Identifying children at high risk for a child maltreatment report

Child Abuse & Neglect 35 (2011) 96–104

Contents lists available at ScienceDirect

Child Abuse & Neglect

Identifying children at high risk for a child maltreatment report夽 Howard Dubowitz ∗ , Jeongeun Kim, Maureen M. Black, Cindy Weisbart, Joshua Semiatin, Laurence S. Magder Department of Pediatrics, University of Maryland School of Medicine, 520 W. Lombard Street, Baltimore, MD 21201, USA

a r t i c l e

i n f o

Article history: Received 11 September 2009 Received in revised form 7 September 2010 Accepted 10 September 2010 Available online 4 March 2011 Keywords: Child maltreatment Child abuse Neglect Risk factors Prevention

a b s t r a c t Objective: To help professionals identify factors that place families at risk for future child maltreatment, to facilitate necessary services and to potentially help prevent abuse and neglect. Method: The data are from a prospective, longitudinal study of 332 low-income families recruited from urban pediatric primary care clinics, followed for over 10 years, until the children were approximately 12 years old. Children with prior child protective services involvement (CPS) were excluded. The initial assessment included sociodemographic, child, parent and family level variables. Child maltreatment was assessed via CPS reports. Risk ratios (RRs) and their 95% confidence intervals (CIs) were estimated using Cox regression models. Results: Of the 224 children without a prior CPS report and with complete data who were followed for an average of 10 years, 97 (43%) later had a CPS report. In a multivariate survival analysis, 5 risk factors predicted CPS reports: child’s low performance on a standardized developmental assessment (RR = 1.23, 95% CI = 1.01–1.49, p = .04), maternal education ≤ high school (RR = 1.55, CI = 1.01–2.38, p = .04), maternal drug use (RR = 1.71, CI = 1.01–2.90, p < .05), maternal depressive symptoms (RR per one standard deviation higher score = 1.28, CI = 1.09–1.51, p < .01), and more children in the family (RR per additional child = 1.26, CI = 1.07–1.47, p < .01). Conclusions: Five risk factors were associated with an increased risk for later maltreatment. Child health care and other professionals can identify these risk factors and facilitate necessary services to strengthen families, support parents and potentially help prevent child maltreatment. © 2011 Published by Elsevier Ltd.

Introduction The prevalence and toll of child maltreatment (i.e., all forms of abuse and neglect) have been amply established (Pilowsky et al., 2008; Gaudin, 1999; Sedlak et al., 2010; Teicher et al., 2004; US DHHS, 2010). The challenge is to develop and evaluate promising strategies to help prevent this complex problem (Dubowitz & Guterman, 2005; Macmillan et al., 2009). Preventive interventions need to build on a clear understanding of the risk factors and etiology of child maltreatment.

夽 This grant was funded by the U.S. Department of Health and Human Services, Administration on Children and Families, Office on Child Abuse and Neglect, 90CA1749. ∗ Corresponding author. 0145-2134/$ – see front matter © 2011 Published by Elsevier Ltd. doi:10.1016/j.chiabu.2010.09.003

