Abstract
Objective
To compare the risk of hospitalization for adult Medicaid beneficiaries with bipolar I disorder treated with lurasidone vs. other oral atypical antipsychotics (AAPs) as monotherapy.
Methods
A retrospective cohort study of the IBM MarketScan Multi-State Medicaid Claims database identified adults with bipolar I disorder who initiated an AAP (index date) between 1 January 2014 and 30 June 2019. Patients were continuously enrolled 12 months pre- and 24 months post-index date. Each month during the post-index period was categorized as monotherapy with lurasidone, aripiprazole, olanzapine, quetiapine or risperidone, no/minimal treatment, or other. Marginal structural models were performed to estimate hospitalization risk and length of stay (LOS) (all-cause and bipolar I disorder-related) compared to lurasidone.
Results
The analysis included 8262 adults. Compared to lurasidone, the adjusted odds ratios (aORs) of all-cause hospitalization were significantly higher for olanzapine (aOR = 1.60, 95% CI = 1.09–2.10) and quetiapine (aOR = 1.54, 95% CI = 1.18–1.89). The risk was significantly higher for bipolar I disorder-related hospitalization for quetiapine (aOR = 1.57, 95% CI = 1.10–2.04) and risperidone (aOR = 1.80, 95% CI = 1.04–2.56) compared to lurasidone. The bipolar I disorder-related LOS per 100 patient-months was more than twice as long for quetiapine (8.42 days) compared to lurasidone (3.97 days, p < .01).
Conclusions
Lurasidone-treated adult Medicaid patients with bipolar I disorder had significantly lower risk of all-cause hospitalization than those treated with olanzapine and quetiapine and lower risk of bipolar I disorder-related hospitalization than quetiapine and risperidone. Bipolar I disorder-related hospital LOS was significantly shorter for patients treated with lurasidone compared to quetiapine.
Transparency
Declaration of funding
This study was funded by Sunovion Pharmaceuticals Inc.
Declaration of financial/other relationships
X.N., C.D., K.L, G.R.W. and A.L. have disclosed that they are employees of Sunovion. P.V., S.D. and Y.L. have disclosed that they are employees of PRECISIONheor, which received funding from Sunovion to conduct this study.
Author contributions
All authors were directly involved in the design of the study, interpretation of results, drafting of the manuscript and providing final review. S.D. and Y.L. were involved in study conception and design, conducting the analyses, interpretation of study findings, drafting/editing the manuscript and providing final approval. P.V. oversaw the data analysis.
Acknowledgements
We thank Barbara Blaylock PhD from Blaylock Health Economics LLC for providing medical writing support.
Availability of data and material
This retrospective database study used Medicaid claims data from the IBM MarketScan Research Database (IBM, Somers, NY, USA) spanning 1 January 2014 to 30 June 2019. The claims data that support the findings of this study are from a proprietary administrative claims database and are not publicly available. However, summary data tables are available from the authors upon reasonable request.
Previous presentation
An earlier version of this work was presented as a poster at the Psych Congress; 2020 Sep 10–13; Virtual Conference, USA.