I've fitted a latent growth mixture model to time series data. It consists of a value (population prevalence) at 11 time points for a sample of 150 areas. Said model was fitted using the lcmm package in R and identified a two class model as optimal - reflecting an hlme model as follows:
gmm2