We employed the systematic review of randomized controlled
trials of psychoeducation in participants with bipolar disorder not in an
acute illness episode, compared with treatment-as-usual, and placebo or
active interventions. Pooled odds ratios (ORs) for non-relapse into any
episode, mania/hypomania, and depression were calculated using an
intent-to-treat (ITT) analysis, assigning dropouts to relapse, with a
sensitivity analysis in which dropouts were assigned to non-relapse
(optimistic ITT).