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

97

Child healthcare and other professionals can help prevent child maltreatment (Dubowitz, 2002). A comprehensive approach to children’s health, especially in primary care, offers an excellent opportunity to help ensure children’s safety, health and development, and to prevent maltreatment (Green, 1994; Hoekelman, 1997). Within populations considered at risk of maltreatment (e.g., low income), most families do not abuse or neglect their children. Professionals involved with children and families can screen for risk factors to identify those in need of services. They can also pinpoint the nature of specific problems needing help, such as parental depression. Thus, interventions that reduce the extent and severity of risk factors should strengthen families, support parents, enhance the care of children, and potentially help prevent maltreatment. The present study was guided by the ecological theory that the etiology of child maltreatment (Belsky, 1980) includes multiple and interacting risk factors, at the levels of the child, parent, family, community, and society. We utilized data gathered during infancy and toddlerhood from a large prospective study examining the antecedents and outcomes of maltreatment. We were guided by prior research that has linked multilevel, ecological risk factors to physical abuse and neglect; risk factors for these two problems often overlap (Coulton, Corbin, & Su, 1999). We prioritized risk factors that can somewhat readily be identified in child health care settings: Child. Children with disabilities, cognitive impairment, or failure to thrive may be at increased risk for maltreatment (Hibbard & Desch, 2007; Skuse, Gill, Reilly, Wolke, & Lynch, 1995). Parent. Multiple factors can impair parents’ abilities to care for their children. Maternal depression has been linked to neglect (Hien, Cohen, Caldeira, Flom, & Wasserman, 2010; Windham et al., 2004). Low maternal education (Kotch et al., 1995) and young maternal age (Zhou, Hallisey, & Freymann, 2006) are risk factors for maltreatment. Parental drug use is another concern (Connell, Bergeron, Katz, Saunders, & Tebes, 2007; Ondersma, 2002). One study found that children born to women who used cocaine during pregnancy were 6.5 times more likely to be maltreated than a matched comparison group (Leventhal et al., 1997). Chaffin and colleagues (1996) reported that half the maltreating parents in their sample had a history of substance abuse. Family factors can jeopardize children’s safety and health—directly and indirectly. Single parenthood (mostly mothers) and associated burdens are risk factors for physical abuse and neglect (Dubowitz, Hampton, Bithoney, & Newberger, 1987; Sedlak et al., 2010). A large number of children in the family, perhaps confounded by poverty, have also been linked to neglect (Sedlak et al., 2010; Zhou et al., 2006). Community. The social environment influences families and parenting. Low levels of involvement in formal and informal community agencies may be associated with maltreatment (Sidebotham, Heron, & Golding, 2002). Higher levels of social support are associated with less physical neglect, and more use of non-physical disciplinary methods (Lyons, Henly, & Schuerman, 2005; McCurdy, 2005). However, because many of these risk factors were identified at the time of maltreatment, often in samples reported to child protective services (CPSs), they are associated with maltreatment, but not necessarily causative or predictive. Risk factors identified through prospective longitudinal research may also not be causal and addressing them may not prevent child maltreatment (CM). Identifying risk factors prospectively, however, can help target those where further assessment and services may be beneficial. A strength of this prospective study is that the risk data were gathered before maltreatment was identified. Another strength is that the data on potential risk factors were gathered from multiple sources. Professionals working with high risk populations may have difficulty identifying families who are in special need of additional support and services. The study objective was to identify risk factors that can be readily assessed by child health care and potentially other professionals that should guide them in better serving families, and potentially help prevent physical abuse and neglect. Methods Sample Parents of 332 children agreed to participate and were recruited from 3 university-based pediatric clinics serving low income, urban families from 1989 through 1992. Eligibility criteria included age < 40 months, gestational age > 36 weeks, birth weight ≥ 2500 grams, and no congenital problems, disabilities, or chronic illnesses. One group consisted of children meeting criteria for non-organic failure-to-thrive (FTT) (n = 132), based on age and gender-specific National Center for Health Statistics (NCHS) growth charts: weight-for-age below the 5th percentile or weight-for-length below the 10th percentile. A pediatrician examined all children and reviewed their medical records; no specific medical causes were identified to explain their poor growth. All children enrolled with FTT were treated and followed in a subspecialty clinic; over 83% of eligible children were recruited (Black, Dubowitz, Hutcheson, Berenson-Howard, & Starr, 1995). We evaluated the children’s care giving environment to determine the likelihood of neglect, demonstrating FTT may occur with and without neglect (Mackner, Starr, & Black, 1997). A second group was identified in the perinatal period as being at risk for HIV mostly due to a history of maternal drug use (HIV Risk; n = 89). These children were followed in a special primary care clinic; over 80% of those who were eligible were recruited. The third group was recruited from a general primary care clinic and had no identified risk factors, other than being low income (Primary Care; n = 111). They were group matched for age, gender, race,

98

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

Table 1 Socio-demographic characteristics of participants at the time of recruitment (n = 224) and involvement with Child Protective Services (CPS). No CPS report (n = 127) Child Age (months), mean (SD) Male African American Failure to thrive Bayley Mental Development Index, mean (SD) Parent Age at child’s birth (years), mean (SD) Education* High school Employed Marital status* Married Separated/divorced/widowed Never married Ever used drugs No Yes Uncertaina Depressive symptoms, mean (SD)** Family Receiving AFDCb Receiving WICc Receiving Food Stamps** # of children in home, mean (SD)** # of adults in home, mean (SD) Social Support, mean (SD) a b c * **

CPS report (n = 97)

14.1 (6.8) 51% 92% 40% 99.2 (14.9)

14.0 (7.7) 55% 96% 43% 95.8 (14.8)

22.7 (5.6)

23.1 (5.3)

40% 43% 17% 20%

59% 34% 7% 13%

19% 9% 72%

6% 9% 85%

60% 15% 25% 0.2 (0.4)

47% 22% 31% 0.4 (0.6)

73% 58% 73% 2.4 (1.2) 2.4 (1.1) 21.0 (8.8)

