How to Design a Markov Model to Address Cost-Depactiveness

There is a progression-free survival benefit and some folks might change their practice based on a PFS benefit, so your question is very relevant. Historically for oncology, partition survival models were created. These are very simple models but they aren't able to incorporate a lot of the nuances that we see in lymphoma. A Markov model allows us to create very complex health states. We could create as many health states as we want. It also allows us to calibrate each part of the model using published survival curves.