Blinding in Physical Therapy Trials and Its Association with Treatment Effects: A Meta-epidemiological Study. (January 2017)

• Record Type: Journal Article
• Authors: Armijo-Olivo, Susan; Fuentes, Jorge; da Costa, Bruno R.; Saltaji, Humam; Ha, Christine; Cummings, Greta G.
• Is Part Of: American journal of physical medicine & rehabilitation. Volume 96:Number 1(2017)
• Publication Date: 2017-01
• Subjects: Blinding -- Physical Therapy -- Randomized Controlled Trials -- Risk of Bias; Rehabilitation -- Periodicals; Medicine, Physical -- Periodicals; 617.062
• DOI: 10.1097/PHM.0000000000000521
• Languages: English
• ISSNs: 0894-9115

Abstract

Abstract : Objective: The aim of this study was to examine whether blinding of participants, assessors, health providers, and statisticians have an effect on treatment effect estimates in physical therapy (PT) trials. Design: This was a meta-epidemiological study. Randomized controlled trials in PT were identified by searching the Cochrane Database of Systematic Reviews for meta-analyses of PT interventions. Assessments of blinding in PT trials were conducted independently following established guidelines. Results: Three hundred ninety-three trials and 43 meta-analyses that included 44, 622 patients contributed to this study. Only a quarter of the trials were adequately blinded (n = 80; 20%). Most individual components of blinding as well as what they were blinded to were also poorly reported. Although trials with inappropriate blinding of assessors and participants tended to underestimate treatment effects when compared with trials with appropriate blinding of assessors and participants, the difference was not statistically significant (effect size, −0.07; 95% confidence interval, −0.22 to 0.08; effect size, −0.12; 95% confidence interval, −0.30 to 0.06, respectively). Conclusions: The lack of statistical significance between blinding and effect sizes should not be interpreted as meaning that an impact of blinding on effect size is not present in PT. More empirical evidence in a larger sample is needed to determine which biases are likely to influence reported effect sizes of PT trials and under which conditions. Abstract : Supplemental digital content is available in the text.