Analysis of CCRC Data

This paper presents an approach to analyzing continuing care retirement community (CCRC) data, and demonstrates the methods using data from a large CCRC. It is assumed that residents make "transitions" among a number of "states" that represent the levels of care required by residents. There is randomness associated with both the transition times and the states entered at these times. The model is conveniently characterized in terms of "transition intensity functions", which represent the instantaneous rates of transition between pairs of states. Statistical methods for estimating these functions are discussed, and estimates are obtained from the data set. A simulation approach to determining probabilities and other interesting quantities based on the estimated intensity functions is also described and illustrated.