Abstract
Unlike conventional meta-analyses, individual patient data (IPD) meta-analysis assesses moderator variables at the level of each participant, which generates more precise and biased estimates. The objective of this study was to investigate whether psychological therapy reduces depression symptoms in people with Bipolar I and II disorders and examine whether baseline depression has a moderating effect on treatment outcomes. Through the use of several electronic databases, a systematic search was conducted. Eligible studies were randomized controlled trials evaluating a psychological intervention for adults diagnosed with Bipolar I or II disorder. Titles and abstracts were screened, followed by full texts. The authors of the included studies were asked to provide IPD from their trials. A multilevel model approach was used to analyze the data. From the 7552 studies found by our searches, six studies with 668 study participants were eligible. Intervention significantly reduced depression scores. There was a significant association between baseline depression and post treatment depression scores. There was no statistically significant interaction between condition allocation and baseline depression score. When IPD from the two most comparable studies were analyzed, CBT had reduced depression scores relative to the comparator condition. The study included patient data from only six studies which were heterogeneous in terms of intervention type, outcome measure, and comparators. Overall, the psychological interventions tested significantly reduced bipolar depression scores. There was no evidence of moderation by baseline depression scores.