| This dissertation evaluated the health outcomes of second-generation antidepressant medication use in patients with bipolar disorder, and applied different methodologies including the propensity score to control for sample selection bias.;Bipolar disorder is a severe, life-long, chronic illness, the depressed phase of which is a significant cause of suffering, disability, and mortality, but it has been extraordinarily understudied. The rationale for current treatment guidelines is influenced by a concern that continued antidepressant use might induce a switch into mania. However, there is also a possibility that antidepressant discontinuation itself could contribute to a depressive relapse.;The dissertation selected a bipolar population from the Integrated HealthCare Information Solutions, a large longitudinal national managed care claims database, to establish the association between the use of second-generation antidepressants and the outcomes of mania-related visits, depression-related visits, healthcare costs, and depressive relapse. The propensity score method and difference-in-difference approach were employed. The empirical results were compared with the standard regression results applying covariate adjustment.;This dissertation study adds to the literature of bipolar depression treatment by providing empirical evidence at the national managed care level. Second-generation antidepressant monotherapy and antidepressant-mood stabilizer combination therapy were not associated with a higher risk of manic-switching in bipolar depressed patients when compared to mood stabilizer monotherapy. In addition, both antidepressant monotherapy and antidepressant-mood stabilizer combination therapies were associated with reduced number of depression-related visits, although the combination therapy was associated with higher bipolar-related costs and overall healthcare costs. Finally, this dissertation suggests a potential adverse outcome (depressive relapse) of removing a second-generation antidepressant treatment after depressive remission.;Since the safety and efficacy of antidepressants in bipolar disorder have not been firmly established, medical practitioners and policy makers should use caution when using second-generation antidepressant medications in treating bipolar disorder. The methodological comparison in the dissertation helps researchers be aware that different methodologies might produce different results. Overall, outcomes of the dissertation further the understanding of consequences of treating bipolar illness using different common therapies, and broaden our understanding of appropriate methodological applications in health services research focused on treatment of mental health conditions using large databases. |
