This work may not be copied, distributed, displayed, published, reproduced, transmitted, modified, posted, sold, licensed, or used for commercial purposes. By downloading this file, you are agreeing to the publisher’s Terms & Conditions.

Letters to the Editor

Dr Andrade Replies

See letter by Citrome and Ketter and article by Andrade

This work may not be copied, distributed, displayed, published, reproduced, transmitted, modified, posted, sold, licensed, or used for commercial purposes. By downloading this file, you are agreeing to the publisher’s Terms & Conditions.

Dr Andrade Replies

To the Editor: Drs Citrome and Ketter provide a vigorous defense of the usefulness of the numbers needed to treat and harm (NNT, NNH) statistics. My article,1 on which they comment, primarily sought to present a simple, easy-to-understand explanation for the calculation of these statistics, describe how to interpret them, and illustrate their applications. However, my article also sought to discuss the limitations of these statistics because just as the P value has its uses and limitations, so too do the NNT and NNH. Clinicians would be doing themselves a disservice if they adopted the NNT and NNH as evidence-based mental health evaluation tools without understanding their limitations. In this regard, Drs Citrome and Ketter offer excellent suggestions for the additional information that should accompany the presentation of the NNT and NNH so that these statistics can be properly interpreted. I support their suggestions with enthusiasm and hope that the suggestions will be adopted by the international community so that the limitations that I described1 can be offset.

The above notwithstanding, I retain my personal concern that the NNT and NNH have limited value for the practicing clinician. This is because what the practicing clinician really needs to know are the actual response and adverse event rates with drug and placebo. These rates are clinically meaningful, informative, easier to understand, and (most important of all) less likely to be misunderstood than statistics that describe how many patients need to be treated for 1 “extra” patient to be benefited or harmed. Furthermore, if these rates are provided, as Drs Citrome and Ketter suggest, then the NNT and NNH statistics become superfluous because they are derived from the stated rates.

On a parting note, I accept the concern that Drs Citrome and Ketter express about clinician unfamiliarity with differences in means on clinical ratings scales and hence the practical need for dichotomized outcomes such as response and remission.

Reference

1. Andrade C. The numbers needed to treat and harm (NNT, NNH) statistics: what they tell us and what they do not. J Clin Psychiatry. 2015;76(3):e330-e333. doi:10.4088/JCP.15f09870 PubMed

Chittaranjan Andrade, MD

andradec@gmail.com

Author affiliations: Department of Psychopharmacology, National Institute of Mental Health and Neurosciences, Bangalore, India.

Potential conflicts of interest: Included with Clinical and Practical Psychopharmacology columns at Psychiatrist.com (http://www.psychiatrist.com/documents/andrade-financial-disclosure.pdf).

Funding/support: None reported.

J Clin Psychiatry 2015;76(9):e1137

dx.doi.org/10.4088/JCP.15lr10001a

© Copyright 2015 Physicians Postgraduate Press, Inc.

Volume: 76

Quick Links: Psychiatry

References