People living with type 2 diabetes have a higher risk of coronary heart disease (CHD). In a CHDS seminar, Bart Ferket discussed using risk stratification to prevent CHD among diabetics. Ferket is an Associate Professor in the Department of Population Health Science and Policy at the Icahn School of Medicine at Mount Sinai, NYC.
The level of CHD risk is heterogenous among those with diabetes, which may influence the optimal combination of testing by computerized tomography coronary artery calcium scoring (CT-CACS) and the use of statins and aspirin to prevent CHD. Ferket described how he developed a prediction model based on the Multi-Ethnic Study of Atherosclerosis (MESA). The model was trained on 10 years of follow-up data within MESA, and subsequently validated internally (against observed risks in the same 10 years of MESA data), predictively (against observed risks in the following 15 years of MESA data), and externally (against observed risks in the following 15 years of NHANES data).
Ferket used the CHD risk prediction model to identify cost-effective strategies for preventing CHD in those with type 2 diabetes. In doing so, he used a weighting approach to overcome the discrepancy in the risk factor distributions between MESA – which is oversampled for certain groups – and the nationally-representative NHANES data. His results show that adding CT-CACS to risk-based prevention strategies may have limited value; in most scenarios, an “intensive” strategy is cost-effective, in which all individuals with diabetes receive a statin. If aspirin is also considered as an additional preventive measure, then the cost-effective strategy would be to give aspirin to all patients with diabetes except those with a CT-CACS score of zero.
Learn more: Read the publication, Lifetime Cardiovascular Disease Risk by Coronary Artery Calcium Score in Individuals with and Without Diabetes
Learn more: Read about the Multi-Ethnic Study of Atherosclerosis
Learn more: Read the publication, Patient Variability Seldom Assessed in Cost-Effectiveness Studies
Related news: Learning Optimal Treatment Rules
Related news: Deep Data Analysis and Intelligent Policy Design