33. Robustness of Simulation Results

All simulations include many elements of randomness, so multiple identical runs will generate different results (assuming seeding is off). Given this "noise" among model runs, how do we know if a single run has generated results that are stable enough to draw robust conclusions.

This chapter focuses on techniques that can help assess how robust the results are.

Note that the contents of this chapter apply to all Monte Carlo simulations - Patient Simulation, Probabilistic Sensitivity Analysis, Patients & PSA, and EVPPI.