Education And Debate

Post-randomisation exclusions: the intention to treat principle and excluding patients from analysis

BMJ 2002; 325 doi: http://dx.doi.org/10.1136/bmj.325.7365.652 (Published 21 September 2002) Cite this as: BMJ 2002;325:652
  1. Dean Fergusson (dafergusson@ohri.ca), scientista,
  2. Shawn D Aaron, assistant professora,
  3. Gordon Guyatt, professorb,
  4. Paul Hébert, associate professora
  1. a Ottawa Health Research Institute, 501 Smyth Road, Box 201, Ottawa, ON, Canada K1H 8L6
  2. b Departments of Clinical Epidemiology and Biostatistics and Medicine, McMaster University, Hamilton, ON, Canada L8N 3Z5
  1. Correspondence to: D Fergusson
  • Accepted 17 April 2002

When is it legitimate to exclude randomised patients from the analysis of data in clinical trials? Basing their analysis on the desirability of minimising bias and random error, the authors consider the circumstances when it may be possible to exclude patients, even in an intention to treat trial

Most clinical researchers and statisticians agree that the primary analysis of data in a randomised clinical trial should compare patients according to the group to which they were randomly allocated, regardless of patients' compliance, crossover to other treatments, or withdrawal from the study. Such an analysis is referred to as an intention to treat or an “as randomised” analysis. Proponents argue that the intention to treat approach

  • Helps preserve prognostic balance in the study arms

  • Limits inferences based on arbitrary or ad hoc subgroups of patients in the trial

  • Emphasises greater accountability for all patients entered into the study and consequently minimises the influence of withdrawals, non-compliers, and patients lost to follow up

  • Is the most cautious approach and so minimises type 1 error, and

  • Allows for the greatest generalisability.15

Critics say, however, that an intention to treat approach is too cautious and more susceptible to type II error. 6 7 They argue that such an analysis is less likely to show a positive treatment effect, especially in studies that randomise patients who have little or no chance of benefiting from the intervention. These critics maintain that an efficacy or explanatory approach to an analysis is more important than an effectiveness or pragmatic approach.

Experts have documented the strengths and weaknesses of the different analytical approaches.8 However, one issue that has only rarely been addressed in the literature is post-randomisation exclusions unrelated to non-compliance, withdrawal, or losses to follow up.9 These exclusions occur when patients are inappropriately …

Sign in

Subscribe