80% 59% 88% 2.9 (1.3) 2.4 (1.1) 20.8 (10.2)

The status of those who did not complete questions pertaining to possible substance use was coded as “uncertain.” Aid to Families with Dependent Children. The Special Supplemental Nutrition Program for Women, Infants, and Children. p < .05. p < .01.

and socioeconomic status to children in the first 2 groups; approximately 90% of eligible families who were approached were recruited. Our approach in combining two high risk groups with a normative group of low-income children was guided by the observation that there is considerable overlap in the contributors to several problems concerning children, such as FTT, noninflicted injuries, ingestions and child maltreatment (Newberger, Hampton, Marx, & White, 1986). These authors suggested that it would be useful to recognize the commonalities of these symptoms as “pediatric social illnesses,” rather than viewing them as unrelated problems. The data analysis was designed to examine both commonalities and differences among the three risk groups. Seventy-three children who had been reported to CPS prior to recruitment were excluded leaving 259 children for the analyses; 224 had complete data (see Table 1). Approximately half of the children were male, most were African-American, and their average age was 14 months. The average age of mothers (including a few other primary caregivers; mothers used for brevity) was slightly less than 25 years, 48% had less than a high school education, few were employed, and 13% were married. Most families received Aid to Families with Dependent Children, the federal welfare program supporting children in low income families at the time. Fifty-nine percent participated in the Special Supplemental Nutrition Program for Women, Infants and Children (WIC), and 80% participated in the Food Stamp Program. The children and families were followed for a mean of 10 years.

Procedure Mothers agreed to participate in a longitudinal study, following consent procedures approved by our Institutional Review Board; including permission to review CPS records. The evaluation for mothers involved a 1-h interview using standardized questionnaires on demographic information and child and family functioning, administered by a research assistant. The evaluation for children included anthropometrics and a developmental assessment. Mothers were paid $25 for participating. CPS data were obtained at regular intervals in the ensuing years from the Maryland State Department of Human Resources.

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

99

Measures Independent variables Cognitive development. The Mental Development Index (MDI) of the Bayley Scales of Infant Development (Bayley, 1969) was used to assess cognitive development (M = 100, SD = 16). The Bayley has acceptable internal consistency, inter-rater reliability, and short-term test–retest reliability (Bayley, 1969). Reliability coefficients based on split-half correlations with this sample, ranged from .81 to .93 (median = .88) (Black et al., 1995). Parental depression. The Brief Symptom Inventory (Derogatis & Spencer, 1982) was administered to mothers who rated their feelings regarding 17 symptoms during the prior week on a 5-point Likert-type scale. We used the depression dimension for this paper; it has good internal consistency (alpha = .85). Test–retest reliability ranges from .68 to .91; construct validity is also good (Derogatis & Spencer, 1982). Maternal age, marital status and education, and number of children in the family were ascertained on a Demographic Questionnaire. Parental drug use. Mothers were asked about their drug use. The stem, “Have you ever used” was paired with a list of 10 illicit drugs. Participants also had the opportunity to list additional illegal drugs they had used. If a participant endorsed any substance, they were coded as having used drugs. Lack of social support. Social support was measured using the Family Support Scale (FSS), an 18-item scale measuring perceived support from formal and informal familial and non-familial (e.g., friends, church, child’s physician) sources in the past 3–6 months (Dunst, Jenkins, & Trivette, 1984). Mothers used a 6-point scale to rate the support that they received. Formal and informal supports scores were summed to provide a Total Support score. The FSS has adequate reliability and validity (Dunst et al., 1984). Dependent variable Child maltreatment report. The state CPS computerized system was examined approximately every 2 years for families’ possible involvement with the agency. All data until the study child was approximately 12 years old were combined. Time to the first CPS report in any given family was treated as the outcome variable using survival analytic methods. In these analyses, families with no CPS reports were treated as right censored at the time of last follow-up. For the three families whose child died during the follow-up period, the outcome was considered right censored at the date of the child’s death. The first (if any) CPS report pertaining to any child in the family was used as the outcome of interest. Many of the reports were for neglect; clinical and empirical evidence suggests that neglect generally involves all children in the household (Hines, Kantor, & Holt, 2006). We considered all reports, regardless of substantiation because evidence has shown that there are few differences between children and families based on substantiation (e.g., Hussey et al., 2005; Kohl, Jonson-Reid, & Drake, 2009). Data analysis Survival analysis methods were used to estimate the probability of a CPS report and to identify risk factors predicting a future CPS report. Survival analysis is useful when subjects are followed for different lengths of time as in our study. To estimate the probability that a child would have a CPS report over time we used the Kaplan–Meier approach. To identify factors that put children at higher risk of a future CPS report, we fit Cox proportional hazards regression models. To develop the final model predicting a CPS report, we first conducted bivariate analyses, separately examining each independent variable in a different Cox regression. The final model, based on the 224 families with complete data, was fit using the backward stepwise variable selection method using Wald statistics (p-values for inclusion and exclusion were .05 and .10, respectively). We used stepwise regression to derive a more parsimonious model to identify the key predictors, with more precise estimates of the regression coefficients than the full model. Standardized Z-scores were used for continuous variables. Statistical Software SPSS (SPSS Inc., Version 17.0, Chicago, IL) was used for the analyses. Results Ninety-seven (43%) of the families were reported to CPS at least once. Sixty-five percent of the 97 CPS reports were for neglect, 27% for physical abuse, and 8% for sexual abuse. In comparing the 3 risk groups (FTT, HIV Risk, Primary Care) combined for the current analysis, no differences were found regarding child’s gender and ethnicity, the number of children and adults in the household, mother’s education, employment status, and marital status, depression level at study intake, the Bayley MDI, and the receipt of Food Stamps. The following differences were found: children in the Primary Care group were older than children in both the FTT and HIV Risk groups, mothers in the HIV Risk group were older at their child’s birth than mothers in the Primary Care group, and mothers in the FTT group reported less social support than mothers in the Primary Care group. As expected, children in the FTT group were also more likely to have received WIC than the other two risk groups, and HIV Risk group mothers reported significantly more drug use than the FTT and Primary Care group mothers, both of which were similar.

