Multi-state outcome analysis of treatments (MOAT): application of a new approach to evaluate outcomes in longitudinal studies of bipolar disorder

Abstract

Survival analyzes are usually based on a single point in time predefined event. Dissatisfied with this approach to evaluating maintenance treatment outcomes, we developed the Multi-state Outcome Analysis of Treatments (MOAT) methodology using a combined database from two FDA registration studies of lamotrigine, lithium and placebo. MOAT partitions total survival time into clinically distinct periods operationally defined by cutpoints on rating scales. For bipolar disorder (BD), the clinical states are remission, subsyndromal and syndromal mania, mixed states or depression. MOAT results can be crossed with information about tolerability and functioning to yield an outcome system integrating efficacy and tolerability. As found in the original analysis, both drugs were associated with longer time in study compared with the placebo. MOAT supplements this by finding that both drugs increased the time remitted compared with placebo. However, a substantial amount of time in all three treatments was spent in subsyndromal depression. Time with manic symptoms was reduced with lithium, but not lamotrigine. Patients on placebo neither benefitted nor had adverse effects from the assignment but experienced more syndromal levels of symptoms and were terminated from the study sooner than either drug treated group. Lithium was associated with both benefit in time manic and worse tolerability compared with placebo. In summary, lamotrigine was associated with limited therapeutic benefit but not harm; lithium with both benefit and harm; and placebo with neither. MOAT describes not only quantity but also quality of time spent in longitudinal studies, providing a more clinically informative picture than Kaplan-Meier survival analysis.

Figure 1

Integrated symptom-adverse events (AE) outcomes. LTG, lamotrigine.