Abstract
Objectives. We sought to examine the extent to which children in the child welfare system receive mental health care consistent with national standards.
Methods. We used data from 4 waves (3 years of follow-up) of the National Survey of Child and Adolescent Well-Being, the nation's first longitudinal study of children in the child welfare system, and the Area Resource File to examine rates of screening, assessment, and referral to mental health services among 3802 youths presenting to child welfare agencies. Weighted population-averaged logistic regression models were used to identify variables associated with standards-consistent care.
Results. Only half of all children in the sample received care consistent with any 1 national standard, and less than one tenth received care consistent with all of them. Older children, those exhibiting externalizing behaviors, and those placed in foster care had, on average, higher odds of receiving care consistent with national standards.
Conclusions. Adverse consequences of childhood disadvantage cannot be reduced unless greater collaboration occurs between child welfare and mental health agencies. Current changes to Medicaid regulations that weaken entitlements to screening and assessment may also worsen mental health disparities among these vulnerable children.
Ensuring that children in the child welfare system receive high-quality mental health services has been a policy priority for the past 2 decades, and professional organizations have proposed standards of care specifically tailored for this population. However, there is no information on the extent to which children in the child welfare system actually receive care that is consistent with these national standards.
Attention to standards-consistent mental health care is important because children in the child welfare system have very high needs for mental health services. Nearly half of all children coming into contact with child welfare agencies nationally have clinically significant emotional or behavioral problems,1 and 40% to 60% of children in the child welfare system may have a psychiatric diagnosis.2 Local and regional studies have also documented high rates of problems across several domains among these children.3–5 Children in foster care have between 10 and 20 times the rate of utilization of mental health services when compared with children maintained within their homes,6–8 and national estimates suggest that the point prevalence of psychotropic medication use is 13.5%9—far above the 4% seen among privately insured children10 and the 5% to 6% seen among Medicaid-enrolled children.11,12
Given this magnitude of service use and the unique vulnerabilities of children in child welfare, the Child Welfare League of America proposed standards for the provision of health and mental health services for children in foster care in 1988.13 The American Academy of Pediatrics proposed similar standards in 1994 and in 2002,14,15 as did the American Academy of Child and Adolescent Psychiatry in 2001.16,17 These recommendations are highly convergent with regard to mental health care and suggest that all children should receive a mental health screening when placed into foster care, a subsequent comprehensive mental health assessment by a mental health professional within a month of being placed into foster care, and a systematic, coordinated approach to the delivery of services to meet children's ongoing mental health needs.
These standards since have been operationalized for use by child welfare agencies and extended to children who are not placed in foster care. For example, the Council on Accreditation, presumably informed by the literature documenting significant mental health needs among non–foster-care child welfare populations,1 evaluates child protective services agencies on their case-management abilities, including needs assessment and the provision of counseling and intensive mental health services.18 These standards govern processes of care—not quality of clinical care—and are primarily designed to assist child welfare agencies in implementing organizational structures and processes that can improve child well-being. Passage of the 1997 Adoption and Safe Families Act19–21 provided additional federal impetus to standards development, and states are currently evaluated on the extent to which children “receive adequate services to meet their … mental health needs.”22
Despite professional consensus and federal legislation and regulation, there is little information on whether children in the child welfare system actually receive care that is consistent with these standards. In 2002, Halfon et al. reported on the overall inadequacy of services available to children in foster care23; a key informant survey of state and county mental health agencies found that only one third of all mental health agency respondents reported being aware of these standards24; and another study found that fewer than half of all counties had policies regarding mental health assessments for children in child welfare.25 However, there was no information obtained at the child level to determine whether this lack of knowledge and policymaking adversely affects the care received by individual children.
In an attempt to quantify the extent to which children in the child welfare system receive mental health services consistent with national standards, we analyzed data from the first national study of children coming into contact with child welfare agencies—the National Survey of Child and Adolescent Well-Being (NSCAW). We conducted longitudinal data analyses to estimate the prevalence of standards-consistent care among this national sample and identify child-, caregiver-, and caseworker-level characteristics that place children at most risk for receiving care inconsistent with these national standards. Through these analyses we attempt to provide policy guidance for child welfare and child mental health policymakers to better construct systems of care for these highly vulnerable children.
METHODS
NSCAW is the first national probability study of children investigated by their local child protective services for possible abuse and neglect in 92 primary sampling units in 97 counties throughout the United States.26 We used data from NSCAW's wave 1 (1999–2001), wave 2 (2000–2002), wave 3 (2001–2002), and wave 4 (2002–2004) for these longitudinal analyses.
We deleted from our sample all children younger than 2 years at wave 1, because the version of the Child Behavior Checklist (CBCL),27 the measure of need for mental health services used in NSCAW, is not normed for that age group. Our final sample contained 3802 children at wave 1, with information obtained from available caregivers (birth parent if the child is being maintained within his or her own home, or foster parent if the child is placed outside of his or her own home) and caseworkers. Except where otherwise noted, all variables were obtained from caregiver report.
