Latent classes of symptom trajectories during partial hospitalization for major depressive disorder and generalized anxiety disorder

A variety of treatments have been empirically validated in the treatment of major depressive disorder and generalized anxiety disorder. Researchers commonly evaluate symptom change during treatment using single model curves, however, modeling multiple curves simultaneously allows for the identification of subgroups of patients that progress through treatment on distinct paths. Latent growth mixture modeling was used to identify and characterize distinct classes of symptom trajectories among two samples of patients with either MDD or GAD receiving treatment in a daily partial hospital program. Four depression symptom trajectories were identified in the MDD sample, and three anxiety symptom trajectories were identified in the GAD sample. Both samples shared symptom trajectory classes of responders, rapid responders, and minimal responders, while the MDD sample demonstrated an additional class of early rapid responders. In both samples, low symptom severity at baseline was associated with membership in the responder class, though few other patterns emerged in baseline characteristics predicting trajectory class membership. At treatment discharge, those in the minimal responder class reported poorer outcomes on every clinical measure. Patients within each class reported similar scores at discharge as compared to each other class, indicating that class membership affects clinical measures beyond symptom severity. Patient demographic characteristics were relatively homogeneous. Group-based trajectory modeling inherently involves some degree of uncertainty regarding the number and shape of trajectories. Identifying symptom trajectories can provide information regarding how patients are likely to progress through treatment, and thus inform clinicians when a patient deviates from expected progress.