The Prevalence of Childhood Disability in a National Child Welfare Sample: A Latent Profile Analysis

Purpose:  It is well established that children who are maltreated are more likely to have physical, developmental, and emotional impairments.  However, it is still unclear how many children served by child welfare have a disability. Previous research attempting to ascertain this prevalence rate has primarily relied on dichotomous measures of disability, usually with poor measures of disability.  However, disability is a multifaceted phenomenon that cannot be reduced to a bifurcated “disabled vs. not disabled” categorization.  The goal of this study is to take a rigorous quantitative approach to classifying disability in child welfare.  We identified disability subtypes using latent profile analysis (LPA) of national data of children in families investigated by Child Protective Services (CPS).   

Methods:  This study used the 2010 National Survey of Child and Adolescent Well-Being (NSCAW II), a national probability study of 5,873 children who were subjects of child abuse and neglect investigations.  LPA were conducted on two groups: those still residing in-home following the investigation (n=1702) and those living in out-of-home (n=852).  All measures of disability were standardized and continuous, with higher scores indicating better functioning.  Measures included the verbal and nonverbal sections of the Kaufman Brief Intelligence test; the Social Skills Rating Scale; the daily-living questionnaire of the Vineland Adaptive Behavioral Scale; and the internalizing and externalizing scales of the Child Behavioral Checklist.  LPA was used to identify smaller homogenous subgroups (i.e. latent classes) of disabling conditions based on the above set of 6 continuous variables. 

Results:  A three-class model was chosen as the final solution.  Class 1 was the smallest subgroup identified (13.0% in-home and 8.8% in out-of-home).  This class, labeled cognitively impaired, had a mean score on both verbal and nonverbal intelligence of more than three standard deviations below the population mean.  A greater number of children residing in-home had a cognitive disability compared to children living out-of-home.  Children in this class also had a delay (i.e. more than one standard deviation below the mean) in daily-living functionality in both in-home and out-of-home placements.  Class 2 included slightly more than one-third of children in-home and children in out-of-home care (36.9% and 39.4%) and is labeled behavioral problems.  Children in this class were between 1 and 1.5 standard deviations below the mean for both internalizing and externalizing behaviors irrespective of placements.  Class 3 was the largest class and represented slightly over half of the total population in both in-home (53.3%) and out-of-home (51.8%) and is labeled typically developing.  This class had cognitive, social, functional, and behavioral scores at or slightly below the population mean. 

Conclusion: Our findings from LPA show that when using standardized measures of disability, nearly half of children investigated by the child welfare system have a disability. There are considerably more children with cognitive impairments than found in previous studies, and on average, these children have significant impairments, and fewer with behavioral problems. Understanding the prevalence and severity of disabling conditions among children can help the child welfare system better modify their services to serve this large proportion of clients.