Partitioned Survival Model

A partitioned survival model is an economic modeling technique used to track a theoretical cohort over time as they transition between a set of exhaustive and mutually exclusive health states. Unlike Markov models, where the number of individuals in each state is determined by transition probabilities, partitioned survival models estimate the proportion of the cohort in each state using parametric survival equations.

Key features of partitioned survival models include:

– Health States: The model includes a set of health states that are both exhaustive (covering all possible states) and mutually exclusive (an individual can only be in one state at any time).

– Survival Equations: The proportions of individuals in each health state are derived from parametric survival equations rather than transition probabilities. Common parametric functions used include exponential, Weibull, and Gaussian distributions.

– Application in Cancer Treatments: These models are frequently employed to assess cancer treatments, with separate survival equations for overall survival and progression-free survival.

– Sensitivity Analysis: By varying the parameters of the survival equations, sensitivity analysis can be conducted to explore the impact of different assumptions on the model’s outcomes. However, care must be taken to avoid logical inconsistencies, such as overall survival exceeding progression-free survival if the equations are treated independently.

Partitioned survival models are particularly useful in health economic evaluations where detailed survival data are available, allowing for a more accurate representation of patient outcomes over time. These models provide a robust framework for assessing the cost-effectiveness of interventions, especially in conditions like cancer where survival is a critical outcome measure.

Click HEOR TERMINOLOGY