Open access Original research BMJ Open: first published as 10.1136/bmjopen-2019-035168 on 12 April 2020. Downloaded from Definitions of blinding in randomised controlled trials of interventions published in high- impact anaesthesiology journals: a methodological study and survey of authors Antonija Penić,1 Dinka Begić,2 Karolina Balajić,3 Martin Kowalski,3 Ana Marušić ,3 Livia Puljak 4 To cite: Penić A, Begić D, ABSTRACT Strengths and limitations of this study Balajić K, et al. Definitions Objectives To analyse the completeness of reporting of blinding in randomised of blinding in randomised controlled trials (RCTs) of ► We analysed a large group of randomised controlled controlled trials of interventions interventions in anaesthesiology, the actual blinding published in high- impact trials published in seven high-impact anaesthesiol- status of various persons associated with an RCT and trial anaesthesiology journals: a ogy journals. authors’ interpretation of blinding terminology related to methodological study and ► Reporting about blinding was insufficient in the ma- RCTs. survey of authors. BMJ Open jority of analysed trials. 2020;10:e035168. doi:10.1136/ Methods This was a methodological study and a ► The term ‘double- blind’ is ambiguous and research- cross- sectional survey. We analysed reporting related to bmjopen-2019-035168 ers interpret it in different ways. blinding in published RCTs of interventions published in ► Prepublication history and ► We focused on anaesthesiology trials as a specific seven highly cited anaesthesiology journals from 2014 additional material for this research field. to 2016 and registered protocols in ClinicalTrials. gov. We paper are available online. To ► Future studies should monitor whether reporting surveyed corresponding authors of included RCTs about view these files, please visit about blinding in trials will improve. the journal online (http:// dx. doi. their definitions of blinding. The primary outcome was the org/ 10. 1136/ bmjopen- 2019- number of RCTs that explicitly described who was blinded 035168). in a trial. Secondary outcomes were definitions of blinding terminology in the trials; trial authors’ interpretation observational studies and studies with a Received 24 October 2019 of blinding terminology; discrepancies in the blinding randomised experimental design, that is, http://bmjopen.bmj.com/ Revised 20 January 2020 description within registered protocols and between randomised controlled trials (RCTs).1 RCTs Accepted 17 March 2020 registered protocols and publications. are considered the highest level of primary Results Out of 622 analysed RCTs, 38% were not evidence for analysing medical interventions explicitly described as either open label or blinded studies because randomisation ensures the alloca- and 10% did not report any information about blinding tion of participants by chance, and not by or lack of blinding. Only one manuscript fully reported choice, which minimises bias.2 Apart from the the status of blinding for various individuals that may be method of allocating participants into study © Author(s) (or their involved with a trial. The most common descriptor was that arms, there are other methodological aspects on September 27, 2021 by guest. Protected copyright. employer(s)) 2020. Re- use a trial was double- blind. We found discrepant information permitted under CC BY- NC. No regarding blinding in the majority of registered protocols. of an RCT that can contribute to a risk of commercial re- use. See rights Even when there were no discrepancies in the registration, bias, including performance bias (blinding and permissions. Published by we found discrepancies in the reporting of blinding of participants and personnel) and detection BMJ. between the majority of registered protocols and published bias (blinding of outcome assessors).3 1 Vujanovic Pharmacy, Sinj, manuscripts. The survey of authors (40 responses from Blinding is very important for the validity of Croatia 231 eligible authors; 17% response rate) of analysed RCTs 2Department of Endocrinology, RCT results. Multiple studies reported signifi- showed that they differed in how they defined different cant differences in the effect size of treatment University Hospital Tuzla, Tuzla, levels of blinding in trials. Bosnia and Herzegovina estimates depending on the use of blinding 3 Conclusions Reporting of the blinding status of key Department of Reasearch of key individuals.4–8 individuals involved in analysed anaesthesiology RCTs in Biomedicine and Health, The term ‘double- blind’ is sometimes taken University of Split School of was insufficient. For reporting guidelines, peer reviewers 9 Medicine, Split, Croatia and editors should insist on clear information on who was as a token of the