Standards-Consistent Mental Health Services
We examined attainment of these standards, beginning with wave 2, as operationalized for child welfare agencies in prior literature.24 We did not examine developmental or cognitive services among these youths. The measure for assessment was a dichotomous variable elicited from the caseworker asking whether “a formal assessment … for an emotional, behavioral, or attention problem” was done by a “psychologist or a medical professional.” This assessment could either have been an initial assessment, or one performed after a change in placement.
The measure for a referral to mental health services was elicited from caseworkers using a series of questions asking whether “agency staff [referred the child] to counseling for an emotional, behavioral, or attention problem.”
The weakest index was a dichotomous variable for screening elicited from the caseworker in response to a question about “major special needs or behavior problems … at the time of investigation” with subsequent questions following up on the type of problem. Because of validity concerns with the screening variable, we constructed a binary variable of services consistent with at least 1 standard and used this (admittedly liberal) indicator of standards-consistent care in regression analyses.
Recognizing the potential for upward bias on caseworker-obtained indices, we triangulated this information from caregivers' reports. These data were available for the referral variable; all discrepancies were coded in favor of the caregiver.
Covariates
Choice of covariates was based on the behavioral model of access to health services,28 as applied to mental health services in the literature.29 Child-level covariates included child age, gender, and race/ethnicity. Dichotomous emotional and behavioral need variables were based on the identification of a probable behavioral disorder if the child scored in the clinical range (t score ≥ 64) on the internalizing or externalizing scales of the CBCL,27 which is a well-established measure of childhood behavior problems among child welfare populations.30,31
Maltreatment history was obtained from the child's caseworker and was based on a modified Maltreatment Classification Scale.32 Categories of physical abuse, sexual abuse, neglect, and abandonment were dichotomized such that a child could have more than 1 type of abuse coded. For multivariate analyses, we aggregated maltreatment into positive abuse (physical or sexual abuse) and negative abuse (neglect or abandonment) in accordance with prior literature that suggested increased rates of service use (and possibly higher-quality care) for children exposed to positive maltreatment.31
Each child's placement status was grouped into 3 mutually exclusive categories: in home (i.e., living with their permanent primary caregiver, usually their birth parent), in family foster care (either with a relative or nonrelative), and in congregate care (i.e., in a group home or residential treatment center). In multivariate analysis, we grouped children in family foster care and those in congregate care into a single category. We categorized children into mutually exclusive insurance categories of Medicaid, private insurance, federal (Civilian Health and Medical Program of the Uniformed Services [CHAMPUS]), and uninsured to control for differences in access. We then constructed a binary variable indicating whether the child was insured at each wave. Information on whether the child lived in an urban or rural area was obtained from NSCAW data as an additional control for the availability of health care resources in the child's community.
Caregivers' ability to exercise agency roles on behalf of their child was assessed using proxies of highest level of education, current employment status, and annual family income from all sources. Caseworker variables included highest educational level (undergraduate, master's, or doctorate) of the caseworker and number of months of experience in the child welfare system, both obtained from the child's caseworker. We assumed that highly educated and experienced caseworkers may be better able to secure standards-consistent care for their dependents.
County health resources were derived from the Area Resource File,33 which is a collection of health-related data from over 50 sources. We divided the number of adult psychiatrists and pediatricians by the county population to obtain population-adjusted mental health–provider ratios, as described elsewhere.29,34 These ratios served to adjust provider supply across counties that display variations in amount of available provider resources.
Analyses
All analyses were weighted to account for varying probabilities of selection. We lagged the mental health standards variables such that each outcome was regressed on predictors obtained at the prior wave.
NSCAW data revealed decreasing sample sizes from 3802 (wave 1) to 3280 (wave 4), but none of our covariates were predictive of missingness between waves 1 through 4. Given the lack of relationship between observed variables and missingness, we imputed missing variables using the method of imputation by chained equations (ice).35,36 Ice is a multiple imputation technique that uses a series of switching regressions to impute missing values.37 Five data sets were created, each data set was reshaped from “wide” to “long” format, weights were applied, and logistic regression models estimated in each of these data sets. Unstandardized parameter estimates from each data set were averaged into a single coefficient, and their standard errors were adjusted for both between- and within-imputation variance.
We estimated weighted population-averaged logistic regression models, which allow testing of effects of predictors on attainment of national standards over time. Population-averaged models are a type of longitudinal data analysis model that provides mean odds of attainment of national standards for the entire sample. The child observation was the unit of analysis. All models included a year-of-interview variable to control for secular trends (data not shown). We conducted model sensitivity analyses by estimating unweighted random effects models, unimputed weighted models, and full-information maximum likelihood models in Mplus.38 Concerns with bias in all these models led us to use the weighted population-averaged logistic regression models for our final analyses.
