Hi, i am looking to analyse a set of data which has yearly health-related measurements for patients with a particular disease. Each patient is also categorised due to the severity of their disease status, either mild, moderate or severe. I want to test if changes in health status over time are related to disease severity, so need to use some sort of repreated measures approach. However, there will be different numbers of measurements for patients because some will have had the disease for longer, and thus had more annual readings taken, and some will have died so will clearly have a limited number of measurements. The number of annual measurements per patient ranges from 2 to 11, with a total of around 80 patients. What would be the best approach to analysing this data? Many thanks in advance, Michael Roughton