Advertisement

Pharmaceutical Medicine

, Volume 24, Issue 3, pp 145–149 | Cite as

Misleading Statistics

The Problems Surrounding Number Needed to Treat and Number Needed to Harm
  • Jane L. Hutton
Current Opinion

Abstract

Collection of good-quality clinical data is expensive. It is important to choose methods for data analysis and presentation of results that allow clear assessment. For studies that compare rates of infection, or other adverse or beneficial outcomes, there are several ways in which the results can be summarized. The difference between the rates in two groups, say new treatment versus standard, can be used; this can be expressed as the absolute risk reduction (ARR). The ratio of rates, the relative risk (RR), is often used in epidemiological and survival analyses. Odds ratios (ORs) and log ORs are not so easy to understand but are useful in the analysis stage of research.

A statistic called the ‘number needed to treat’ (NNT) has been proposed, and is now included in some textbooks of Pharmaceutical Medicine and used in research articles and guidelines. The NNT is the inverse of the difference in rates and is usually expressed as a whole number. If the difference between the infection rates on two treatments is 17%, then 100÷17 = 6 is the NNT.

Other desirable aspects of a summary statistic were given by the authors who introduced NNT. It is important to be able to compare treatments, both in terms of benefits and harms and to identify patients who are most likely to benefit or to experience adverse effects. Subsequent claims were that NNT can be used to extrapolate research findings, is easy to calculate and allows the cost of treatment to be assessed. However, the claim that onlyNNTcan provide a measure of benefit, harm and cost is simply false.Afurther issue, that RR and ORs combine two rates into one summary statistic, applies to NNT as well.

An important problem with NNT relates to its supposed main advantage: it is claimed that it provides a measure in terms of patients treated rather than in terms of probabilities. However, this is not true. An NNT gives the average number of patients, among whom, if they were treated with one therapy rather than another, exactly one patient will benefit. The NNT statistic is biased, and reliable confidence intervals cannot be provided. Furthermore, there is no simple value that indicates no difference between treatments. For meta-analysis, NNT cannot be used directly because simple arithmetic, such as addition, does not give correct results on the NNT scale. If the baseline rate, e.g. the mortality rate on standard treatment, is given, some transformations are required to be able to find the mortality rate on a new treatment using the NNT. If the risk reduction, the difference in rates, is given, the rate on new treatment is found by subtraction.

The major claim for NNT was that it is easy to understand, but this claim has been refuted by observational and experimental studies. It is best, for both theoretical and practical reasons, to use the baseline risk and the difference in success or event rates, the ARR, to present results.

Keywords

Relative Risk Fluconazole Capsaicin Itraconazole Baseline Rate 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Notes

Acknowledgements

No sources of funding were used to assist in the preparation of this review. The author has no conflicts of interest that are directly relevant to the content of this review.

References

  1. 1.
    Winston DJ, Maziarz RT, Chandrasekar PH, et al. Intravenous and oral itraconazole versus intravenous and oral fluconazole for long-term antifungal prophylaxis in allogeneic hematopoietic stem-cell transplant recipients: a multicenter, randomized trial. Ann Intern Med 2003; 138: 705–13PubMedGoogle Scholar
  2. 2.
    Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures of the consequences of treatment. N Engl J Med 1988; 318: 1728–33PubMedCrossRefGoogle Scholar
  3. 3.
    Cook RJ, Sackett DL. The number needed to treat: a clinically useful measure of treatment effect. BMJ 1995; 310: 452–4PubMedCrossRefGoogle Scholar
  4. 4.
    Sackett DL, Deeks JJ, Altman DG. Down with odds ratios. Evid Based Med 1996; 1: 164–6Google Scholar
  5. 5.
    Altman DG. Confidence intervals for the number needed to treat. BMJ 1998; 317: 1309–12PubMedCrossRefGoogle Scholar
  6. 6.
    Mason L, Moore RA, Derry S, et al. Systematic review of topical capsaicin for the treatment of chronic pain. BMJ 2004; 328: 991PubMedCrossRefGoogle Scholar
  7. 7.
    CONSORT. A proposal for structured reporting of randomized controlled trials: the Standards of Reporting Trials Group [published erratum appears in JAMA 1995; 273: 776]. JAMA 1994; 272: 1926–31CrossRefGoogle Scholar
  8. 8.
    Hutton JL. Number needed to treat: properties and problems. J R Stat Soc Ser A 2000; 163: 403–19CrossRefGoogle Scholar
  9. 9.
    Chatellier G, Zapletal E, Lemaitre D, et al. The number needed to treat: a clinically useful nonogram in its proper context. BMJ 1996; 312: 426–9PubMedCrossRefGoogle Scholar
  10. 10.
    Christensen PM, Kristiansen IS. Number-needed-to-treat (NNT): needs treatment with care. Basic Clin Pharmacol Toxicol 2006; 99: 12–6PubMedCrossRefGoogle Scholar
  11. 11.
    Lesaffre E, Pledger G. A note on the number needed to treat. Control Clin Trials 1999; 20: 439–47PubMedCrossRefGoogle Scholar
  12. 12.
    Tramer MR, Reynolds DJM, Moore RA, et al. When placebo controlled trials are essential and equivalence trials are inadequate. BMJ 1998; 317: 875–80PubMedCrossRefGoogle Scholar
  13. 13.
    Stang A, Poole C, Bender R. Common problems related to the use of number needed to treat. J Clin Epidemiol. Epub 2009 Oct 31Google Scholar
  14. 14.
    Stovring H, Gyrd-Hansen D, Kristiansen IS, et al. Communicating effectiveness of intervention for chronic diseases: what single format can replace comprehensive information? BMC Med Inform Decis Mak 2008; 8: 25PubMedCrossRefGoogle Scholar
  15. 15.
    Halvorsen PA, Kristiansen IS. Decisions on drug therapies by numbers needed to treat: a randomized trial. Arch Intern Med 2005; 165: 1140–6PubMedCrossRefGoogle Scholar
  16. 16.
    Guidelines on science and health communication. London: The Royal Institution and Social Issues Research Centre and The Royal Society, 2001Google Scholar

Copyright information

© Adis Data Information BV 2010

Authors and Affiliations

  1. 1.Department of StatisticsUniversity of WarwickCoventryUK

Personalised recommendations