In these models, time-varying covariates included age, CBCL scores (both internalizing and externalizing), placement, and insurance status. To assess partial effects of caseworker and health resources variables, we introduced these variables in staged models. Because imputed models cannot be formally tested for fit, we used unimputed (complete case) models to test for model fit and to refine our specification (data not shown).
All bivariate analyses were performed using a weighted χ2 test for categorical variables and a weighted paired t test for continuous variables. All bivariate and regression analyses were performed in STATA version 9 (StataCorp, College Station, TX).39
RESULTS
Children in our sample (N = 3802) had a mean age of 8.1 years (SD = 3.8 years) and were equally divided between the genders. Most were White (47.5%), followed by African American (27.8%), and Hispanic (17.8%); all percentages were weighted, as measured at baseline wave. One third (31.2%) reportedly displayed emotional distress as measured by a score in the clinical range on the CBCL's externalizing scale, whereas 20.7% had a clinical score on the CBCL's internalizing scale. Virtually all children in the sample had a history of abuse or neglect, which included physical abuse (36.0%), sexual abuse (13.8%), moderate neglect (55.4%), and abandonment (21.6%). Most (89.9%) were resident in their own homes. The majority of children were covered by some type of insurance, including private insurance, Medicaid, or another state-funded program. Most children resided in urban areas (76.2%).
Caregivers of many children had a high school or equivalent education (43.3%), and 28.3% reported having attended college. Many caregivers were employed full-time (46.4%), and 38.4% reported not being formally employed. Less than half of all caregivers reported family incomes of less than $15 000 per year (41.3%), and one third (32.3%) reported family incomes between $15 000 and $30 000.
Most (74.9%) caseworkers of these children had an undergraduate degree, and 25.1% had master's degree or higher. The lengths of experience of these caseworkers ranged from 1 month to 36 years, with a mean of 6.8 years (SD = 6.5 years). Counties had an average of 1.1 adult psychiatrists and 1.52 pediatricians per 10 000 residents.
Rates of Receipt of Services Consistent With National Standards
Increasing percentages of children received standards-consistent care with increasing duration of follow-up (Figure 1). Percentages of children screened showed modest increases from 41.8% at wave 2 to 48.3% at wave 4 (all percentages weighted). Rates of assessment increased from 33.2% to 34.5%, and rates of referral also increased from 36.8% to 38.3%, in the same time interval. Across all waves of NSCAW, only around half (54.6%) of all children received services consistent with at least 1 national standard; less than one tenth (9.8%) received care consistent with all these standards.
FIGURE 1.
Percentage of children receiving services consistent with national standards, by wave: National Survey of Child and Adolescent Well-Being, 1999–2004.
Note. The percentages were weighted using complete case analyses at each wave. Wave 1 = 1999–2001, wave 2 = 2000–2002, wave 3 = 2001–2002, and wave 4 = 2002–2004.
Characteristics of Children Receiving Standards-Consistent Services
Table 1 shows bivariate relationships between child, caregiver, and caseworker characteristics and receipt of the standards of initial screening for mental health needs, comprehensive assessment (both initial as well as at the time of placement), and referral to mental health services (all percentages are weighted row percentages). Age, mental health need as measured by the CBCL, abuse history, and placement status were significantly associated with receipt of all of the standards. A greater proportion of children in older age cohorts, children with CBCL scores in the clinical range (indicating probable behavioral disorder), and those in foster care, residential care, or some other congregate care environment displayed receipt of care consistent with national standards. A greater percentage of abused children received standards-consistent care when compared with children without a history of abuse.
TABLE 1.
Characteristics of Child Welfare Children Meeting National Standards in Waves 1 Through 4: National Survey of Child and Adolescent Well-Being, 1999–2004
| Characteristics | Screening, No. (%) | Assessment, No. (%) | Referral, No. (%) | Any Standard, No. (%) |
| Total | 1458 (100) | 1348 (100) | 1346 (100) | 1680 (100) |
