Microsimulation Modeling of Coronary Heart Disease

Abstract

We use microsimulation to forecast changes in coronary heart disease (CHD) among adults 45 or above over a 20-year time horizon in Los Angeles County (N = 3.4 million), a county with 12 635 CHD deaths in 2010. We simulate individuals’ life course and calibrate CHD trends to observed trends in the past. Using the Health Forecasting Community Health Simulation Model, we simulate CHD prevalence and CHD mortality in 2 CHD prevention scenarios: (1) “comprehensive hypertension intervention” and (2) “gradual reduction of the average adult body mass index back to the year 2000 level.” We use microsimulation methodology so that nonprofit hospitals can easily use our model to forecast intervention results in their specific hospital catchment area. Our baseline model (without intervention) forecasts an increase in CHD prevalence that will reach 13.01% among those 45+ in Los Angeles County in 2030. Under scenario 1, the increase in CHD prevalence is slower (12.47% in 2030), and the prevalence in scenario 2 reaches 12.83% in 2030. The baseline scenario projects a number of 21 300 CHD deaths in 2030, whereas there will be 20 070 CHD deaths under scenario 1 and 20 970 CHD deaths under scenario 2. At the population level, the CHD mortality outcome, as compared with the metric of CHD prevalence, might be more sensitive to preventive lifestyle interventions. Both CHD prevalence and CHD mortality might be more sensitive to the hypertension intervention than to the obesity reduction in the time horizon of 20 years.