The impact of blinding on estimated treatment effects in randomized clinical trials on acupuncture: A meta‐epidemiological study

Youlin Long,Na Zhang,Xinyao Wang,Ruixian Tang,Qiong Guo,Jin Huang,Liang Du
DOI: https://doi.org/10.1111/jebm.12589
2024-03-13
Journal of Evidence-Based Medicine
Abstract:Objective To evaluate the sole impact of blinding patients and outcome assessors in acupuncture randomized controlled trials (RCTs) on treatment effects while considering the type of outcome measures. Methods We searched databases for the meta‐analyses on acupuncture with both blinded and non‐blinded RCTs. Mixed‐effects meta‐regression models estimated the average ratio of odds ratios (ROR) and differences in standardized mean differences (dSMD) for non‐blinded RCTs versus blinded mixed‐effects meta‐regression model. Results The study included 96 meta‐analyses (1012 trials). The average ROR for lack of patient blinding was 1.08 (95% confidence intervals 0.79–1.49) in 18 meta‐analyses with binary patient‐reported outcomes. The average ROR for lack of outcome assessor blinding was 0.98 (0.77–1.24) in 43 meta‐analyses with binary subjective outcomes. The average dSMD was −0.38 (−0.96 to 0.20) in 10 meta‐analyses with continuous patient‐reported outcomes. The average dSMD was −0.13 (−0.45 to 0.18) in 25 meta‐analyses with continuous subjective outcomes. The results of the subgroup analysis were consistent with the primary analysis findings. Conclusions Blinding of participants and outcome assessors does not significantly influence acupuncture treatment efficacy. It underscores the practical difficulties of blinding in acupuncture RCTs and the necessity to distinguish between trials with and without successful blinding to understand treatment expectations' effects. Enhancing blinding procedures' quality and assessment in future research is crucial for improving RCTs' internal validity and reliability.
medicine, general & internal
What problem does this paper attempt to address?