validity of an RCT. However, 4Center for Evidence- Based blinded in a trial instead of using the term ‘double- blind’ it has been shown that the definitions of Medicine and Health Care, for different blinding practices. blinding may vary. Devereaux et al showed Catholic University of Croatia, that both physicians and textbooks provided Zagreb, Province, Croatia highly inconsistent interpretations of the Correspondence to INTRODUCTION blinding terminology used in clinical trials Dr Livia Puljak; The available evidence about the efficacy and and suggested that the current ambiguous livia. puljak@ gmail. com safety of interventions in medicine includes terminology should be replaced by specific Penić A, et al. BMJ Open 2020;10:e035168. doi:10.1136/bmjopen-2019-035168 1 Open access BMJ Open: first published as 10.1136/bmjopen-2019-035168 on 12 April 2020. Downloaded from descriptions of who was blinded.10 According to the reference management software (Clarivate Analytics, Consolidated Standards of Reporting Trial (CONSORT) Boston, Massachusetts, USA). Two authors independently checklist for reporting RCTs, the following should be analysed titles and abstracts to screen them for eligibility. reported for blinding: ‘If done, who was blinded after We retrieved potentially eligible full- text manuscripts and assignment to interventions (eg, participants, care two authors screened them independently. We resolved providers and those assessing outcomes) and how’.11 any disagreements through discussion. The aim of this study was to analyse the completeness of the reporting of blinding in RCTs of interventions in the Outcomes field of anaesthesiology, the actual blinding status of the The primary outcome was the number of RCTs that various persons involved in an RCT, information about explicitly described who (which individuals or group/s of the blinding reported in the registered trial protocol and individuals) was blinded in a trial. Secondary outcomes the trial authors’ interpretation of the blinding termi- were the definitions of the blinding terminology in the nology related to RCTs. The study focused specifically on trials (how the authors defined‘single-blinding’, ‘double- the field of anaesthesiology because pain is a subjective blinding’ and ‘triple- blinding’ in the manuscript), trial outcome and there is no objective method for analysing authors’ interpretation of the blinding terminology pain. The analgesic placebo effect has been well docu- related to RCTs, as well as discrepancies in the blinding mented.12 In the area of placebo effects, expectations play description within registered protocols, and between a major role, triggering a cascade of endogenous opioids registered protocols and publications. and non- opioids, and altering the experience of pain.12 For this reason, the field of anaesthesiology and pain Data extraction research is likely biased if blinding is not used adequately. We extracted data into a Microsoft Excel sheet (Microsoft, Redmond, Washington, USA) after the data extraction table was piloted with 10 studies. Data from each trial METHODS were extracted independently by two authors (three Study design authors participated in data extraction: AP, KB and MK). We conducted a methodological cross-sectional study of Independent extractions were compared for consistency published RCTs and registered data from their protocols and discrepancies were resolved via consensus. in public trial registries. Additionally, we surveyed the For each publication, we analysed the abstract, the trials’ corresponding authors. full body of the manuscript (and online supplementary files if the authors indicated that additional details about Ethics the methods were in online supplementary files) to see The research protocol for the survey of trials’ corre- whether and in which part of the manuscript, there was http://bmjopen.bmj.com/ sponding authors was approved by the Ethics Committee an explicit mention that the trial was blinded/masked. of the University of Split, School of Medicine (Klasa: We then analysed the Methods section of each included 003-08/18-03/0001, Ur. br.: 2181-198-03-04-18-0049; date: RCT and extracted information about the blinding termi- October 12, 2018). All invited study participants received nology (‘single- blind’, ‘double- blind’, ‘triple- blind’ or no information that the study was approved by the Ethics such definition) and whether the trial authors explicitly Committee. The participants’ continuation of the survey elaborated on who was blinded in a trial. from the invitation email was considered as consent to We checked whether the RCTs explicitly mentioned participate in the study. The data were fully anonymised using the CONSORT checklist for reporting or mentioned on September 27,