Review ArticleA review of blinding in randomized controlled trials found results inconsistent and questionable
Introduction
In randomized controlled trials (RCTs), blinding is widely accepted to be an important methodological component for avoiding bias [1]. Studies not involving blinding yield exaggerated estimates of treatment effects [2], [3]. Blinding can be difficult to establish and maintain, however, and participants, care providers, or outcome assessors might be able to distinguish the placebo from treatment [4], [5]. For example, when assessing zinc treatment for the common cold, the blinding procedure could fail because the taste and aftertaste of zinc are distinctive [6], [7]. Although describing the blinding procedure in the published report provides some assurance that blinding was successful, the revised Consolidated Standards of Reporting Trials (CONSORT) [8] recommends reporting “how the success of blinding was evaluated.” Fergusson et al. [9], in an examination of reporting and success of blinding in a sample of RCTs, found that only 8% of the trials provided evidence of the success of blinding. The authors claimed that the success of blinding should be assessed and reported in all published blinded RCTs. Methods of assessing and reporting the success of blinding are debatable, however, and no clear guidelines for adequately reporting the success of blinding exist [10], [11]. Our goal for the present study was to use the MEDLINE and Cochrane databases to target studies that attempt to assess blinding and then to describe and discuss the methods used in those studies.
Section snippets
Search strategy and selection of reports for detailed study
The results of our search strategy are shown in Fig. 1. We searched MEDLINE, the Cochrane Controlled Trials Register and the Cochrane Method Register using the following terms: (double blind method[MeSH] OR single blind method[MeSH] OR research design[MeSH:NoExp]) AND ((guess∗ AND blind∗) OR blindness OR blindability) NOT guess[author]. Limits: English AND French.
We also searched the Cochrane Controlled Trials Register and Cochrane Method Register (The Cochrane Library, 2004, issue 1) using the
Selected reports
Reports were identified by electronic search (n = 207), manual search (n = 172), and searching references in the original articles (n = 23). Only 13 (7.6%) of the 172 RCTs identified by manual search mentioned an assessment of blinding success. Finally, 90 reports were included in the analysis. Reasons for excluding articles are given in Fig. 1.
Trial characteristics
Of the reports selected, 28 (31%) were from psychiatry, 13 (14%) from neurology and the remainder from a mix of respiratory and cardiovascular (n = 9),
Discussion
To our knowledge, this is the first systematic review of methods used to assess the success of blinding.
Our study confirmed that the success of blinding is rarely described in reports of RCTs [9] and showed that most reports assessed the success of blinding by asking subjects to guess their treatment allocation. Studies varied widely, however, in who was asked to guess (participants, care providers, outcome assessors), timing of assessment, answer modalities, statistical analyses, and reporting
Acknowledgments
The authors are very grateful to David Sackett for his comments and help in improving this report. This work was supported by a grant from the Caisse Nationale d'Assurance Maladie des Professions Indépendantes (CANAM) and Assistance Publique des Hôpitaux de Paris (AP-HP).
References (32)
What are the main methodological problems in the estimation of placebo effects?
J Clin Epidemiol
(2002)- et al.
Blinding was judged more difficult to achieve and maintain in non-pharmacological than pharmacological trials
J Clin Epidemiol
(2004) - et al.
In the dark: the reporting of blinding status in randomized controlled trials
J Clin Epidemiol
(2002) - et al.
Blinding in randomised trials: hiding who got what
Lancet
(2002) Evaluating preference effects in partially unblinded, randomized clinical trials
J Clin Epidemiol
(2003)- et al.
Assessment of blinding in clinical trials
Control Clin Trials
(2004) - et al.
Matching quality of agents employed in “double-blind” controlled clinical trials
Lancet
(1976) - et al.
Introducing a placebo needle into acupuncture research
Lancet
(1998) - et al.
Empirical evidence of bias: dimensions of methodological quality associated with estimates of treatment effects in controlled trials
JAMA
(1995) - et al.
The impact of blinding on the results of a randomized, placebo- controlled multiple sclerosis clinical trial
Neurology
(1994)