| Child level | ||||
| Age, y | ||||
| 2–5 | 283 (37.6)* | 286 (29.6) | 251 (23.8)*** | 360 (40.1)** |
| 6–11 | 706 (47.2)* | 640 (40.3) | 664 (37.7)*** | 812 (51.4)** |
| 12–16 | 469 (54.7)* | 422 (44.0) | 431 (53.3)*** | 508 (61.5)** |
| Gender | ||||
| Boy | 734 (49.3) | 676 (40.2) | 635 (34.2) | 825 (52.3) |
| Girl | 724 (43.2) | 672 (36.2) | 711 (40.7) | 855 (48.9) |
| Race/Ethnicity | ||||
| White | 682 (49.0) | 617 (40.1) | 620 (38.4) | 773 (53.0) |
| African American | 413 (38.7) | 398 (33.7) | 381 (33.2) | 481 (43.2) |
| Hispanic | 235 (49.7) | 221 (39.1) | 231 (42.8) | 281 (54.8) |
| Other | 125 (50.7) | 109 (43.2) | 111 (34.8) | 142 (54.8) |
| CBCL externalizing score | ||||
| ≥ 64 | 663 (62.7)*** | 603 (50.6)*** | 592 (53.1)*** | 715 (67.0)*** |
| < 64 | 771 (37.1)*** | 721 (31.0)*** | 732 (28.9)*** | 941 (41.8)*** |
| CBCL internalizing score | ||||
| ≥ 64 | 514 (59.8)*** | 467 (49.9)*** | 462 (46.7)* | 549 (60.7)* |
| < 64 | 919 (42.3)*** | 856 (34.4)*** | 861 (34.5)* | 1106 (47.6)* |
| Type of abuse | ||||
| Physical | 466 (53.8)* | 432 (46.2)* | 437 (44.4)* | 526 (56.2) |
| Sexual | 302 (44.6)* | 261 (33.8)* | 311 (48.2)* | 350 (54.4) |
| Neglect | 178 (44.4)* | 172 (45.1)* | 157 (34.6)* | 202 (48.1) |
| Abandonment | 90 (46.2)* | 86 (43.8)* | 77 (37.9)* | 101 (49.7) |
| Placement | ||||
| Out of home | 624 (72.5)*** | 620 (69.7)*** | 610 (72)*** | 701 (79.6)*** |
| In home | 834 (41.2)*** | 728 (32.6)*** | 736 (31.5)*** | 979 (45.6)*** |
| Medical insurance status | ||||
| Uninsured | 100 (41.1) | 79 (19.8)*** | 94 (29.6) | 118 (41.9) |
| Insured | 1326 (46.8) | 1239 (40.3)*** | 1222 (38.2) | 1528 (51.5) |
| Urbanicity | ||||
| Urban | 1162 (48.2) | 1069 (38.4) | 1082 (39.3) | 1341 (52.6) |
| Rural | 296 (38.5) | 279 (37.4) | 264 (30.3) | 339 (42.6) |
| Caregiver level | ||||
| Education | ||||
| Less than high school | 342 (43.7) | 305 (33-0) | 312 (36.7) | 398 (49.4) |
| High school or equivalent | 595 (47.8) | 553 (40.4) | 552 (36.1) | 697 (52.5) |
| Some college or more | 457 (47.3) | 427 (40.7) | 424 (40.5) | 515 (49.8) |
| Employment | ||||
| Not employed | 571 (43.6) | 556 (38.0) | 517 (32.5) | 674 (48.5) |
| Full time | 609 (46.9) | 553 (36.2) | 576 (38.5) | 693 (50.1) |
| Part time | 200 (51.6) | 161 (41.6) | 181 (46.6) | 226 (56.4) |
| Family income, $ | ||||
| < 15 000 | 426 (40.9) | 396 (34.4) | 392 (35.2) | 511 (47.5) |
| 15 000–29 999 | 378 (45.3) | 332 (34.5) | 346 (32.3) | 442 (48.7) |
| ≥ 30 000 | 458 (52.4) | 421 (43.9) | 436 (45.1) | 513 (54.7) |
| Caseworker level | ||||
| Education, degree | ||||
| Undergraduate | 996 (46.6) | 905 (36.2) | 908 (37.4) | 1137 (49.7) |
| Master's or higher | 354 (44.3) | 343 (44.0) | 340 (37.9) | 418 (51.3) |
| Length of experience, y | ||||
| < 5 | 670 (40.4)* | 630 (34.9) | 620 (36.9) | 771 (45.9) |
| 5–9 | 383 (52.4)* | 346 (39.2) | 350 (36.9) | 445 (54.7) |
| ≥ 10 | 295 (49.8)* | 271 (44.4) | 277 (39.6) | 337 (53.1) |
Note. CBCL = Child Behavior Checklist. All predictors are at wave 1. Outcomes were defined as 1 if the child met the standard at any 1 of the 4 waves. Numbers are unweighted; percentages are weighted using complete case analyses.
*P < .05; **P < .01; ***P < .001.
A significantly lower proportion of uninsured children received comprehensive assessments when compared with insured children (19.8% versus 40.3%; P < .001), and children whose caseworkers had less than 5 years of experience had lower rates of screening (40.4%) compared with children with caseworkers with 10 years or more experience in the child welfare system (49.8%; P = .04). There were no significant mean differences in the health resources ratio variables between children who met, and did not meet, at least 1 national standard (data not shown).
On multivariate analysis (Table 2), older age was still associated with receipt of care consistent with at least 1 national standard over time, with the odds of receiving such care increasing by 10% on average with each year's increase in age (odds ratios [ORs] ranging from 1.09 to 1.1, all significant at P = <.001). Children with scores in the clinical range on the externalizing subscale of the CBCL had over twice the odds of receiving such care (ORs between 2.17 and 2.18, all P < .001). Children placed out of home had 3 times the odds of receiving care consistent with at least 1 national standard (ORs between 3.15 and 3.16; P < .001). Results for age and placement were also seen in a model estimated only with children with clinically significant CBCL scores (data not shown).
