ViewpointImplications of trial results: the potentially misleading notions of number needed to treat and average duration of life gained
Section snippets
Number needed to treat
For acute conditions, let us consider thrombolysis for acute myocardial infarction, or antibiotics for pneumonia. Such treatments last for a fixed time. Mortality tends to cluster in time, such that trials can have a follow-up long enough to observe the outcome of the acute episode. NNTs can be calculated by taking the reciprocal of the difference in risk of death between the untreated and treated arm at some time point after randomisation1, 3 (see appendix I).* For example, in ISIS-2 the
Average duration of life gained
If we imagine a placebo-controlled trial in which all patients were followed until death while using assigned treatment, the ADLG would be the difference in mean survival between treated and untreated patients. Survival curves for such a trial would appear as shown in figure 2. As the area under each curve is equal to the mean survival time for each treatment arm, the surface area between the two curves is equal to the ADLG.
Sometimes the ADLG can be obtained directly from published data without
Discussion
No uniformity presently exists in the calculation and presentation of NNTs. For acute conditions with a fixed duration of treatment, NNTs based on differences in risk and expressed as numbers of patients do not materially depend on the duration of follow-up as long as follow-up exceeds the acute phase. For lifelong treatments of chronic conditions, NNTs should not depend on duration of follow-up either. This can be achieved by basing NNTs on differences in hazard and expressing them as
References (9)
- et al.
Shifts in mortality curves: saving or extending lives?
Lancet
(1999) - et al.
Number-needed-to-treat and placebo-controlled trials
Lancet
(1998) - et al.
The number needed to treat: a clinically useful measure of treatment effect
BMJ
(1995) - et al.
An assessment of clinically useful measures of the consequences of treatment
N Engl J Med
(1988)
Cited by (50)
The averted infections ratio: a novel measure of effectiveness of experimental HIV pre-exposure prophylaxis agents
2018, The Lancet HIVCitation Excerpt :An AIR equal to 1 implies that the two agents are equally effective, an AIR of less than 1 implies that the experimental agent is less effective than the control agent, and an AIR greater than 1 implies that the experimental agent is more effective than the control. Alternative interpretations of the AIR are the ratio of the number needed to treat (to prevent one infection) for the active-control agent relative to the number needed to treat for the experimental agent,12 or the ratio of the effectiveness of the experimental agent relative to the effectiveness of the active-control agent. The AIR can be expressed as
Risk-difference curves can be used to communicate time-dependent effects of adjuvant therapies for early stage cancer
2014, Journal of Clinical EpidemiologyEstimation of numbers needed to treat should be based on absolute risks
2014, Journal of Clinical EpidemiologyAbsolute risks rather than incidence rates should be used to estimate the number needed to treat from time-to-event data
2013, Journal of Clinical EpidemiologyCitation Excerpt :The resulting NNTRD(t) is interpreted as the average number of patients needed to be treated to observe one more event-free patient in the treatment group compared with the control group at time point t. A second method was proposed by Lubsen et al. [11] and Mayne et al. [12], independently of each other. Herein, the NNT is estimated by means of the reciprocal of the difference of two incidence rates, that is, the number of events per person-years in the two groups (ID approach).
Authors' reply
2012, The LancetCommon problems related to the use of number needed to treat
2010, Journal of Clinical Epidemiology