Abstract 4986: A Physiology-Based Mathematical Model of Coronary Heart Disease Accurately Predicts CHD Event Rates in Real Populations
The high societal burden of coronary heart disease (CHD) and prohibitive costs associated with clinical trials mandate the need for other methods to investigate new diagnostic protocols, disease management strategies and treatments. The Archimedes Model is a clinically detailed simulation model of human physiology, disease progression, and healthcare systems designed to support such investigations. The model has been validated against more than 50 of the most important clinical trials involving both primary and secondary prevention populations. The model includes emerging biomarkers and conventional CHD risk factors. Here we assess the accuracy of the model by simulating the evolution of CHD events in two primary prevention populations: the Atherosclerosis Risk In Communities (ARIC) and the Framingham Heart Study (FHS) prospective observational studies. Initial populations were created by matching the baseline information in FHS (N=4960) and ARIC (N=12800), and individual-level outcomes of incident and recurrent CHD events (non-fatal and fatal myocardial infarction) were tracked over time. These data were used to calculate Kaplan-Meier curves. Two-tailed p-values were derived from comparison of the simulated and actual survival curves using the log-rank test. P-values for incident (ARIC: p=0.349; FHS: p=0.811) and recurrent (ARIC: p=0.879; FHS: p=0.951) events indicate that there are no significant differences between the simulated and real populations. In conclusion, the Archimedes Model is capable of predicting both incident and recurrent CHD events with a high degree of accuracy in general populations, making it a valuable tool for healthcare applications.