Measuring Area-Socioeconomic Disparities in Mental Health Service Use Among Children in the Child Welfare System

Background: It is well documented that there is unequal availability of and access to mental health (MH) services among children in the child welfare system (CWS). Of various dimensions of disparity, county-level poverty has been considered a source of generating geographic disparities in MH service use, rendering questions about whether poverty contributes to racial/ethnic disparities in MH service use. However, little empirical research has been conducted, at a national level, to measure area-socioeconomic disparities in the CWS. In addition, studies that rely on disparity indices for ordinal social groups determined by county-level poverty rates are lacking. Thus, this study analyzed different approaches to the measurement of MH service disparities, measuring area-socioeconomic disparities in MH service use among a nationally representative sample of youth referred to the CWS.

Methods: Data were drawn from the National Survey of Child and Adolescent Well-being (NSCAW), a nationally representative sample of children referred to the CWS between October 1999 and December 2000 (Cohort 1) and between February 2008 and April 2009 (Cohort II). The current study focused on 1,005 Cohort 1 and 727 Cohort 2 children (aged 12 and 17), with data collected at Wave 3 for Cohort 1 and Wave 2 for Cohort 2. By then, both cohorts had been referred to the CWS for18 months. Using HD*Calc, statistical software designed to evaluate social disparities, we analyzed changes in area socio-economic disparities between Cohort 1 and 2 relying upon eight different disparity measures.  Three area-socioeconomic groups were derived as follows: counties with a child poverty rate of >20%, 15-20%, and <15%. Subgroup analyses were also conducted to examine whether area-socioeconomic disparities were more pervasive among youth in need for MH services, as determined by the Child Behavior Checklist.

Results: All eight measures of disparity showed that area-socioeconomic disparities in MH service use increased over time; however, the size of the increase differed considerably across disparity measures. Among relative disparity measures, for instance, Rate Ratio increased by a 38%, while Theil index and Mean Log Deviation indicated a 560% increase, respectively. Across absolute disparity measures, the size of the increase also varied, ranging from a 116 percent increase in Absolute Concentration Index to a 670% increase in Between-Group Variance. Our subgroup analyses revealed that youth in need for MH services, experienced even greater increases in area-socioeconomic disparities.

Conclusions: This study suggests that the same underlying data can yield different magnitudes of disparity, depending on which measure is used and whether youth’ need for MH service is considered. Given this substantial inconsistency among disparity measures, a greater understanding of the methodological differences among disparity measures and sound justification for selection of disparity measures are warranted among researchers and policy makers, which will in turn impact how interventions are designed to reduce disparities among children in the CWS.