Adding noise to Markov cohort state-transition model in decision modeling and cost-effectiveness analysis

Area of Science:

• Health economics
• Mathematical modeling
• Biostatistics

Background:

• Markov cohort state-transition models are widely used in health decision modeling and cost-effectiveness analysis.
• These models traditionally represent average population trajectories over time.
• Recent work connected cohort models to the average of continuous-time stochastic processes governed by master equations.

Purpose of the Study:

• To introduce a novel modeling method using stochastic differential equations (SDEs) as an alternative to conventional Markov cohort models.
• To demonstrate that SDEs can capture not only the mean but also the variance of population processes.
• To provide a practical framework for applying SDEs in health decision modeling.

Main Methods:

• Derivation of an SDE model from first principles.
• Development of an algorithm for constructing and simulating SDEs.
• Application of the SDE method to two distinct examples in health decision modeling.

Main Results:

• The SDE model accurately reproduces population trajectories, mean, and variance, matching conventional methods.
• The SDE method requires no additional inputs beyond those for standard cohort models.
• SDE models demonstrate superior computational speed compared to microsimulation models.

Conclusions:

• Stochastic differential equation (SDE) models offer a powerful alternative for population modeling in health economics.
• SDEs incorporate variance information, handle time-varying parameters, and are computationally efficient.
• This approach enhances health decision modeling by providing richer insights into population dynamics.