TABLE 2.
Effects of Child-, Caregiver-, and Policy-Level Characteristics and Area Resources on Meeting Any National Standard: National Survey of Child and Adolescent Well-Being, 1999–2004
| Model 1, OR (95% CI) | Model 2, OR (95% CI) | Model 3, OR (95% CI) | |
| Child level | |||
| Age, y | 1.09*** (1.05, 1.14) | 1.10*** (1.05, 1.14) | 1.10*** (1.05, 1.14) |
| Gender | |||
| Boy | 1.18 (0.92, 1.52) | 1.19 (0.90, 1.56) | 1.19 (0.91, 1.56) |
| Girl (Ref) | 1.00 | 1.00 | 1.00 |
| Race/Ethnicity | |||
| White (Ref) | 1.00 | 1.00 | 1.00 |
| African American | 0.86 (0.59, 1.26) | 0.87 (0.59, 1.30) | 0.89 (0.62, 1.28) |
| Hispanic | 1.03 (0.64, 1.66) | 1.06 (0.64, 1.78) | 1.07 (0.64, 1.78) |
| Other | 1.00 (0.51, 1.97) | 1.00 (0.50, 1.98) | 1.00 (0.50, 1.98) |
| CBCL externalizing score | |||
| ≥ 64 | 2.18** (1.38, 3.46) | 2.17** (1.30, 3.61) | 2.18** (1.32, 3.61) |
| < 64 (Ref) | 1.00 | 1.00 | 1.00 |
| CBCL internalizing score | |||
| ≥ 64 | 1.34 (0.82, 2.21) | 1.35 (0.78, 2.31) | 1.34 (0.78, 2.31) |
| < 64 (Ref) | 1.00 | 1.00 | 1.00 |
| Experienced physical or sexual abuse | |||
| Yes | 1.52 (0.91, 2.53) | 1.51 (0.90, 2.52) | 1.50 (0.87, 2.58) |
| No (Ref) | 1.00 | 1.00 | 1.00 |
| Experienced neglect or abandonment | |||
| Yes | 0.91 (0.63, 1.30) | 0.92 (0.64, 1.34) | 0.91 (0.63, 1.32) |
| No (Ref) | 1.00 | 1.00 | 1.00 |
| Placement | |||
| Out of home | 3.15*** (2.01, 4.92) | 3.15*** (2.00, 4.96) | 3.16*** (2.00, 4.96) |
| In home (Ref) | 1.00 | 1.00 | 1.00 |
| Medical insurance status | |||
| Uninsured | 0.67 (0.42, 1.07) | 0.67 (0.42, 1.05) | 0.67 (0.42, 1.05) |
| Insured (Ref) | 1.00 | 1.00 | 1.00 |
| Urbanicity | |||
| Urban | 0.83 (0.54, 1.26) | 0.80 (0.52, 1.24) | 0.83 (0.48, 1.45) |
| Rural (Ref) | 1.00 | 1.00 | 1.00 |
| Caregiver level | |||
| Education | |||
| Less than high school (Ref) | 1.00 | 1.00 | |
| High school or equivalent | 1.15 (0.77, 1.72) | 1.15 (0.76, 1.72) | |
| Some college or more | 1.14 (0.63, 2.06) | 1.14 (0.63, 2.06) | |
| Employment | |||
| Not employed | 0.94 (0.62, 1.43) | 0.94 (0.62, 1.43) | |
| Full time | 0.89 (0.55, 1.44) | 0.89 (0.55, 1.44) | |
| Part time (Ref) | 1.00 | 1.00 | |
| Caseworker level | |||
| Education, degree | |||
| Undergraduate (Ref) | 1.00 | 1.00 | |
| Master's or higher | 1.06 (0.69, 1.64) | 1.04 (0.69, 1.64) | |
| Length of experience, mo | 1.00 (1.00, 1.00) | 1.00 (1.00, 1.00) | |
| Area resources | |||
| Adult psychiatrist ratio | 0.94 (0.66, 1.37) | ||
| Pediatrician ratio (Ref) | 1.00 | ||
Note. CI = confidence interval; CBCL = Child Behavior Checklist; OR = odds ratio. 3802 children were observed a total of 11 406 times across all waves. Models include a year of interview variable to control for secular trends (data not shown).
*P < .05; **P < .01; ***P < .001.
DISCUSSION
To our knowledge, we describe for the first time the extent to which a national sample of children presenting to child welfare agencies receives mental health services consistent with national standards. These preliminary findings indicate that only half of all children receive care consistent with at least 1 index of these standards; fewer than 1 in 10 receives care consistent with all indices; and older children, those exhibiting externalizing behaviors, and those placed outside their homes of origin are more likely to receive care consistent with these standards.
The observed associations between placement and need and standards-consistent care suggest that child-serving systems are complying with the original intent of these standards to guide services for children in foster care and are targeting their limited resources toward children who present with the greatest caretaking challenges. It is also likely that children in foster care have greater access to a larger number of individuals who can advocate for and leverage needed services.