100

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

Table 2 Effect of potential predictors on risk of having a CPS report based on multivariable Cox regression models (n = 224a ).

Full model

Final model

a b c

Variable

B

SE

Relative risk (95% CI)

Wald

Age Bayley Mental Development Indexb Failure to thrive Mother’s age at child’s birth (years) Education, ≤high school Employed Never married Maternal drug usec Ever used drug Uncertain Depression Score (per SD) Number of children Social Support score (per SD) Chi-square = 33.401, df = 12, p < .001 Bayley Mental Development Index b Education,≤high school Maternal drug usec Ever used drug Uncertain Depression Score (per SD) Number of children Chi-square = 30.331, df = 6, p < .001

<.01 −.21 .05 .12 −.46 .09 .54

.13 .10 .27 .11 .23 .32 .30

1.00 (0.78–1.28) 0.81 (0.66–0.99) 1.05 (0.62–1.80) 1.13 (0.91–1.40) 1.58 (1.02–2.46) 1.09 (0.58–2.05) 1.72 (0.96–3.10)

<.00 4.23 0.04 1.23 4.12 0.08 3.30

.99 .040 .85 .27 .042 .78 .07

p

.52 −.04 .25 .22 .01

.29 .30 .09 .08 .12

1.70 (0.96–2.97) 0.96 (0.54–1.72) 1.30 (1.08–1.54) 1.25 (1.06–1.50) 1.02 (0.80–1.28)

3.28 0.02 8.01 6.95 0.02

.07 .89 .005 .008 .90

−.21 −.44

.10 .22

1.23 (1.01–1.49) 1.55 (1.01–2.38)

4.42 4.04

.035 .044

.54 −.03 .25 .23

.27 .25 .08 .08

1.71 (1.01–2.90) 0.97 (0.60–1.57) 1.28 (1.09–1.51) 1.26 (1.07–1.47)

4.01 0.02 8.81 7.98

.045 .90 .003 .005

Due to missing data the sample was reduced from 259 to 224. Per 1 SD difference, higher scores optimal. Reference category is ‘Never used drug.’

