Author Affiliations: Division of Adult Nephrology, Shaare Zedek Medical Center, Jerusalem, Israel (Dr Gabbay) (email@example.com); and Institute for Clinical Research and Health Policy, Tufts Medical Center, Boston, Massachusetts (Dr Kent).
To the Editor: The study by Dr Yourman and colleagues1 evaluated predictive models for survival in elderly adults as decision-making aids. Better life expectancy estimates may help avoid treatments in patients in whom high competing risks limit potential benefits. Several points merit consideration in this decisional context, beyond the usual metrics of model performance that are the focus of the review. First, predictions are not only point estimates but also, perhaps predominantly, probability distributions.
Gabbay E, Kent DM. Prognostic Models for Older Adults. JAMA. 2012;307(18):1911-1912. doi:10.1001/jama.2012.3568