CONSORT 2010 changes and testing blindness in RCTs.

摘要

On The Lancet's website today and in several other medical journals, the CONSORT 2010 Statement chronicles many changes since the last CONSORT statement for reporting randomised trials. We have highlighted a few important examples (panel 1).One change, however, deserves further explanation. In CONSORT 2010, we have removed mention of how the success of blinding might have been evaluated. Our action runs contrary to views expressed by many investigators over the years, with a recent article nicely crystallising those views. Indeed, blinding in trials, when possible, can reduce bias. Many investigators attempt to blind all participants, investigators, and outcome assessors (often termed double-blinding, or double-masking in The Lancet when the term is applied to a person). When blinding is attempted, some investigators have recommended testing the success of those blinding efforts. Moreover, some, including one of us'(DF), had urged the next update of CONSORT to require such testing. The CONSORT Executive (KFS, DGA, and DM) decided to address the reasoning behind CONSORT not requiring testing of blinding, and our colleague (DF) joined in this Comment to clarify his views.