Improving Value of Publicly Funded Mental Health Care

Specific Aims:

Forty years after chlorpromazine revolutionized the treatment and prognosis of schizophrenia, second-generation antipsychotics [SGAs] entered the U.S. market, radically transforming serious mental illness care.1 Annual antipsychotic treatment visits more than doubled between1995 and 2008, a period during which the share of SGAs rose from 16% to 93%. Although no more efficacious than alternative treatments, SGAs other than clozapine are routinely used as first-line agents for schizophrenia and, increasingly, for bipolar disorder. Antipsychotics, SGAs in particular, are frequently and increasingly used in combination despite a lack of evidence supporting the practice and safety concerns. Moreover, off-label use of SGAs is widespread, exposing patients to unnecessary risks. This growth, influenced by delivery system factors that vary regionally, has been costly to payers. The costs associated with SGA utilization may exceed those related to drug purchasing costs. Several SGAs are associated with increased risk for serious cardiometabolic morbidity, including type 2 diabetes and cardiovascular disease. Although metabolic testing is the standard of care,14 testing rates remain low16 and as a result, many at-risk patients may not be offered lower-risk antipsychotics. The long-term health and economic burden to payers associated with this excess metabolic risk is likely substantial yet unknown.

The health care received by individuals with schizophrenia, bipolar I disorder, and treatment-resistant major depressive disorder, the serious mental illnesses [SMI] for which SGA utilization is FDA-approved, is largely financed by Medicaid and Medicare. These payers bear a large portion of the economic consequences of potentially inappropriate antipsychotic utilization, including overuse (polypharmacy, off-label use) and use of drugs with high metabolic risk. Faced with the twin pressures of reduced revenue and increased enrollment, Medicaid and Medicare need to improve the value of publicly-financed care. Implementation of utilization management tools, quality improvement interventions, and other strategies to decrease antipsychotic overuse and increase safety of antipsychotic utilization offers a mechanism to achieve this goal. Payers will require new evidence to target these strategies to appropriate populations and avoid unintended consequences. Adoption of these strategies should be based not only on assessments of short-term benefits and costs but also long-term health and economic implications.

We propose a program of research to investigate antipsychotic utilization and its impacts among Medicaid, Medicare, and dually eligible beneficiaries, and using these findings, to simulate the impacts of strategies to improve value of care. We will quantify antipsychotic overuse among all adult beneficiaries. Among those with SMI, we will assess the short- and long-term health and economic impacts of cumulative antipsychotic exposure and model the effects of value-enhancing strategies. Health impacts include cardiometabolic disorders (dyslipidemia, hypertension, diabetes and cardiovascular disease); economic impacts include drug purchasing costs, diabetes and cardiovascular prevention or treatment costs, and total health care spending. We use longitudinal Medicaid and Medicare data for adults enrolled in one or both programs in six diverse states (California, Illinois, Louisiana, Missouri, New York, and North Carolina) and complement these with other data sources. Analyses will be conducted separately for each state and for each beneficiary population. Our Specific Aims are to:

AIM 1: Assess the contribution of overuse to state and regional trends in antipsychotic utilization and antipsychotic purchasing costs for the period 2007-2013. We will quantify the share of total antipsychotic utilization and share of total costs attributed to antipsychotic polypharmacy and off-label use.

AIM 2: Quantify the health and economic impacts of antipsychotic utilization in beneficiaries with SMI for the period 2007-2013. We will estimate models that adjust for time-dependent confounders to quantify the effects of cumulative exposure to antipsychotics of varying metabolic risk on health and economic outcomes.

AIM 3: Simulate the health and economic impacts of various strategies that may be used to enhance value of antipsychotic treatment in beneficiaries with SMI. We will develop a microsimulation model to quantify the 6-year and lifetime health and economic effects of cumulative exposure to antipsychotics of varying metabolic risk under varying scenarios. Parameter estimates will be obtained from Aim 2, the Medicaid and Medicare data, nationally representative survey information, and evidence gathered through a two-stage modified Delphi process.

Our findings will provide a comprehensive understanding of the extent of antipsychotic utilization and costs accounted for by overuse; the current and projected metabolic morbidity and cost burden associated with alternative antipsychotic prescribing practices; and the potential gains of implementing targeted strategies aimed at improving value of mental health care. The data and the methodology we adopt permit a multi-faceted assessment of value by empirically combining cohort-based observations, national survey responses, and Delphi-gathered expert opinion using contemporary statistical and microsimulation modeling approaches.