Five sociodemographic characteristics differed significantly between families with and without CPS reports (Table 1). Mothers with CPS reports were less educated and less likely to be married. Families with reports were more likely to receive Food Stamps and to have more children in their homes as well having caregivers with more depressive symptoms. In addition to confirming the above relationships, the bivariate Cox analyses also showed that a history of drug use was associated with later CPS involvement. Children reported to CPS had a lower Bayley MDI, though this did not reach statistical significance (p = .06). When examined together in a multivariate backward stepwise regression, five variables predicted a CPS report (Table 2). A lower Bayley MDI score predicted an increased likelihood of a CPS report (p < .05). A one standard deviation improvement in the Bayley MDI was associated with a 19% lower likelihood of a CPS report. Children whose mothers did not finish high school were 1.55 times more likely to be reported. In addition, children whose mothers ever used drugs were 1.7 times more likely to be reported to CPS compared to those whose mothers never used drugs. More depressive symptoms were also predictive of maltreatment. Specifically, as the depressive symptom score increased by one standard deviation, the risk of being reported increased by 28%. Each additional child in the home increased the risk of being reported to CPS by 26%. The overall model was significant (2 = 30.3, df = 6, p < .01). Of note, FTT was not predictive of later CPS involvement. This model assumes a multiplicative relationship between predictors and the risk of a CPS report. Therefore, the risk ratio due to having multiple risk factors equals the product of the risk ratios for the individual risk factors. For example, for children whose mothers did not complete high school and used drugs, the risk of a CPS report is increased by an estimated factor of 1.55 times 1.71 which equals 2.65. Fig. 1 shows the probability that a child will remain CPS-report free over time, in various subgroups defined by key predictors. Note that among those whose parent ever used drugs, by 48 months of follow-up only approximately 60% will have remained CPS-report free. Alternatively, 40% in that group will have been reported to CPS within 48 months of followup. In contrast, in families where the parents did not use drugs, only approximately 25% will have had a CPS report. These figures also illustrate that the risk of a CPS report is higher if the parent did not finish high school, had more than 2 children, or was in the highest quartile based on the BSI depression score. Discussion This prospective, longitudinal study demonstrated that child maltreatment was a prevalent problem in this low-income sample of urban, mostly African American families recruited from pediatric primary care clinics and followed for approximately 10 years; 43% were reported to CPS for possible maltreatment. Consistent with the multilevel, ecological theory underlying child maltreatment (Belsky, 1980) our survival analysis found that within this at-risk population, maltreatment could be predicted by risk factors at the child, parent, and family levels—identified in first few years of life. This study adds to the literature by identifying risk factors early in life, prior to the maltreatment report to CPS. This is the kind of evidence we need to guide our efforts rather than relying on correlational data obtained from families already involved in the child welfare system. Although statistical prediction does not prove causality; in this study, we conclude that risk

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

101

Fig. 1. Probability of remaining CPS-report free, over time since start of follow-up, in various subgroups defined by risk factors.

factors identified early in life were associated with later CPS involvement. Despite the entire sample being at risk, resources seldom permit intervening with everyone. Thus, it can help to identify those at highest risk and their specific needs. Ideally, such a strategy would complement universal policies and programs, such as those that combat poverty and help support families. The only child level variable that was predictive of a maltreatment report was a low score on a standardized assessment of mental development during the first 3.5 years of life. Although others have reported that children with developmental problems are at increased risk for maltreatment (Hibbard & Desch, 2007), the children in this sample were not developmentally delayed. Their mental scores were below average, consistent with other samples of low-income children (Black, Dubowitz, Krishnakumar, & Starr, 2007). Children with more normative development may pose fewer challenges to their parents, making them less vulnerable to maltreatment. Children with age-appropriate development may also be more likely to be in environments that foster healthy early child development. Professionals working with young children are in a position to assess their development. In addition to identifying children in need of early intervention, assessing children’s development may help identify families’ psychosocial needs and help prevent potential maltreatment. Early FTT was not predictive of a maltreatment report, perhaps because all children with FTT were treated in a special interdisciplinary clinic. By age 6, most children’s growth had recovered (Black et al., 2007). When FTT co-occurred with

102

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

neglect or maltreatment, the consequences for children were more severe when both were present, compared to either condition alone (Kerr, Black, & Krishnakumar, 2000; Mackner et al., 1997). Several parental factors were predictive of a maltreatment report. Low maternal education has been shown to increase the likelihood of maltreatment (Kotch et al., 1995). Without a high school education, mothers may be limited in providing adequate care and protection for their children. They may lack competence in managing child rearing demands, particularly when caring for multiple children (Bugental & Johnston, 2000). Finally, poorly educated caregivers may not have access to financial or community resources to assist with child rearing. Several studies have reported that children of mothers with depressive symptoms are at risk for maltreatment (Hien et al., 2010; Windham et al., 2004). Depression can influence mood, affect, and behavior, disrupting the sensitive motherchild relationship necessary for children’s development (Lovejoy, Graczyk, O’Hara, & Neuman, 2000). Depressed mothers have been described as more aggressive toward their children and less attentive to childrearing needs than non-depressed mothers. Children of depressed mothers are at risk for behavioral and developmental problems, potentially increasing their childrearing demands (Carter, Garrity-Rokous, Chazan-Cohen, Little, & Briggs-Gowan, 2001). Treating maternal depression can result in benefits to children (Weissman et al., 2006) and addressing maternal depression may help prevent child maltreatment. The finding that maternal substance use increases the likelihood of a maltreatment report is consistent with other reports (Chaffin, Kelleher, & Hollenberg, 1996; Connell et al., 2007; Leventhal et al., 1997; Ondersma, 2002). Substance use often involves a lifestyle committed to acquiring and using drugs. Thus, substance using women may be “unavailable” to their children. Low education, maternal depression, and drug use often co-occur, contributing to an environment with few resources for children. When the care giving situation is further complicated by a child’s developmental problems as well as multiple children in the family, the risks for maltreatment increase. The 3rd National Incidence Study on Child Abuse and Neglect found that families with more than four children were at increased risk for maltreatment (Sedlak & Broadhurst, 1996).

