Impact of Variance in Child Maltreatment Exposure to Mental Health Utilization and Subsequent Medicaid Expenditures
Method: We linked data from a national sample of children with histories of abuse and neglect drawn from the National Survey of Child and Adolescent Well-Being (NSCAW) to their Medicaid claims obtained from 36 states, and adjusted all expenditures to 2010 dollars. We examined three different operationalizations of abuse and neglect: (1) indicator variables for whether or not a child suffered a given type of abuse, (2) a count of numbers of types of abuse or neglect reported by a caseworker, and (3) a measure of severity of abuse also as reported by a caseworker. We estimated two-part regression models (logistic followed by generalized linear models) to examine Medicaid expenditures on psychotropic drugs conditional upon the child incurring any expenditures on psychotropic drugs, and controlled for demographic characteristics, mental health need, and urbanicity of residence in all models.
Results: (1) By type of abuse, only physical and sexual abuse had significantly higher Medicaid expenditures on psychotropic drugs. For physical abuse, odds of using psychotropic medication were 1.48 times higher, but conditional on receipt, the level of use was not statistically different, and the net psychotropic cost impact was $60 greater per year vs. non-abused. For sexual abuse, odds were 1.73 higher and net impact was $135 greater per year. (2) Children with 3 or more types of abuse had 3.75 times the odds of psychotropic medication use for a net impact of $260 more in psychotropic expenditures per year relative to non-abused children. There were no net impacts on costs for children with 1 or 2 types of abuse relative to non-abused children. (3) Severity measures both quantity and magnitude of abuse and was categorized as mild, moderate or severe maltreatment. Mild and moderate severity did not incur significantly increased expenditures, but severely maltreated children incurred $178 more expenditures than children with no maltreatment and had odds of receipt that were 1.90 times greater.
Conclusions and Implications: Medicaid agencies should be aware of the much greater likelihood of psychotropic medication for children experiencing quantitatively and qualitatively different types of maltreatment. Such approaches to quantify risk of service can help Medicaid agencies better plan for the service needs of maltreated child beneficiaries within the program.