Latent subtypes of depression in a community sample of older adults: Can depression clusters predict future depression trajectories?

A model with three unique clusters best fit the data. Cluster 1 (59%) had a low probability of any symptom endorsement. Cluster 2 (31%) endorsed as a group some negative affect and somatic symptoms but their endorsement of low positive affect did not differ from Cluster 1. Participants in Cluster 3 (10%) had a higher probability of endorsement of all symptoms compared to Clusters 1 and 2. The results did not appreciably differ when symptom severity was included. Cluster assignment was a significant predictor of change in CES-D score over the ten-year follow-up period, and the effects over time differed by sex. Depressive symptom profiles predict the longitudinal course of depression in a community sample of older adults, findings that are important especially in primary care settings.