[Skip to Content]
[Skip to Content Landing]
March 2, 1994

The Predictive Capabilities of Clinical Tests: The 6-Minute Walk-Reply

Author Affiliations

University of Alabama at Birmingham
University of North Carolina at Chapel Hill
Vanderbilt University School of Medicine Nashville, Tenn
Oregon Health Sciences University Portland
Brigham and Women's Hospital Boston, Mass for the SOLVD Investigators

JAMA. 1994;271(9):662. doi:10.1001/jama.1994.03510330039026

In Reply.  —Dr Mark suggests three criteria by which clinicians should judge a new clinical maneuver. The first is the amount of prediction: We agree that both absolute risks and RRs should be considered and have provided information on both in our article. Relative risks are most useful when comparing one predictive variable with another in the same patient population. In our study, for example, the point estimates of the two statistically significant odds ratios (ORs) for death were 1.74 for a 15% decrement in ejection fraction and 1.5 for a 120-m decrement in walking distance. These results indicate to the clinician that walking distance was the second-best predictor of prognosis among the variables evaluated in our model and was nearly as good a predictor as ejection fraction. Absolute risk is more relevant when the test is applied to a single patient. In our population, the risk of dying increased

First Page Preview View Large
First page PDF preview
First page PDF preview