Testing the success of blinding in randomised clinical trials

Article type
Authors
Hrobjartsson A, Forfang E, Haahr M, Als-Nielsen B, Stig B
Abstract
Background: Blinding participants in randomised clinical trials is regarded an important procedure for minimising bias. Double-blind trials tend to report smaller effects than similar trials that are not double-blind [1]. Also, blinded outcome evaluators assess outcomes less optimistically than unblinded evaluators [2]. In most clinical trials the success of the blinding procedures is assumed, but not tested. Earlier reports on the success of blinding have mainly been based on placebo controlled trials published in selected journals. The number of identified trials has been low, and several did not report the full details of the test of the success of blinding [3, 4].

Objectives: 1. To assess how often the success of blinding is tested in randomised clinical trials, 2. to describe the methods used, and 3. to assess the frequency of trials with successful blinding.

Methods: In May 2003 we searched The Cochrane Central Register of Controlled Trials for references to randomised clinical blinded trials published in 2001. Our search terms were: random* and (blind* or mask* or placebo* or sham* or mock* or fake* or dummy* or vehicle*) and publication year = 2001. We identified 5,079 references. Based on a preliminary analysis of 500 references and the aim of identifying 25 eligible trials we randomly sampled 2,250 references. We retrieved the full paper versions of blinded randomised trials, excluding abstracts, conference proceedings and non-clinical trials. Trials that reported to have tested the success of blinding were selected. In the case of incomplete data on the results of the test we contacted the authors.

Results: We excluded 622 abstracts, conference proceedings, and published studies that were either not blinded, randomised or clinical trials. We identified 1,602 randomised blinded clinical trials, and are in the process of retrieving the last 26 trial reports. We have identified 30 eligible trials that test the success of blinding. Nine trials presented full data sets of at least one group of participants, additional data sets were obtained from a further 11 trials, and authors of the last 10 trials are still being contacted.

Conclusions: Will be presented at the colloquium

References: 1. Juni P, Altman DG, Egger M. Systematic reviews in health care: Assessing the quality of controlled clinical trials. BMJ 2001;323:42-6. 2. Noseworthy JH, Ebers GC, Vandervoort MK, Farquhar RE, Yetisir E, Roberts R. The impact of blinding on the results of a randomized, placebo-controlled multiple sclerosis clinical trial. Neurology 1994;44:16-20. 3. Ney PG, Collins C, Spensor C. Double blind: double talk or are there ways to do better research. Med Hypotheses 1986;21:119-26. 4 Fergusson D, Glass KC, Waring D, Shapiro S. Turning a blind eye: the success of blinding reported in a random sample of randomised, placebo controlled trials. BMJ 2004;328:432-4.