Group-based trajectory modeling (GBTM) is a statistical method created to explore the heterogeneity of clinical groups based on their longitudinal outcomes by identifying distinct trajectories of change. This model can be applied to assess heterogeneity in responses to treatment. This pilot study explored the relevance of the GBTM associated with the dimensional evaluation of mood (MATHYS) to define trajectory of recovery in acute bipolar mood episodes on a short period of time during a naturalistic study.

The sample consisted in 118 bipolar patients and all patients were recruited during an acute phase: 56% had a major depressive episode, 26% a manic or hypomanic episode, and 18% a mixed state using the DSM-IV criteria. Patients were assessed four times with MATHYS during a three weeks follow-up period. It is an observational study and treatment was prescribed as usual. We applied the GBTM method and MATHYS total score to define trajectories of recovery.

This method allows identifying 4 trajectories of recovery. At Baseline, two of them started with a score of inhibition but with quite different evolutive profiles (stable inhibition versus improvement). The two others trajectories started with a score of activation (mild versus moderate) and showed a linear improvement of symptoms but with a more rapid recovery for the patients with the higher activation at baseline.

When considering the diagnosis of patients belonging in each trajectory, there model seems particular relevant to explore the high heterogeneity in response to treatment in bipolar patients during an acute depressive episode.