Statins and Public Health Policy: Modeling and Optimization

Pills.

Diabetes is the seventh leading cause of death in the U.S., and accounted for almost 70,000 deaths in 2010, the majority of which were caused by coronary heart disease (CHD) or stroke.

Currently, in the U.S., the size of the population diagnosed with diabetes is more than 25 million and unless the current trend in the incidence of diabetes changes, by 2050 one in every three adults is expected to be affected directly or indirectly by diabetes. Associated with its health implications, diabetes is a huge burden on the national economy. Its overall cost was $245 billion in 2012. 

Type 2 diabetes is the most common type of diabetes and accounts for more than 95% of the cases in the U.S. While there is no known cure for Type 2 diabetes, the risk of diabetes-related CHD and stroke can be alleviated by treating lipid abnormalities using statins, which are available at negligible cost with no prescription.

Statins reduce total cholesterol and increase high-density lipoprotein levels which in turn decrease the probability of CHD and stroke. Even though statins reduce the risk of CHD and stroke there have been cases of major (extreme muscle inflammation and damage) and minor (headache, nausea, fatigue and fever) side effects. While several studies have suggested a significant impact of disutility of statins in patients’ adherence to prescribed medication, except for few surveys there has not been any emphasis on quantifying these adverse effects to help physicians and patients make better treatment decisions. 

ISE professor Mark Karwan with PhD student Niraj Kumar Pandey use real clinical data and optimization models to quantify the disutility of using statins from a central policy maker’s point of view. Their work not only demonstrates that guidelines from across the world show variation in the disutility they attach to the treatment but also supports the common belief that treatment decisions are based on negligible medication disutility. Their analyses show that CHD and stroke risk equivalent of a life year on treatment is less than 1%. They further evaluate the cross-effectiveness of the guidelines and look at room for improvement for better treatment decision making.