A review of blinding in randomized controlled trials found results inconsistent and questionable

J Clin Epidemiol. 2005 Dec;58(12):1220-6. doi: 10.1016/j.jclinepi.2005.04.006. Epub 2005 Sep 30.

Abstract

Background and objective: To determine methods to assess the success of blinding in randomized controlled trials (RCTs).

Methods: We searched MEDLINE, the Cochrane Controlled Trials Register, and the Cochrane Method Register and performed a manual search to target studies that attempt to assess blinding and describe the methods used in those studies.

Results: A total of 90 reports were selected. Reports assessed the success of blinding participants (n = 58), care providers (n = 36), and outcome assessors (n = 15). Of the 58 reports assessing the success of blinding participants, 54 (93%) reported asking participants to guess their treatment assignment. There was no consistency in timing of assessment (e.g., once at the end of the trial, 57%, or several times during the trial, 26%) or modalities of answering (e.g., "do not know" answers, 43%, or participants forced to guess, 31%). A statistical analysis was performed in 57% of reports. The statistical analysis mainly compared the proportion of correct guesses to those produced by chance (32%) or checked for a relation between participants' guesses and treatment assignment (23%).

Conclusions: Methods of assessing the success of blinding, analysis and reporting the results were inconsistent and questionable.

Publication types

  • Research Support, Non-U.S. Gov't
  • Review

MeSH terms

  • Bibliometrics*
  • Data Interpretation, Statistical
  • Double-Blind Method*
  • Humans
  • Randomized Controlled Trials as Topic / methods
  • Randomized Controlled Trials as Topic / standards*
  • Sensitivity and Specificity