Sequential Multiple Assignment Randomized Trial for Bipolar Depression

This is a sequential multiple assignment randomized trial for adults (ages \> 18) with a bipolar disorder type 1 and type 2 diagnosis currently experiencing a depressive episode. It is a randomized pragmatic trial that will compare four commonly prescribed treatments for bipolar depression, which includes three FDA-approved medications (Cariprazine, Quetiapine and Lurasidone) and one antipsychotic/antidepressant combination (Aripiprazole/Escitalopram).

This is a comparative effectiveness study to address the critical questions of how best to treat people with bipolar disorder who have a major depressive episode: how to get them well, provide second-line treatment when they don't initially get well, and keep them well after they get well. This is a multisite Sequential Multiple Assignment Randomized Trial (SMART) comparative effectiveness design. Investigators will recruit 2726 participants who have BD I or II with a current major depressive episode lasting at least 2 weeks. In Stage 1, investigators will compare four treatment arms, including three FDA approved monotherapies and the most widely used, but understudied, treatment for BD major depressive episode (i.e., a non-FDA approved combination of an antipsychotic and antidepressant). In Stage 2, participants who do not remit will be re-randomized to treatments not used in Stage 1. Investigators will follow all participants for a total of 52 weeks. This study will be conducted in two phases, a feasibility phase and a full study phase. In the feasibility phase, investigators will recruit at 8 of the 19 study sites based on readiness and interest, to ensure the efficacy of the study design before launching the full study phase. Feasibility Aim 1: To ensure that 8 of the 19 selected sites can recruit, randomize, and retain participants. During the feasibility phase, the selected sites will recruit 133 participants, administer baseline assessments, randomize the participants (at baseline and again at 6-weeks for non-remitters), and conduct follow-up assessments at the end of Stage 1 (6-weeks) and the end of Stage 2 (12-weeks) and end of study (52 weeks or end of feasibility phase, whichever is sooner) and all other scheduled visits. Feasibility Aim 2: To refine and finalize all study standard operating procedures for the full-scale study. After Aim 1 is complete, any changes in standard operating procedures will be made as needed to be implemented in the full study phase. Full Scale Study Full-Scale Study Aim 1: To compare the effectiveness, tolerability, and safety of the four treatments for BD major depressive episodes. Hypothesis 1a: There will be significant differences across the Stage 1 treatments in the proportion of remitters at week 6 (primary endpoint, effectiveness) and in the average adverse event burden and suicidal ideation/behavior at week 6 (secondary endpoints, tolerability and safety). Hypothesis 1b: Among non-remitters of a given Stage 1 treatment (i.e., participants who do not remit by week 6), there will be significant differences across the three remaining Stage 2 treatments in the proportion of remitters at week 12 (primary endpoint, effectiveness) and the average side effect burden and suicidal ideation/behavior at week 12 (secondary endpoints, tolerability and safety). Full-Scale Study Aim 2: To characterize the effectiveness, tolerability, and safety of the 12 adaptive interventions for BD major depressive episodes embedded within the SMART design. Hypothesis 2: On average, there will be significant differences across the 12 embedded adaptive interventions in the proportion of remitters at week 12 (primary endpoint, effectiveness) and in the average side effect burden and suicidal ideation/behavior at week 12 (secondary endpoints, tolerability and safety). Full-Scale Study Aim 3: To determine individual-level characteristics that predict heterogeneity of treatment effect (HTE) across the 12 adaptive interventions for BD major depressive episodes embedded within the SMART design. Hypothesis 3a: Individual-level characteristics and symptoms assessed at baseline as well as symptom changes and side effect profiles at week 6 will predict HTE at weeks 12 and 52. Hypothesis 3b: Investigators hypothesize that an optimal personalized adaptive intervention will perform significantly better than the best aggregate embedded adaptive intervention identified in Aim 2 in terms of the proportion of remitters at week 12 (primary endpoint, effectiveness) and the side effect burden and suicidal ideation/behavior at week 12 (secondary endpoints, tolerability and safety).