Evidence-appraisal glossary
Net reclassification improvement
A measure of whether adding a new marker to a risk model correctly shifts people across risk categories: upward for those who later have the event and downward for those who do not.
Also called: NRI.
Net reclassification improvement was proposed because the C-statistic can be insensitive to a genuinely useful new predictor, and it counts how many people move to a more appropriate risk category once the marker is added. It has drawn substantial criticism, however, because it can suggest improvement even when a model is poorly calibrated, and its value depends heavily on where the category boundaries are drawn. Many methodologists now favor decision curve analysis or calibration checks alongside, or instead of, it.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.