Latent variable modelling of recovery trajectories: sexual function following radical prostatectomy

As treatments improve and survival time lengthens, the course of recovery and long-term quality of life (QoL) is of greater interest. The application of latent class growth models to longitudinal QoL data provides unique insights into recovery that are not evident with marginal analyses. We examine this approach in the context of a large population-based observational study. This is the first attempt to characterize individuals' recovery over time following prostate cancer surgery, and to determine factors associated with varying recovery experiences of prostate cancer patients. Four major patterns emerged that illustrate typical patterns of recovery following radical prostatectomy. Given a man's baseline data, this method produces estimates of the probability of belonging to each recovery class. The method is presented as a useful tool for identifying hypotheses associated with recovery and potential antecedents of importance.