Simulation of micropopulations in epidemiology: tutorial. 4. Evaluations of simulation models. A series of tutorials illustrated by coronary heart disease models.

The first tutorial in this series (Ackerman E: Simulation of micropopulations in epidemiology: Tutorial 1. Simulation: an introduction, Int J Biomed Comput, 36 (1994) 229-238) introduces the general approach of simulation of micropopulation models of coronary disease using Monte Carlo techniques. The modeling process includes the selection of functional forms to represent the probability of state transfers. Alternative functional forms are described in the second tutorial (Ackerman E: Simulation of micropopulations in epidemiology: Tutorial 2. Analytic forms of event probabilities, Int J Biomed Comput, 37 (1994) 139-149). The third tutorial (Ackerman E: Simulation of micropopulations in epidemiology: Tutorial 3. Simulation model evaluation methods, Int J Biomed Comput, 37 (1994) 195-204) considers the actual risk factors used in the models and the compartmentalization of the population. However, that tutorial emphasizes the methods of evaluating different models of coronary heart disease. In the current tutorial, the evaluation methods are applied to the models previously introduced. Essential to this is the concept of degrees of freedom. Limitations of that concept are discussed with specific reference to models of coronary heart disease. The estimation of optimal values for the risk coefficients is considered further. The various tests of appropriateness are applied to specific models and the role of sensitivity analysis is further illustrated. Discussions of intervention strategies and their simulation is deferred to the following tutorial.