New Tool Refines Heart Attack Risk Prediction
Heart Attack Risk
For about 10 years, the Framingham risk score has been used to estimate a person's chances of having a heart attack based on just six bits of information -- age, sex, total cholesterol, HDL cholesterol, smoking status, and systolic blood pressure.
Doctors know what to recommend for people whose scores indicate high or low risk. But it's less clear what to do with those in the middle.
Over the years, researchers have experimented with adding additional risk factors to the formula to try to narrow the grey zone of mid-range results. Now, after testing three dozen separate risk factors, Harvard researchers have found that adding just two --a measurement of C-reactive protein and whether a parent had a heart attack before age 60 -- to the Framingham model made the resulting predictions even more accurate, reports the May 2007 issue of the Harvard Heart Letter.
Based on information collected from more than 24,000 women for more than a decade, the researchers created a new tool called the Reynolds risk score. When used on the study group, the Reynolds risk score did as well as the Framingham risk score for women at high and low risk. For those in between, it was better. The new model reclassified almost half of these women into high- risk and low-risk groups. The new assignments, done by computer, corresponded almost perfectly to what actually happened to these women over the next 10 years.