Limitations There are limitations to be considered in interpreting the findings. First, we prioritized variables that are reasonably accessible to health care professionals working with young children and their families For example, parental depression can be screened easily in primary care (Dubowitz et al., 2007), whereas parental IQ is difficult to assess. Second, the study did not include all possible predictors of maltreatment, such as intimate partner violence; we were constrained by available data. Third, the sample consists of 3 groups of primarily low-income, urban, African American families. Although the groups are unique, they share many common problems facing minority families in poverty, living in dangerous, depleted, urban neighborhoods. Generalizability is limited to similar populations. Fourth, the sample did not include children whose families did not seek primary care or children with disabilities, other likely risk factors. Fifth, we relied on CPS reports as indicators of maltreatment. However CPS reports are an imperfect proxy for maltreatment; there may be erroneous reports or maltreated children who were not reported. There may also be professional bias with minority families at increased risk of being reported (Hampton & Newberger, 1985). However, 93% of the children were African-American, reducing the possibility of professional reporting bias. Finally, it is possible that families with 1 or more of the risk factors were more likely to be reported to CPS, presenting a tautological problem.

Implications and future directions Child health care and other professionals face the quandary of identifying families at high risk for maltreatment, who may benefit from additional support and services. This prospective study offers valuable guidance with five risk factors that can be readily identified or assessed via brief screens (Dubowitz et al., 2007; Lane et al., 2007) and useful brief screens of children’s development are available (Brothers, Glascoe, & Robertshaw, 2008). Identification of risk is only the first step. Comprehensive assessment, generally done by other community agencies, should clarify families’ needs. At the same time, identifying for example a probable substance abuse problem helps professionals link parents with specific resources to address that problem. In addition to potentially helping prevent child maltreatment, helping address risk factors such as maternal depression or drug use should strengthen families and enhance parent–child interactions and children’s development. The findings should not be interpreted as enabling clinicians to clearly predict future maltreatment; rather, they suggest how certain risk factors can be used to identify those who could most likely benefit from further assessment and help. In addition, advocacy to mitigate systemic social problems such as poverty and violence is also much needed to truly tackle the problem of child maltreatment. Future research is necessary to identify useful risk and protective factors, practical screening approaches, and effective strategies to help children and families obtain needed services to help prevent child maltreatment.

