Premium
Number needed to treat (NNT): estimation of a measure of clinical benefit
Author(s) -
Walter S. D.
Publication year - 2001
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.1173
Subject(s) - number needed to treat , estimator , context (archaeology) , event (particle physics) , medicine , randomized controlled trial , statistics , absolute risk reduction , clinical trial , confidence interval , econometrics , computer science , relative risk , mathematics , surgery , paleontology , physics , quantum mechanics , biology
Abstract The number needed to treat (NNT) is becoming increasingly popular as an index for reporting the results of randomized trials and other clinical studies. It represents the expected number of patients who must be treated with an experimental therapy in order to prevent one additional adverse outcome event (or, depending on the context, to expect one additional beneficial outcome), compared to the expected event rates under the control therapy. Although NNT is a clinically useful measure, little work has been done on its statistical properties. In this paper, alternative NNT‐type measures are defined for use with discrete or continuous data. Estimators and their variances are obtained for these measures in cross‐over or parallel group designs. The ideas are illustrated with data on quality of life in asthma patients. Copyright © 2001 John Wiley & Sons, Ltd.