 The paper provides a step-by-step introduction to interpreting decision-curve analysis, which is used to evaluate prediction models and diagnostic tests, by re-labeling the y-axis as benefit and the x-axis as preference, and recommending a model or test for clinical use if it has the highest level of benefit across a range of clinically reasonable preferences. This article was authored by Andrew J. Vickers, Ben van Kolster and you, W. Steyerberg.