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

103

References Bayley, N. (1969). Manual for the Bayley Scales of Infant Development. New York, NY: The Psychological Corporation. Belsky, J. (1980). Child maltreatment: An ecological integration. American Psychologist, 35, 320. Black, M. M., Dubowitz, H., Hutcheson, J., Berenson-Howard, J., & Starr, R. J. (1995). A randomized clinical trial of home intervention for children with failure to thrive. Pediatrics, 95, 807–814. Black, M. M., Dubowitz, H., Krishnakumar, A., & Starr, R. H. (2007). Early intervention and recovery among children with failure to thrive: Follow-up at age 8. Pediatrics, 120(1), 59–69. Brothers, K. B., Glascoe, F. P., & Robertshaw, N. S. (2008). PEDS: Developmental milestones—an accurate brief tool for surveillance and screening. Clinical Pediatrics, 47(3), 271–279. Bugental, D. B., & Johnston, C. (2000). Parental and child cognitions in the context of the family. Annual Review Psychology, 51, 315–344. Carter, A. S., Garrity-Rokous, F. E., Chazan-Cohen, R., Little, C., & Briggs-Gowan, M. J. (2001). Maternal depression and comorbidity: Predicting early parenting, attachment security, and toddler social–emotional problems and competencies. Journal American Academy Child Adolescent Psychiatry, 40(1), 18–26. Chaffin, M., Kelleher, K., & Hollenberg, J. (1996). Onset of physical abuse and neglect: Psychiatric, substance abuse and social risk factors from prospective community data. Child Abuse & Neglect, 20, 191–200. Connell, C. M., Bergeron, N., Katz, K. H., Saunders, L., & Tebes, J. K. (2007). Re-referral to child protective services: The influence of child, family, and case characteristics on risk status. Child Abuse & Neglect, 31, 573–588. Coulton, C. J., Corbin, J. E., & Su, M. (1999). Neighborhoods and child maltreatment: A multi-level study. Child Abuse & Neglect, 23(11), 1019–1040. Derogatis, L. R., & Spencer, R. (1982). The Brief Symptom Inventory (BSI): Administration, scoring, and procedure manual. Baltimore: Clinical Psychometric Research. Dubowitz, H. (2002). Preventing child neglect and physical abuse: A role for pediatricians. Pediatrics in Review, 23(6), 191–196. Dubowitz, H., Feigelman, S., Lane, W., Prescott, L., Blackman, K., Grube, L., Meyer, W., & Tracy, J. K. (2007). Screening for depression in an urban pediatric primary care clinic. Pediatrics., 119, 435–443. Dubowitz, H., & Guterman, N. (2005). Prevention of physical abuse and child neglect. In A. Giardino, & R. Alexander (Eds.), Child maltreatment: A clinical guide and reference. St. Louis, MO: GW Medical, Publications. Dubowitz, H., Hampton, R., Bithoney, W. G., & Newberger, E. H. (1987). Inflicted and non-inflicted injuries: Differences in child and family characteristics. American Journal of Orthopsychiatry, 57(4), 525–535. Dunst, C. J., Jenkins, V., & Trivette, C. M. (1984). The Family Support Scale: Reliability and validity. Journal of Individual, Family & Community Wellness, 1, 45–52. Gaudin, J. M. (1999). Child neglect: Short-term and long-term outcomes. In H. Dubowitz (Ed.), Neglected children: Research, practice and policy (3rd ed., pp. 89–108). Thousand Oaks, CA: Sage Publications. Green, M. (1994). Bright futures: Guidelines for health supervision of infants, children, and adolescents. Arlington, VA: National Center for Education and Maternal and Child Health. Hampton, R. L., & Newberger, E. H. (1985). Child abuse incidence and reporting by hospitals: Significance of severity, class and race. American Journal of Public Health, 75, 56–60. Hibbard, R. A., & Desch, L. W. (2007). American Academy of Pediatrics Committee on Child Abuse and Neglect; American Academy of Pediatrics Council on Children With Disabilities. Maltreatment of children with disabilities. Pediatrics, 119(5), 1018–1025. Hien, D., Cohen, L. R., Caldeira, N. A., Flom, P., & Wasserman, G. (2010). Depression and anger as risk factors underlying the relationship between maternal substance involvement and child abuse potential. Child Abuse & Neglect, 34(2), 105–113. Hines, D. A., Kantor, G. K., & Holt, M. K. (2006). Similarities in siblings’ experiences of neglectful parenting behaviors. Child Abuse & Neglect, 30(6), 619–637. Hoekelman, R. A. (1997). Child health supervision. In R. A. Hoekelman, S. B. Friedman, N. M. Nelson, H. M. Seidel, & M. L. Weitzman (Eds.), Primary pediatric care. St. Louis, MO: Mosby. Hussey, J. M., Marshall, J. M., English, D. J., Knight, E. D., Lau, A. S., Dubowitz, H., & Kotch, J. B. (2005). Defining maltreatment according to substantiation: Distinction without a difference? Child Abuse & Neglect, 29(5), 479–492. Kerr, M., Black, M. M., & Krishnakumar, A. (2000). Failure-to-thrive, maltreatment and the behavior and development of 6-year-old children from low-income urban families: A cumulative risk model. Child Abuse & Neglect, 24(2000), 587–598. Kohl, P. L., Jonson-Reid, M., & Drake, B. (2009). Time to leave substantiation behind: Findings from a national probability study. Child Maltreatment, 14(1), 17–26. Kotch, J. B., Browne, D. C., Ringwalt, C. L., Stewart, P. W., Ruina, E., Holt, K., Lowman, B., & Jung, J. W. (1995). Risk of child abuse or neglect in a cohort of low-income children. Child Abuse & Neglect, 19(9), 1115–1130. Lane, W. G., Dubowitz, H., Feigelman, S., Kim, J., Prescott, L., Meyer, W., & Tracy, J. K. (2007). Screening for parental substance abuse in an urban pediatric primary care clinic. Ambulatory Pediatrics, 7(6), 458–462. Leventhal, J. M., Forsyth, B. W., Qi, K., Johnson, L., Schroeder, D., & Votto, N. (1997). Maltreatment of children born to women who used cocaine during pregnancy: A population-based study. Pediatrics, 100(2), E7. Lovejoy, M. C., Graczyk, P. A., O’Hara, E., & Neuman, G. (2000). Maternal depression and parenting behavior: A meta-analytic review. Clinical Psychology Review, 20(5), 561–592. Lyons, S. J., Henly, J. R., & Schuerman, J. R. (2005). Informal support in maltreating families: Its effect on parenting practices. Children Youth Services Review, (1), 21–38. Mackner, L. M., Starr, R. H., & Black, M. M. (1997). The cumulative effect of neglect and failure to thrive on cognitive functioning. Child Abuse & Neglect, 21, 691–700. Macmillan, H. L., Wathen, C. N., Barlow, J., Fergusson, D. M., Leventhal, J. M., & Taussig, H. N. (2009). Interventions to prevent child maltreatment and associated impairment. Lancet, 17, 373(9659), 250–266. McCurdy, K. (2005). The influence of support and stress on maternal attitudes. Child Abuse & Neglect, 29, 251–268. Newberger, E. H., Hampton, R. L., Marx, T. J., & White, K. M. (1986). Child abuse and pediatric social illness: An epidemiological analysis and ecological reformulation. American Journal of Orthopsychiatry, 56(4), 589–601. Ondersma, S. J. (2002). Predictors of neglect within low socioeconomic status families: The importance of substance abuse. America Journal of Orthopsychiatry, 72, 383–391. Pilowsky, D. J., Wickramaratne, P., Talati, A., Tang, M., Hughes, C. W., Garber, J., Malloy, E., King, C., Cerda, G., Sood, A. B., Alpert, J. E., Trivedi, M. H., Fava, M., Rush, A. J., Wisniewski, S., & Weissman, M. M. (2008). Children of depressed mothers 1 year after the initiation of maternal treatment: findings from the STAR*D-Child Study. America Journal of Psychiatry, 165(9), 1136–1147. Sedlak, A. J., & Broadhurst, D. D. (1996). The Third National Incidence Study of Child Abuse and Neglect (NIS-3). Washington, DC: U.S. Government Printing Office. Sedlak, A. J., Mettenburg, J., Basena, M., Petta, I., McPherson, K., Greene, A., & Li, S. (2010). Fourth National Incidence Study of Child Abuse and Neglect (NIS-4): Report to congress. Washington, DC: US Department of Health and Human Services, Administration for Children and Families. Sidebotham, P., Heron, J., & Golding, J. (2002). Child maltreatment in the “Children of the Nineties.” Deprivation, social class, and social networks in a UK sample. Child Abuse & Neglect, 26, 1243–1259. Skuse, D. H., Gill, D., Reilly, S., Wolke, D., & Lynch, M. A. (1995). Failure to thrive and the risk of child abuse: A prospective population survey. Journal of Medical Screening, 2(3), 145–149.

