Author Affiliation: Department of Internal Medicine, University of Michigan, Ann Arbor (email@example.com).
To the Editor: Another factor to consider when evaluating prognostic indices for older adults1 is whether their predictions for an individual agree. It is not well recognized, but different prognostic indices provide different estimates for the same individual and these differences can be substantial.2,3 Based on this lack of reliability, Feinstein4 suggested that clinicians instead rely on pertinent resemblance subgroups instead of multivariable methods. Thus, life expectancy estimates for patients of the same age, sex, and race from life tables may remain preferable, even when otherwise ideal prognostic indices are developed in the future.5
Stern R. Prognostic Models for Older Adults. JAMA. 2012;307(18):1911. doi:10.1001/jama.2012.3538