This is the second in a series of papers dealing with models of coronary heart disease. Three different types of statistical models are considered as risk functions: the multivariate logistic model, the Cox proportional hazard model and the Neyman exponential risk avoidance model. The types of models differ in the form hypothesized for the probability of occurrence of coronary heart disease outcomes: incident myocardial infarct, cardiac death, and death from other causes. Although the three risk functions are strikingly different, they can all be tested using the CRISPERS chronic disease simulation system. Simulations were performed using data from North Karelia, Finland. The polychotomous multivariate logistic risk function is convenient for studies involving increasing numbers of risk factors. The Cox proportional hazard regression model is shown to be unsuitable for the cohort dataset used as well as for some of the intended uses of the simulation models. The Neyman exponential risk avoidance model involves time in a quite different fashion. It has the inherent advantage of being easier to relate to underlying biological mechanisms because it is the integral of first order rate equations. It is concluded that more than one risk function should be evaluated for simulations of coronary heart disease.

Adult↗

PubMedDate not supplied

Polychotomous multivariate models for coronary heart disease simulation. III. Model sensitivities and risk factor interventions.

This is the third in a series of papers dealing with models of coronary heart disease. Sensitivity analyses of the logistic risk function and the Neyman risk function are reported. The resulting response surfaces are also used to investigate the optimality of the set of values for the risk coefficients. It is shown that the coefficients estimated by maximum likelihood are preferable to the sets from an optimisation procedure. Two different sets of risk coefficients estimated using short periods and entire epochs for the logistic risk function are shown to lead to similar conclusions concerning simulated primary intervention strategies. However, the corresponding risk factor reductions using the Neyman risk function lead to somewhat different effects. Additional information is needed to distinguish between these two assumptions of the risk function used to model coronary heart disease. This underscores the need to understand the effects of the underlying risk function assumed when interpreting simulated outcomes of intervention strategies.

Computer Simulation↗

PubMedDate not supplied

Simulation of micropopulations in epidemiology: tutorial. 3. Simulation model evaluation methods. A series of tutorials illustrated by coronary heart disease models.

This is the third in a series of tutorials concerning the simulation of micropopulation models to support epidemiological research. The series emphasizes techniques used in studies at the National Micropopulation Simulation Resource at the University of Minnesota. For pedagogic purposes, applications to coronary heart disease (CHD) models are used to illustrate the principles and methodologies employed. All of the models presented are implemented using available software. A variety of tests and techniques are used to evaluate these models. This tutorial presents some of those methods stressing the advantages and limitations of the tests rather than the formal definitions. To make the evaluation methods more understandable, some of the risk factors for CHD models are introduced. The interpretation of the evaluation depends critically on the goals of the modeling effort. The subset of evaluation methods presented includes investigations of the epidemiological, mathematical and statistical aspects of the models.

Computer Simulation↗

PubMedDate not supplied

Polychotomous multivariate models for coronary heart disease simulation. IV. The impact of physiological aging.

This is an extension of a series of papers dealing with certain models used in the simulation of coronary heart disease. The current study investigates implications of including age as a risk factor in the models discussed in the preceding papers. The effects of using age as a risk factor were investigated in two ways. In one of these, age is interpreted as age of entry into the study; it is similar to the other risk factors in that it is assumed to be constant throughout the study. In the other, age is interpreted as the actual age; thus it increases during the course of simulations. Two polychotomous, multivariate risk functions developed in previous studies, the logistic risk and the Neyman exponential risk, were used to explore the effects of including age as a risk factor. The estimated risk coefficient for age was found to be statistically significant for both functions. The model performance was evaluated by comparing the observational data with outcomes simulated using Monte Carlo techniques. It was found that the logistic risk function failed to describe the observations either with age as a constant or with aging during the simulations. The models including the Neyman exponential risk avoidance fit the data well. The evaluation of the results indicates that aging during the simulations is better than using only the age as the constant value at entry to the study.

Adult↗