Modelling and predicting complex patterns of change using growth component models: An application to depression trajectories in cancer patients

In this paper we present a general and flexible framework for constructively defining growth components to model complex change processes. Building on the concepts of the latent state-trait theory (LST theory; Steyer, Ferring, & Schmitt, 1992), we develop structural equation models containing latent variables that represent latent growth (change) components of interest. We formulate these models based on an approach presented by Mayer, Steyer and Mueller (2012). We discuss an application to the longitudinal course of depression in 2,794 individuals from the Health and Retirement Study, who experienced cancer diagnosis over the course of the study. We found that (1) on average, the depression trajectories showed a steep increase after diagnosis as well as an adaptation phase where levels returned back to levels prior to diagnosis, and (2) individual differences in change were large and could be partly explained by marital status and cognitive functioning.