Abstract
Because it can be easily understood by clinicians, the “number needed to treat” (NNT) has recently received considerable attention in the clinical sciences. When the NNT is applied to summarize findings, it is important to understand the limitations of its use. First, the important confusion and criticisms found in the past on using the confidence interval of the NNT are summarized. To eliminate the confusion and criticisms, the NNT should be used only when there is some prior knowledge to rank the order of the response rates under comparison. An example is included to illustrate the use of the proposed approach and show how a number of the criticisms, limitations, and undesirable properties of the interval estimator for the NNT can be avoided. To test the equality of two response rates, the use of test statistics based on other well-established indices is recommended in lieu of the NNT.
Get full access to this article
View all access options for this article.