I have a multinomial logistic regression model where I am predicting the probability of having a low, normal or high blood pressure based on weight, gender, exercise routine, stressful event in the last three months, etc. I've observed people over a six year time period but I have people who dropped off the study at various time over the six years, people who joined the study at various times and people who've been in the study the whole time. I tested every person who was actively involved in the study at the time once every year. So I have as many as 6 observations from the same person, most of which are identical. On the other hand, for some people, I only have one observation. Does JMP have a way of handling these repeated measures in the context of multinomial logistic regression?