 1920, Houston, TX 77030. E-mail: bethanie.s.carr{at}uth.tmc.edu
Dr Van Horne designed the study, analyzed the data, and prepared the manuscript; Ms Moffitt prepared, cleaned, and linked the data; Drs Canfield and Case conceptualized the study, assisted in acquiring the data from multiple agencies, and reviewed the manuscript; Dr Greeley assisted with the design and interpretation of results and reviewed drafts of the manuscript; Dr Morgan assisted with the design and data analysis and reviewed drafts of the manuscript; Dr Mitchell assisted in the design of the study and interpretation of results and reviewed multiple drafts of the manuscript; and all authors approved the final version of the manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported in part through a cooperative agreement (#5U01DD000494-04) between the Centers for Disease Control and Prevention and the Texas Department of State Health Services, as well as through funding from the Title V Block Grant at the Texas Department of State Health Services.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
References
- Copyright © 2015 by the American Academy of Pediatrics
In this issue
