Abstract: The number needed to treat (NNT) is widely used to present the results of clinical trials. However, incorrect calculations, misuse in specific study situations and misleading interpretation of NNTs are not uncommon in practice, for example in the case of non-significant results or in the situation of survival time data.
This workshop will provide an introduction to the use and interpretation of NNTs as effect measure in clinical trials and systematic reviews. It will be shown how to calculate NNTs with corresponding confidence intervals from Summary of Findings (SoF) tables. We will discuss in particular the limitations of NNTs including problems associated with the inconvenient scale, the sensitivity to changes of the baseline risk and the dependence on the follow-up time.
Small group practical work will be an integral part of the workshop focusing on adequate calculation and interpretation of point and interval estimates of NNTs in practical examples. Participants should bring pocket calculators, pens and paper.
Reference:
Bender, R. (2005): Number needed to treat (NNT). In: Armitage, P. & Colton, T. (Eds.): Encyclopedia of Biostatistics, 2nd Ed., Vol. 6, pp. 3752-3761. Wiley, Chichester.