—Dr Marks raises two issues to which we take exception. One, a statistical matter, is the notion that R2 and the confidence interval for the predicted outcome are similar measures of predictability in regression analysis. The other, a matter of clinical judgment, is that ± 2.00 D is a reasonable level of predictability for radial keratotomy.We disagree with the statement in Arrowsmith and Marks' article that "a low R2 would indicate a poor predictive model, while a high R2 would indicate good predictability." While R2 does indicate the percentage of the variability in the outcome measure that can be "explained" by the factors in the regression equation, R2 does not indicate the precision with which the outcome can be predicted.1 The confidence interval, which provides a range in which the outcome for an individual patient is likely to fall, clearly shows
Lynn MJ, Hall EC, Waring GO, Sperduto RD, Binder PS. Predictability of Radial Keratotomy-Reply. Arch Ophthalmol. 1987;105(6):742. doi:https://doi.org/10.1001/archopht.1987.01060060019005
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: