The Effect of an Expanded Medical Home Model for Children in Foster Care
Methods: The study uses linked administrative data from the Illinois Integrated Database from FY 2002-09. The study populations were limited to children who were continuously eligible for Medicaid over a 2 year period. The health services of children in foster care (all enrolled in the Healthworks program, n=45,562 child-years) are compared with children with histories of abuse and neglect without foster care placement (n=12,980 child-years) and all other Illinois children eligible for Medicaid (n=3,354,517 child-years). The analysis focused on the following outcomes: well-child visits, ER visits, and hospitalizations. Logistic regression and propensity score matching was used to compare the health outcomes of foster children with the comparison groups.
Results: Children in foster care enrolled in Healthworks had 3.3 times the odds of well-child visits (p<.001) compared with abused and neglected children not in foster care and 3.6 times the odds compared with other Medicaid eligible children (p<.001). In relation to both comparison groups, there was a significant increase in well-child utilization over time (p<.001). There were no significant differences in ER service use between children in foster care and the two comparison groups; however, there was evidence of increasing odds of ER visits over time for children in foster care versus other Medicaid eligible children (p<.05). With respect to general inpatient hospitalization, there were no significant differences in utilization between children in foster care and abused and neglected children not in foster care, but some evidence of increased odds of hospitalization for children in foster care compared with other Medicaid eligible children at baseline (FY 2003, p<.05).
Conclusions and Implications:
Consistent with the goals of the medical home model provided through Healthworks, the study finds that children in foster care are more likely to have well-child visits and that there exists an increasing trend over time. In contrast, there was no evidence that ER service utilization or hospitalization were less likely or decreasing over time among foster children. This may be due in part to higher rates of chronic conditions in the foster care group not completed ameliorated with the propensity score methods. Future analysis needs to explore the condition specific triggers of ER visits and hospitalization to identify additional ways that primary care providers can limit the use of such services.