Are Risk Stratification Tables the Best Way to Evaluate Model Performance?

  1. Ralph H. Stern, PhD, MD; and
  2. Dean E. Smith, PhD
  1. From University of Michigan, Ann Arbor, MI 48109.

    TO THE EDITOR:

    Janes and colleagues (1) advocate using risk stratification tables, not the area under the receiver-operating characteristic (ROC) curve or c-statistic, and valuing models that assign a wide range of risks to individuals who were assigned a narrow range of risks by another model. A recent review of this topic (2) reached different conclusions.

    Simply providing a risk distribution curve (frequency vs. risk in the population) is a more informative way to present the …

    [Full Text of this Article]
    « Previous | Next Article »Table of Contents

    This Article

    1. Ann Intern Med March 17, 2009 vol. 150 no. 6 427-428
    1. » Excerpt
    2. Full Text
    3. Full Text (PDF)

    Rapid Responses

    1. Submit a response
    2. No responses published

    Current Issue

    1. November 17, 2009, 151 (10)
    1. Alert me to current issues