Author + information
- Ralph Stern, MD, PhD⁎ ()
- ↵⁎University of Michigan, Department of Internal Medicine, CVC Cardiovascular Medicine, 1500 East Medical Center Drive, SPC5853, Ann Arbor, Michigan 48109-5853
The excellent review by Berger et al. (1) perfectly illustrates a problem with individual risk prediction, which is that different risk scores provide different risk estimates for the same patient. Table 2 of their paper (1) shows estimated risks ranging from 2% to 39% for a patient asking about her risk of a future cardiovascular event. Much of this discordance represents the different outcomes and different time periods used in the various cardiovascular risk scores. But even 10-year global cardiovascular risk ranges from 6% to 14%. This is a large discordance, and different preventive measures may be offered depending on the choice of a cardiovascular risk score.
When the U.S. Preventive Services Task Force published guidelines on aspirin for prevention of cardiovascular disease, correspondents raised the same point, suggesting the guidelines be reissued “with a specific risk-assessment tool that has been thoroughly studied to ensure the clinically appropriate application of these important guidelines” (2). The authors responded that development of a “gold-standard” cardiovascular disease risk calculator was a pressing priority (3). However, a recent review found many examples of the discordance of individual risk estimates and concluded that such discordance was unavoidable (4). The reason for this is that individuals do not have definite, pre-existing probabilities of a cardiovascular outcome that can be precisely estimated. In mathematical terms, the problem of risk stratification does not have a unique solution.
How do the authors manage such discordant individual risk estimates in the clinic?
- American College of Cardiology Foundation
- Berger J.S.,
- Jordan C.O.,
- Lloyd-Jones D.,
- Blumenthal R.S.
- Stern R.H.