Given the paucity of Medicaid and child welfare budgets, this kind of triage may actually be the most cost-effective allocation of resources if the children who gain the greatest marginal benefit are the ones receiving these services. However, treating placement as a proxy for need causes other children to become underserved. Focusing on children with externalizing disorders ignores the mental health needs of children who present with internalizing symptoms,40 and focusing on children placed out-of-home ignores the considerable mental health needs of those maintained within their own homes.41 To the extent that child welfare and mental health agencies implement standards in a targeted, as opposed to universal, fashion, they are likely to underserve many children within their care.
That a considerable proportion of children in child welfare are, in fact, being underserved is evident from the fact that less than two thirds of children with clinically significant CBCL scores receive care consistent with any 1 national standard (Table 1). Overall, only a little over one third of all children are being assessed, either at their initial entry into child welfare or during a transition in placement. This low overall rate of assessment, coupled with a focus on children in foster care, may diminish the rate of referral and, as a consequence, children with mental health needs are likely being underidentified and undertreated by child-serving agencies nationwide. Such underidentification and undertreatment may hinder efforts to reduce the adverse consequences of childhood disadvantage and may worsen behavioral health inequalities among these children.
The reasons for such standards-inconsistent care are not readily apparent from this study—they do not, however, seem to lie in the nonavailability of appropriate health care professionals within the child's county. Instead, it is likely that they lie within the systems and processes that child welfare and mental health agencies construct (or fail to construct) as they attempt to serve the mental health needs of children in the child welfare system. For example, it is known that care coordination between child welfare and mental health agencies promotes appropriate mental health care,42 yet only one third of mental health agencies report being aware of quality standards for children in child welfare.25 Greater coordination between child welfare and mental health agencies specifically directed toward improving the ability of caseworkers to identify and refer internalizing children, and assuring the availability and adequacy of comprehensive assessments for these children may enhance the overall appropriateness of care available to these children.
Current policymaking also endangers the ability of child welfare systems to provide standards-consistent care. Children in the child welfare system are largely dependent upon Medicaid,29 and Medicaid's Early and Periodic Screening, Diagnosis, and Treatment program establishes an entitlement to assessments and treatment that could be used to pay for care consistent with national standards. Unfortunately, federal policymaking over the past several decades has diminished the effectiveness of the program, and passage of the Deficit Reduction Act of 200543 (PL 109–171) has replaced Medicaid's defined benefits with industry-derived “benchmarks.”44,45 Because these benchmarks may not be consistent with Medicaid's own coverage, careful consideration of proposed and ongoing changes to Medicaid financing is necessary to ensure that financing assists rather than hinders the ability of child welfare systems to provide services consistent with national standards.
These policy recommendations are tempered by a few limitations. First, the lack of an exact match between NSCAW variables and national standards diminished our ability to present estimates of each of the indices. Hence, these findings cannot be construed as definitive and should be interpreted as the best indication of how well current practice is adhering to these national standards. Second, the potential for upward bias of indices elicited from caseworkers exists, in which case fewer children than reported in this study receive standards-consistent care. Third, 10% of children were placed outside their homes of origin, and the ratings on the CBCL by their caregivers may have been influenced by cultural issues or by insufficient familiarity with the child. Fourth, current software does not permit the full incorporation of stratification corrections when conducting weighted, panel data analyses on binary outcomes. Hence, the true prevalence of standards-consistent care is likely to be lower than suggested by our estimates. Finally, the national standards we use are not practice parameters directed toward clinicians, and so these findings cannot be used to establish estimates of the quality of clinical care received by these children.
Despite these limitations, these data indicate that large numbers of children in contact with child welfare nationwide may be receiving care inconsistent with national standards. Expanding screening and assessment protocols to children maintained within their own homes, ensuring adequacy of financing to child welfare agencies, and enhancing the ability of caseworkers to identify and refer children with internalizing as well as externalizing disorders are necessary to assure the well-being of children in the child welfare system.
Acknowledgements
This study was funded by the National Institute of Mental Health (grant R03 MH082117). R. Raghavan is partly supported by the National Institute of Mental Health (award 5P30 MH068579).
The National Survey on Child and Adolescent Well-Being was developed under contract with the Administration on Children, Youth, and Families, US Department of Health and Human Services (ACYF/DHHS). The data have been provided by the National Data Archive on Child Abuse and Neglect.
The authors thank Peichang Shi for assistance with statistical analyses.
Note. The information and opinions expressed herein reflect solely the position of the authors. Nothing herein should be construed to indicate the support or endorsement of its content by ACYF/DHHS.
Human Participant Protection
These analyses were approved by the Washington University Human Research Protection Office.
