SMDM Webinar: Calibrating Markov Models to Clinical Survival Data

About this Event

Building accurate Markov models is challenging because they cannot easily be matched up with clinical survival data. Partitioned Survival models are built directly on survival curves that can be fitted to clinical data through regression analysis. However, Partitioned Survival models are limited to forward progression, and they do not support critical events beyond implicit progression and death. Markov models are instead built on distinct events including, but not limited to, progression and death. However, Markov models cannot be directly fitted to clinical survival data because disease progression results from a combination of event risks applied over time. For Markov models, curve-fitting calibration techniques can refine event probabilities to match multi-cause disease progression to clinical survival data.