Methods: A literature review was conducted with a clear inclusion criteria developed in accordance with PRISMA guidelines. A two-step inclusion/exclusion process was used to eliminate unrelated articles and create the final sample. The quality of each article was assessed as recommended by the Cochrane reviewers handbook. A random-effects meta-analysis was conducted for weighted means estimate of incidence rates and correlations associated with medication. Furthermore, an analysis of publication bias was conducted.
Results: The database search returned 374 entries published between 1990 and 2016. After deleting duplicates and excluding articles that failed the inclusion criteria, 40 studies were extracted for analysis. Most studies were retrospective cross-sectional studies and Medicaid claims were a common data source. Most had a moderate assessment of quality according EPHPP ratings with a mean score of 1.77 (SD=0.58) out of 3.
From the 23 studies that reported rates of psychotropic medication for children in foster care, the overall rate of psychotropic medication was 32.9% (95% CI=.29-.27) with a range of 10.3%-82.0%. Meta-analysis of correlates of psychotropic medication for children in foster care was conducted when there were at least 3 studies contributing to the analysis. This analysis found that foster children who are male, older, White, in fair or poor health, living in out-of-home placements, and those with clinical CBCL scores for externalizing and internalizing behavior were more likely to receive psychotropic medication. Conversely, foster children with a history of neglect, Black, and in Medicaid systems that offers primary care case management had lower odds of receiving psychotropic medication. Publication bias analysis revealed mixed results.
Conclusions and Implications: This study confirms that the foster children receive psychotropic medication at much higher rates than the general public. Also, there is evidence that either not all foster children are at risk of psychotropic medication treatment equally or there may be a disparity in treatment options. As abuse history had differential effects on medication, abuse history is relevant information for mental health clinicians as they treat children in the child welfare system. Most studies used a cross-sectional design with limited measures to reduce bias due to confounding variables. The focus on correlational results also prohibits conclusions of causality. The field would greatly benefit from more experimental or quasi-experimental studies in this area.