References
- 1.Burns BJ, Phillips SD, Wagner HR, et al. Mental health need and access to mental health services by youths involved with child welfare: a national survey. J Am Acad Child Adolesc Psychiatry 2004;43(8):960–970 [DOI] [PubMed] [Google Scholar]
- 2.Landsverk J, Garland AF, Leslie LK. Mental health services for children reported to Child Protective Services. : Myers JEB, Berliner L, Briere J, Hendrix CT, Jenny C, Reid TA, APSAC Handbook on Child Maltreatment Thousand Oaks, CA: Sage; 2002 [Google Scholar]
- 3.Leslie LK, Gordon JN, Meneken L, Premji K, Michelmore KL, Ganger W. The physical, developmental, and mental health needs of young children in child welfare by initial placement type. J Dev Behav Pediatr 2005;26(3):177–185 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Urquiza AJ, Wirtz SJ, Peterson MS, Singer VA. Screening and evaluating abused and neglected children entering protective custody. Child Welfare 1994;73(2):155–171 [PubMed] [Google Scholar]
- 5.Milburn NL, Lynch M, Jackson J. Early identification of mental health needs for children in care: a therapeutic assessment programme for statutory clients of child protection. Clin Child Psychol Psychiatry 2008;13(1):31–47 [DOI] [PubMed] [Google Scholar]
- 6.Halfon N, Berkowitz G, Klee L. Mental health service utilization by children in foster care in California. Pediatrics 1992;89(6 Pt 2):1238–1244 [PubMed] [Google Scholar]
- 7.Halfon N, Berkowitz G, Klee L. Children in foster care in California: an examination of Medicaid reimbursed health services utilization. Pediatrics 1992;89(6 pt 2):1230–1237 [PubMed] [Google Scholar]
- 8.Harman JS, Childs GE, Kelleher KJ. Mental health care utilization and expenditures by children in foster care. Arch Pediatr Adolesc Med 2000;154(11):1114–1117 [DOI] [PubMed] [Google Scholar]
- 9.Raghavan R, Zima BT, Andersen RM, Leibowitz AA, Schuster MA, Landsverk J. Psychotropic medication use in a national probability sample of children in the child welfare system. J Child Adolesc Psychopharmacol 2005;15(1):97–106 [DOI] [PubMed] [Google Scholar]
- 10.Stein BD, Sturm R, Kapur K, Ringel J. Datapoints: psychotropic medication costs among youth with private insurance in 1998. Psychiatr Serv 2001;52(2):152. [DOI] [PubMed] [Google Scholar]
- 11.Martin A, Van Hoof T, Stubbe D, Sherwin T, Scahill L. Multiple psychotropic pharmacotherapy among child and adolescent enrollees in Connecticut Medicaid managed care. Psychiatr Serv 2003;54(1):72–77 [DOI] [PubMed] [Google Scholar]
- 12.Zito JM, Safer DJ, DosReis S, et al. Psychotropic practice patterns for youth: a 10-year perspective. Arch Pediatr Adolesc Med 2003;157(1):17–25 [DOI] [PubMed] [Google Scholar]
- 13.Child Welfare League of America Standards for Health Care Service for Children in Out-of-Home Care Washington, DC: Child Welfare League of America; 1988 [Google Scholar]
- 14.American Academy of Pediatrics Committee on Early Childhood Adoption and Dependent Care Health care of young children in foster care. Pediatrics 1994;93:335–338 [PubMed] [Google Scholar]
- 15.American Academy of Pediatrics Committee on Early Childhood Adoption and Dependent Care Health care of young children in foster care. Pediatrics 2002;109:536–54111875156 [Google Scholar]
- 16.American Academy of Child and Adolescent Psychiatry Psychiatric Care of Children in the Foster Care System Washington, DC: American Academy of Child and Adolescent Psychiatry; 2001 [Google Scholar]
- 17.American Academy of Child and Adolescent Psychiatry AACAP/CWLA Policy Statement on Mental Health and Use of Alcohol and Other Drugs, Screening and Assessment of Children in Foster Care Washington, DC: American Academy of Child and Adolescent Psychiatry; 2003 [Google Scholar]
- 18.Council on Accreditation Accreditation Standards—Public Agencies Eighth Edition. Available at: http://www.coastandards.org/standards.php?navView=public. Accessed August 29, 2008
- 19.Baker DR, Rauber DB. Making Sense of the ASFA Regulations: A Roadmap for Effective Implementation Washington, DC: American Bar Association, National Child Welfare Resource Center on Legal and Judicial Issues, Center on Children and the Law; 2001 [Google Scholar]
- 20.Wulczyn F. Beyond Common Sense: Child Welfare, Child Well-Being, and the Evidence for Policy Reform New Brunswick, NJ: AldineTransaction; 2005 [Google Scholar]
- 21. Adoption and Safe Families Act PL 105-89. 111 S 2115, amending 42 USC §671–675, 1997. [Google Scholar]
- 22.Administration for Children and Families, US Department of Health and Human Services. Procedures Manual Available at: http://www.acf.hhs.gov/programs/cb/cwmonitoring/tools_guide/proce_manual.htm. Accessed August 21, 2006
