Modelling and monitoring cluster transitions in R; detecting if users “disappear” or “move” to another cluster over different time points

Question

I do a lot of k-means clustering with longitudinal user generated data in R and one of the key things I would like to look at is can I detect when users "disappear" fr