104

H. Dubowitz et al. / Child Abuse & Neglect 35 (2011) 96–104

Teicher, M. H., Dumont, N. L., Ito, Y., Vaituzis, C., Giedd, J. N., & Andersen, S. L. (2004). Childhood neglect is associated with reduced corpus callosum area. Biological Psychiatry, 56, 80–85. US Department of Health and Human Services, Administration on Children Youth and Families. (2010). Child maltreatment—2008. Washington, DC: US Government Printing Office. http://www.acf.hhs.gov/programs/cb/pubs/cm08/summary.htm. Weissman, M. M., Pilowsky, D. J., Wickramaratne, P. J., Talati, A., Wisniewski, S. R., Fava, M., Hughes, C. W., Garber, J., Malloy, E., King, C. A., Cerda, G., Sood, A. B., Alpert, J. E., Trivedi, M. H., Rush, A. J., & STAR*D-Child Team. (2006). Remission in maternal depression and child psychopathology. Journal of the American Medical Association, 295(12), 1389–1398. Windham, A. M., Rosenberg, L., Fuddy, L., McFarlane, E., Sia, C., & Duggan, A. K. (2004). Risk of mother-reported child abuse in the first 3 years of life. Child Abuse & Neglect, 28(6), 645–667. Zhou, Y., Hallisey, E. J., & Freymann, G. R. (2006). Identifying perinatal risk factors for infant maltreatment: An ecological approach. International Journal of Health Geography, 5, 53.