PP040 Hospitalization Costs In Schizophrenia: Long-acting Injectable Antipsychotics Versus Oral Antipsychotic Use

Abstract

INTRODUCTION:Existing findings on effectiveness of long-acting injectable antipsychotics (LAIs) versus oral antipsychotics in preventing hospitalizations are inconclusive. This study was conducted to compare hospitalization costs between Medicaid patients diagnosed with schizophrenia who initiated a LAI and those who changed from one oral antipsychotic to another.METHODS:This retrospective cohort analysis used the Truven Health Analytics MarketScan® Medicaid claims database to study patients ≥18 years with schizophrenia. The two cohorts were: “LAI”, defined as initiating LAI (no prior LAI therapy) between 1 January 2013 and 30 June 2014; and “oral”, defined as changing from one oral antipsychotic to another during the same period. The first day of LAI or the new oral antipsychotic was the index date. A linear regression model was conducted to estimate hospitalization costs.RESULTS:The final sample included 2,861 (36.7 percent) LAI and 4,926 (63.3 percent) oral users. Compared to oral users, LAI patients were younger (mean (Standard Deviation, SD): 39.9 (13.2) versus 42.7 (13.1); p<.001) and had a lower mean Charlson Comorbidity Index score (mean (SD): 1.1 (1.9) versus 1.7 (2.3); p<.001). Of the 877 LAI initiators and 1,688 oral users who were hospitalized during the 1-year post-index follow-up period, the unadjusted mean hospitalization costs for LAI and oral users were USD32,626 and USD36,048, respectively. After adjusting for patient demographic and clinical characteristics, baseline medication use, and baseline ED or hospitalizations, the adjusted average hospitalization costs were USD1,170 lower in LAI initiators than oral users. None of the unadjusted or adjusted differences were statistically significant.CONCLUSIONS:This real-world study suggests that among hospitalized patients, hospitalization costs are lower in LAI initiators than in oral antipsychotic users, although the difference is not statistically significant. Our study is limited as our results are reflective of a multi-state Medicaid population. Future studies are warranted to confirm the results in non-Medicaid patient populations.