Effects of the Colorado Medicaid Accountable Care Collaborative on Psychotropic Polypharmacy for Foster Children
High rates of psychotropic medication use (e.g., antipsychotics, antidepressants, mood stabilizer, stimulants, and antianxiety medications) among youths in foster care is a major national concern, which has led to intense scrutiny about its appropriateness in this vulnerable population. Rubin and colleagues showed a doubling of psychotropic polypharmacy associated with foster care status among Medicaid enrolled children (2009). Domino, Olfson, and Zito have highlighted increasing rates of antipsychotic use among those without diagnosed indications, and have especially pointed to poly-pharmacy within and outside the same therapeutic classes.
A natural experiment in value-based health insurance offers an opportunity to study changes in health care financing and organization for children in Colorado foster care. In 2009 the Colorado legislature passed a budget action authorizing the Medicaid Value-Based Care Coordination Initiative; this Medicaid system reform was implemented in May 2011 as the Accountable Care Collaborative (ACC) Program. After the first year the ACC program 21% of the total Medicaid population had enrolled (620K in FY2012). The ACC Program contracted for all Medicaid beneficiaries within seven geographical Regional Care Collaborative Organizations (RCCOs) via a competitive bidding process. Medicaid is approximately one-third adults and two-thirds children; by June 2012 enrollment reached 133,000 members; nearly 50 thousand were eligible as children or in foster care.
The ACC shifted reimbursement from fee-for-service to capitation (per member per month). In FY2013, one dollar of the PMPM was redirected to an incentive pool for performance targets. Each Primary Care Medical Provider (PCMP) contracted with the ACC also earns a PMPM as a medical home; this pool also set aside an incentive pool. We will investigate the impact of this system change on polypharmacy, psychotropic medication use, service utilization, and process indicators of quality. The study population is ACC-enrolled children in foster care; comparison groups are non-foster eligible children, and non-enrolled foster and non-foster children. Children are assigned to a PCMP at the time of enrollment if they have a clear pattern of use with that provider (RCCO enrollees). Children with a clear pattern of use with a provider who is not in the ACC Program are not enrolled (Non-enrollees). The exogenous changes are from fee-for-service to capitated regional care organizations, and additional performance incentives. We will use difference-in-differences.
Strata include age (< 5 years, 6-11 years, 12-17 years of age) and type of psychotropic medication (antipsychotic, antidepressant, mood stabilizer, stimulant, antianxiety medication). During the implementation period (2012), 25% of foster children and 5% of non-foster children had at least one prescription fill for a psychotropic medication (out of 406K non-foster plus 17K foster). Out of all psychotropic medications in the youth Medicaid population, 50% were stimulants, 35% antidepressants; 25% each were antipsychotics or mood stabilizers; and 10% were antianxiety medications (percentages exceed 100% due to multiple medications). Foster youth aged 6-18 years had significantly higher rates of two or more psychotropic medications in the same class, and also across different classes, compared to non-foster youth (p<0.01).