Family Practice Vol. 19, No. 5, 566
© Oxford University Press 2002
NNT is not easily understood
Institute of Public Health University of Southern Denmark Odense Denmark E-mail: isk{at}sam.sdu.dk
In a recent editorial, Misselbrook and Armstrong comment upon the concept of risk in relation to John Everyman1 and claim that the number needed to treat (NNT) for Mr Everyman is 11. It is not! If the baseline risk of cardiovascular disease (CVD) events is 27.5% (‘25–30%’) and 24.5% (‘22–27%’) after treatment, the absolute risk reduction is 27.5 – 24.5% = 3% and the NNT = 1/0.03 = 33. The relative risk reduction is (27.5 – 24.5) x 100/27.5 which is ~11% and not 9%.
The authors state that "the ‘personal probability of benefit’ (PPB) may make the probability as meaningful as possible to the individual patient".1 In their example, the NNT is 11, and %PPB is claimed to be 1/NNT x 100 = 9%. This implies, according to Misselbrook and Armstrong, that the patient "would have a 9% chance of benefiting from treatment". Unfortunately, this interpretation is not correct.
For interventions aimed at postponing adverse events (e.g. CVD events, fractures, death) in chronic diseases (e.g. hypertension, osteoporosis), risk estimates, whether they be NNT, relative risk reduction or others, only seldom inform about the individual person’s probability of having benefit from the intervention. A simple example may clarify the point. In a hypothetical trial, 52 patients in the control group die one by one with an interval of 1 week. In an identical intervention group, death is postponed by 1 week in all patients. Whenever the absolute risk reduction is measured during the trial, one more patient is alive in the intervention group, and the absolute risk reduction is 1/52. Consequently, NNT is 52 but the ‘personal probability of (some) benefit’ is 100%.
More generally, the proportion of patients having benefit from interventions cannot be inferred from risk or survival analysis because small postponements in many patients can create the same risk reductions or survival curves as large postponements in a few patients. Lay people2,3 and doctors4 misunderstand NNT in this respect. Interestingly, advocates of ‘evidence-based medicine’ fall victim to the same misunderstanding when they promote the use of NNT in medical decision making.5–7
References
1 Misselbrook D, Armstrong D. Thinking about risk. Can doctors and patients talk the same language? Fam Pract 2002; 19: 1–2.
2 Hux JE, Naylor CD. Communicating the benefits of chronic preventive therapy: does the format of efficacy data determine patients’ acceptance of treatment? Med Decis Making 1995; 15: 152–157.
3 Misselbrook D, Armstrong D. Patients’ responses to risk information about the benefits of treating hypertension. Br J Gen Pract 2001; 51: 276–279.[Web of Science][Medline]
4 Nexoe J, Gyrd-Hansen D, Kragstrup J, Kristiansen IS, Nielsen JB. Danish GPs’ perception of disease risk and benefit of intervention. Fam Pract 2002; 19: 3–6.
5 Sackett DL, Richardson WS, Rosenberg W, Haynes RB. Evidence-based Medicine. How to Practice and Teach EBM. New York: Churchill Livingstone, 2000.
6 McAlister FA, Straus SE, Guyatt GH, Haynes RB. Users’ guides to the medical literature: XX. Integrating research evidence with the care of the individual patient. Evidence-Based Medicine Working Group. J Am Med Assoc 2000; 283: 2829–2836.
7 Sinclair JC, Cook RJ, Guyatt GH, Pauker SG, Cook DJ. When should an effective treatment be used? Derivation of the threshold number needed to treat and the minimum event rate for treatment. J Clin Epidemiol 2001; 54: 253–262.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||