Comparison of antipsychotic polypharmacy use between academic and community hospitals

Background: Antipsychotic polypharmacy (APP) is not proven to be either safe or efficacious for the treatment of mental illness. When mentioned in treatment guidelines, APP is recommended only after trials of clozapine and long acting preparations. In one of the few pharmacoeconomic studies of APP, when compared with receiving 1 antipsychotic, medication costs were 2.5 and 4 times higher for patients receiving 2 or 3 antipsychotics, respectively. However, rates of APP as high as 50% are reported in the literature. Differences within healthcare systems make cross study comparisons of APP problematic. The purpose of this study was to directly compare APP in a community (CH) and academic hospital (AH).

Institution: Cambridge Health Alliance (CHA) is a public health system consisting of an AH, two CHs, 20 primary care clinics and the Network Health Plan. Administration for hospitals is centralized. Hospitals utilize one formulary. Patients are admitted to hospitals based on bed availability not clinical presentation. Within this setting it is possible to isolate the effects of academic affiliation on medication use.

Methods: This was a retrospective observational quality improvement project. Records for patients admitted during 3 months to adult psychiatric units receiving antipsychotics at 2 hospitals within CHA were reviewed. Baseline and clinical data were collected.

Results: Baseline characteristics did not differ between institutions. A total of 197 and 192 patients received at least one antipsychotic at the CH and AH respectively. Use of ≥2 antipsychotics was greater at the CH (39.1%) vs. the AH (26%). Use of >2 antipsychotics was relatively infrequent in both settings. Use of 2 typicals or typical/atypical was more common in the CH (8.9%, 63.7%) vs. the AH (6.7%, 46%). Use of 2 atypicals was more common in the AH (47.3%) vs. the CH (27.4%). The 3 most commonly used antipsychotics in rank order at both facilities were quetiapine, risperidone and olanzapine. Average dosages of quetiapine, risperidone and olanzapine were low at both the CH and AH (226.8mg, 246.87mg; 2.08mg, 2.66mg; 14.27mg, 12.03mg; respectively). Use of long acting preparations and clozapine was infrequent in both settings.

Conclusions: Use of APP was relatively common in both settings, although more so at the CH. Education regarding use of APP should be completed at both types of institutions. Emphasis should be placed on the evidence base supporting maximization of antipsychotic dosages and utilization of clozapine and long acting preparations.