profession
Risk of major coronary event miscalculated in millions of patients
■ The point-based Framingham risk assessment tool classified about 5.7 million people differently than the original equation-based model, a new study says.
By Christine S. Moyer — Posted Sept. 28, 2010
Physicians who use a simplified version of the Framingham risk assessment tool may be miscalculating patients' likelihood of a major coronary event, a new study says.
The original Framingham model relies on a complex equation to estimate a person's 10-year risk of myocardial infarction and coronary death. The formula has been simplified to a point-based system that allows physicians to determine a patient's risk without using a calculator or computer.
The study, published online Sept. 8 in the Journal of General Internal Medicine, said that about 5.7 million people were classified as being at a different risk level under the point-based system than under the original Framingham model. Those varying classifications would result in different medical treatment recommendations.
"The key message is not that the point-based model is bad. It's definitely better to use that than nothing at all," said senior study author Michael A. Steinman, MD, an associate professor of medicine at the University of California, San Francisco.
But he recommends that physicians use the equation-based model if they have computers or hand-held devices in exam rooms that can calculate the complex formula. He said the model may be slightly more accurate than the point-based system. Using the equation-based model also would create greater consistency in determining a patient's risk for a major coronary event, he said.
The study recommends that doctors who made treatment decisions based on the point-based method recalculate patients' risk using the original Framingham model (link).
Coronary heart disease caused about one of every six deaths in 2006, according to the most recent data from the American Heart Assn.
For the study, researchers examined data on 2,543 people age 20 to 79 who participated in the 2001-2006 National Health and Nutrition Examination Surveys. The study group's data were weighted to be representative of about 39 million adults.
Patients were excluded if they took lipid-lowering medications. Also omitted were people who had coronary heart disease or risk equivalents, including myocardial infarction. The Framingham models are not designed to predict risk in those populations, the study said.
Researchers calculated the 10-year risk of major coronary events for each participant, using the original model and the point-based system.
They found that the point-based method classified 15% of participants (equivalent to about 5.7 million people) into different risk groups than did the original Framingham model. Of those patients, 10% were misclassified into higher risk groups and 5% into lower groups.
"This does not say that certain patients are getting the wrong treatment. ... But it certainly raises potential for concern," Dr. Steinman said.