Abstract: Examining Contextual Factors to Understand Residential Care Utilization (Society for Social Work and Research 14th Annual Conference: Social Work Research: A WORLD OF POSSIBILITIES)

49P Examining Contextual Factors to Understand Residential Care Utilization

Friday, January 15, 2010
* noted as presenting author
Sigrid James, PhD , Loma Linda University, Associate Professor, Loma Linda, CA
John Landsverk, PhD , Rady Hospital San Diego, Child and Adolescent Services Research Center, Director, San Diego, CA
Jin Jin Zhang, MS , Child and Adolescent Services Research Center, Data Analyst, San Diego, CA

Placement into residential care (RC) is considered an adverse outcome due to its cost, restrictiveness and weak evidence base. Yet despite explicit efforts by service systems to reduce RC utilization, placement into RC remains a common occurrence. This study uses two national data sets – the National Survey of Child and Adolescent Well-Being (NSCAW) and its companion study Caring for Children in Child Welfare (CCCW) – to investigate the impact of contextual variables on RC utilization among youth in the CW system (CWS).


Data. - The NSCAW study is the first national longitudinal study to examine the characteristics, needs, and outcomes of children referred to CWS. NSCAW used a stratified two-stage cluster sampling strategy (n=5,501). - The CCCW study was funded as a supplement study to NSCAW, using the same sampling frame. The study collected detailed contextual data at state/local levels on CW policies, the organization of MH care, etc., and when linked to NSCAW data, can estimate the number of children affected by different kinds of policies and organizational structures.

Study sample. – This analysis includes children, two and older, who experienced any stays in out-of-home care during the 4 study waves. Of the 981 children in the study sample, 22.8% experienced entries into RC.

Analysis. - To examine the effect of contextual variables on RC utilization (defined as number of days in RC), negative binomial regression is used. This technique is appropriate for count data, and when the poisson distribution is overdispersed. Variables entered into the model included were: demographics (gender, age, race/ethnicity), clinical factors (behavioral problems – CBCL<64; developmental problems - ‘delayed' if score 2 SD< the mean on at least one standardized measure in any of 3 developmental domains; family risk - score based on caseworker report); placement/service-related (maltreatment – modified Maltreatment Classification Scale; prior out-of-home placement; number of placements; length of stay in non-RC placements; onset of specialty outpatient MH services); system/contextual (urbaniticty, county poverty level; insurance type; mental health linkages between CW and MH - count of 26 indicators of linkages between the two local agencies; mental health provider supply - Area Resource File variables were used to examine variation in provider supply). Results

Results indicate that after controlling for salient sociodemographic, clinical and placement-related variables, MH linkages accounted significantly for greater RC utilization. Counties with a high degree of linkage had the lowest rate of RC utilization. Other factors associated with RC utilization were male gender, older age, a higher rate of behavior problems, a higher number of placements as well as longer stays in other types of placements.


RC is a form of service delivery whose use is affected by multiple factors. Yet few studies to date have empirically examined the relationship between organizational factors and RC utilization. Findings of this analysis support the efforts of service systems to create explicit linkages between CW and MH through steps such as improved mental health screening and assessment and provision of psychosocial interventions. Existing barriers to such linkage are discussed.