Doctors are constantly telling us that prevention is the best medicine, especially when it comes to chronic problems like heart disease. The key to a good prevention strategy is, of course, controlling the risk factors that contribute to disease. But now a new study shows that a commonly used measure to predict — and therefore help prevent — heart disease may not be as accurate as previously thought.
Researchers in Spain report this week in the Annals of Family Medicine that the Framingham risk score, a standard score that doctors use to gauge a patient’s risk of heart disease, overestimated that risk in a population of healthy Spanish men and women.
A version of the Framingham model specially calibrated to the Spanish population also failed to accurately predict risk. That could mean that doctors using such scores may be basing treatment decisions on inaccurate numbers.
The Framingham model is designed to predict heart risk based on five factors — age, gender, cholesterol levels, blood pressure and smoking status. In the Spanish study, which tracked 447 healthy men and women aged 35 years to 74 years for 10 years, the Framingham score predicted that 15% of the study population would have a heart event. In fact, 10% of participants did. Meanwhile, the customized model estimated that only about 6% would develop heart trouble.
Overall, that means the original Framingham risk assessment overestimated heart disease risk by 73%, while the Spanish version underestimated risk by 64%.
As discouraging as that sounds, the findings don’t necessarily mean the Framingham score isn’t worth a beat. The model is based on a study that was begun in the 1960s (and still continues today, with surviving members of the original cohort and their children), at a time when the risk factors for heart disease — high-fat diets, high blood pressure and smoking, for example — weren’t as well known as they are today.
Given that patients today are constantly told to eat better and quit cigarettes, the authors say it’s not surprising that the original model would overestimate risk in people who were more likely to get many of these risk factors under control between the start and end of the study.
Plus, says Dr. Robert Eckel, a professor medicine at the University of Colorado and past president of the American Heart Association, the number of participants in the current study was small, which means the number of actual heart events was small as well. So it’s hard to say how valid the conclusions are.
The findings do hint, however, that perhaps it’s time for an update to the decades-old model. Taking into account a person’s history of heart disease, weight (via body mass index), and use of drugs to control hypertension, for example, could refine the Framingham model and make it more accurate.
That’s exactly what heart experts in the U.S. are doing now. There is some discussion about how globally useful the Framingham risk score remains, since its original cohort was not as ethnically diverse as is our current population. Plus, heart experts have been concerned about the fact that the model does not include a consideration of family history of heart disease, which recent studies show can play a role in an individual’s risk.
“My opinion is that ultimately we need to expand on the current way of assessing global risk,” says Eckel. “This study is just an appetizer for the full meal to follow.”
Alice Park is a writer at TIME. Find her on Twitter at @aliceparkny . You can also continue the discussion on TIME’s Facebook page and on Twitter at @TIME.