Prevalence, utilization patterns, and predictors of antipsychotic polypharmacy: experience in a multistate Medicaid population, 1998-2003.

OBJECTIVES: This study was conducted to estimate the prevalence of antipsychotic polypharmacy among fee-for service state Medicaid beneficiaries initiating antipsychotic drug therapy and to investigate psychiatric and demographic predictors of such polypharmacy.

METHODS: This was a retrospective cohort study employing Medicaid claims data from California, Nebraska, Oregon, Utah, and Wyoming for patients who filled >1 antipsychotic prescription between 1998 and 2003 and who were continuously eligible for benefits from 180 days before to 1 year after the index antipsychotic claim. Antipsychotic Polypharmacy was defined as initiation of multiple antipsychotic medications or at least 60 consecutive days of concomitant antipsychotic medication overlapping the index antipsychotic prescription at any time during the 365 days after the index drug claim. Primary and secondary diagnosis codes (International Classification of Diseases, Ninth Revision, Clinical Modification) were used to identify patients with mental disorders and mental health-related hospitalizations. Multivariate logistic regression, with adjustment for sex, age, race/ethnicity, state, mental health diagnoses, hospitalization, year, and type of index antipsychotic, was performed to identify predictors of polypharmacy. A multivariate Cox proportional hazards model was used to compare the cumulative incidence of polypharmacy by index antipsychotic drug.

RESULTS: The study cohort consisted of 55,481 individuals with > or =1 prescription claim for an antipsychotic drug. The mean prevalence of long-term antipsychotic polypharmacy in the year after initiating antipsychotic medication was 6.4%. Approximately half of those with polypharmacy were started on multiple antipsychotic drugs and half were started on monotherapy but received > or =2 antipsychotic drugs concomitantly in the year after drug initiation. Among the stronger predictors of polypharmacy were a diagnosis of schizophrenia (odds ratio [OR] = 2.95; 95% Cl, 2.43-3.58), recent mental health hospitalization (OR = 1.17; 95% Cl, 1.02-1.33), and the number of mental health diagnoses (OR = 1.07 per diagnosis; 95% CI, 1.06-1.08). Polypharmacy was more likely among male than female patients (OR = 1.26; 95% Cl, 114-1.39) and among those between the ages of 18 and 24 years. The cumulative incidence of polypharmacy was greater among patients initiating clozapine compared with those initiating other antipsychotics (P < 0.001).

CONCLUSIONS: In these fee-for-service Medicaid beneficiaries from 5 states, the prevalence of chronic antipsychotic polypharmacy was low in the year after the initiation of therapy. Polypharmacy was more common in patients with indicators of more severe mental illness.