Systematic review of blinding assessment in randomized controlled trials in schizophrenia and affective disorders 2000-2010

Abstract

Background: Blinding is an integral part of many randomized controlled trials (RCTs). However, both blinding and blinding assessment seem to be rarely documented in trial reports.

Method: Systematic review of articles on RCTs in schizophrenia and affective disorders research during 2000-2010.

Results: Among 2,467 publications, 61 (2.5%; 95% confidence interval: 1.9-3.1%) reported assessing participant, rater, or clinician blinding: 5/672 reports on schizophrenia (0.7%; 0.3-1.6%) and 33/1,079 (3.1%; 2.1-4.2%) on affective disorders, without significant trends across the decade. Rarely was blinding assessed at the beginning, in most studies assessment was at the end. Proportion of patients' and raters' correct guesses of study arm averaged 54.4 and 62.0% per study, with slightly more correct guesses in treatment arms than in placebo arms. Three fourths of responders correctly guessed that they received the active agent. Blinding assessment was more frequently reported in papers on psychotherapy and brain stimulation than on drug trials (5.1%, 1.7-11.9%, vs. 8.3%, 4.3-14.4%, vs. 2.1%, 1.5-2.8%). Lack of assessment of blinding was associated with: (a) positive findings, (b) full industrial sponsorship, and (c) diagnosis of schizophrenia. There was a moderate association of treatment success and blinding status of both trial participants (r = 0.51, p = 0.002) and raters (r = 0.55, p = 0.067). Many RCT reports did not meet CONSORT standards regarding documentation of persons blinded (60%) or of efforts to match interventions (50%).

Conclusions: Recent treatment trials in major psychiatric disorders rarely reported on or evaluated blinding. We recommend routine documentation of blinding strategies in reports.