All models are wrong, but some are useful.—George Box
Only a few of the many clinical problems in medicine can be addressed by performing a randomized clinical trial. Consequently, we need additional tools to identify the optimal clinical management strategies for patients. One set of tools is the analyses of clinical registries and databases to assess diagnostic tests or to compare treatments—activities now termed comparative effectiveness research. Another powerful tool is the use of a model, or simulation, of a clinical decision to examine the likely consequences of alternative choices. A model is particularly useful in scenarios with multiple clinical options because it is difficult to study several strategies in a single clinical trial.