- 23.Halfon N, Inkelas M, Flint R, Shoaf K, Zepeda A, Franke T. Assessment of Factors Influencing the Adequacy of Health Care Services to Children in Foster Care Los Angeles: University of California, Los Angeles, Center for Healthier Children, Families, and Communities National Center for Infancy and Early Childhood Health Policy; 2002 [Google Scholar]
- 24.Raghavan R, Inkelas M, Franke T, Halfon N. Administrative barriers to the adoption of high-quality mental health services for children in foster care: a national study. Adm Policy Ment Health 2007;34(3):191–201 [DOI] [PubMed] [Google Scholar]
- 25.Leslie LK, Hurlburt MS, Landsverk J, Rolls JA, Wood PA, Kelleher KJ. Comprehensive assessments for children entering foster care: a national perspective. Pediatrics 2003;112(1 pt 1):134–142 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.NSCAW Research Group Methodological lessons from the National Survey of Child and Adolescent Well-Being: the first three years of the USA's first national probability study of children and families investigated for abuse and neglect. Child Youth Serv Rev 2002;24(6–7):513–541 [Google Scholar]
- 27.Achenbach TM. Manual for the Child Behavior Checklist/4-18 and 1991 Profile Burlington: University of Vermont Department of Psychiatry; 1991 [Google Scholar]
- 28.Andersen RM, Davidson PL. Improving access to care in America: individual and contexutal indicators. : Andersen RM, Rice TH, Kominski GF, Changing the US Health Care System San Francisco, CA: Jossey-Bass; 2001:3–30 [Google Scholar]
- 29.Raghavan R, Leibowitz AA, Andersen RM, Zima BT, Schuster MA, Landsverk J. Effects of Medicaid managed care policies on mental health service use among a national probability sample of children in the child welfare system. Child Youth Serv Rev 2006;28(12):1482–1496 [Google Scholar]
- 30.Zima BT, Bussing R, Yang X, Belin TR. Help-seeking steps and service use for children in foster care. J Behav Health Serv Res 2000;27(3):271–285 [DOI] [PubMed] [Google Scholar]
- 31.Raghavan R, Aarons GA, Roesch SC, Leslie LK. Longitudinal patterns of health insurance coverage among a national sample of children in the child welfare system. Am J Public Health 2008;98(3):478–484 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Manly JT, Cicchetti D, Barnett D. The impact of subtype, frequency, chronicity, and severity of child maltreatment on social competence and behavior problems. Dev Psychopathol 1994;6:121–143 [Google Scholar]
- 33.US Department of Health and Human Services Area Resource File. Available at: http://www.arfsys.com. Accessed July 20, 2006
- 34.Raghavan R, Leibowitz AA. Medicaid and mental health care for children in the child welfare system. : Haskins R, Wulczyn F, Webb MB, Practical Knowledge for Child Welfare Practitioners: Findings from the National Survey of Child and Adolescent Well-being Washington, DC: Brookings Institution; 2007 [Google Scholar]
- 35.Royston P. Multiple imputation of missing values: update. Stata J 2005;5(2):188–201 [Google Scholar]
- 36.Royston P. Multiple imputation of missing values: Update of ice. Stata J 2005;5(4):527–536 [Google Scholar]
- 37.van Buuren S, Boshuizen HC, Knook DL. Multiple imputation of missing blood pressure covariates in survival analysis. Stat Med 1999;18(6):681–694 [DOI] [PubMed] [Google Scholar]
- 38.Muthén LK, Muthén BO. Mplus User's Guide 4th ed.Los Angeles, CA: Muthén & Muthén; 2007 [Google Scholar]
- 39.STATA Statistical Software, Release 9 [computer program]. College Station, TX: StataCorp; 2005 [Google Scholar]
- 40.Birmaher B, Brent D, Bernet W, et al. Practice parameter for the assessment and treatment of children and adolescents with depressive disorders. J Am Acad Child Adolesc Psychiatry 2007;46(11):1503–1526 [DOI] [PubMed] [Google Scholar]
- 41.Farmer EM, Burns BJ, Chapman MV, Phillips SD, Angold A, Costello EJ. Use of mental health services by youth in contact with social services. Soc Serv Rev 2001;75:605–624 [Google Scholar]
- 42.Hurlburt MS, Leslie LK, Landsverk J, et al. Contextual predictors of mental health service use among children open to child welfare. Arch Gen Psychiatry 2004;61(12):1217–1224 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Deficit Reduction Act of 2005, 42 USC §1305 (2006)
- 44.Rosenbaum S, Wise PH. Crossing the Medicaid-private insurance divide: the case of EPSDT. Health Aff (Millwood) 2007;26(2):382–393 [DOI] [PubMed] [Google Scholar]
- 45.Rubin DM, Halfon N, Raghavan R, Rosenbaum S, Johnson K. The Deficit Reduction Act of 2005: Implications for Children Receiving Child Welfare Services Seattle, WA: Casey Family Programs; 2006 [Google Scholar]