Partitioned Survival Model

Definition

A partitioned survival model (also called a partition model or area-under-the-curve model) estimates the proportion of patients in each health state at each time point directly from survival curves, rather than using transition probabilities. It is commonly used in oncology, where overall survival (OS) and progression-free survival (PFS) curves are available from clinical trials. The model is simpler to construct than a Markov model but offers less flexibility.