Work

Robustness of positive findings and associated factors in acupuncture RCTs: exemplified by trials for lumbar disc herniation

Year: 2026
Type: study-registration
Abstract: This study examines the statistical fragility of significant findings from randomized controlled trials of acupuncture for lumbar disc herniation. Its purpose is to assess how many changes to individual participants’ outcomes would be needed to reverse statistical significance and to explore whether fragility is associated with study and acupuncture treatment characteristics. The study evaluates the robustness of reported findings; it is not designed to estimate the overall effectiveness of acupuncture. We searched PubMed, Embase, the Cochrane Library, Web of Science, CNKI, Wanfang Data, VIP Database, and SinoMed from inception to September 30, 2026. Eligible studies are two-arm, parallel-group randomized controlled trials involving adults with lumbar disc herniation. Interventions include manual acupuncture or electroacupuncture, used alone or alongside conventional treatment; eligible comparators include conventional treatment, sham acupuncture, or no treatment. Reviewers independently screen records and extract study, treatment, and outcome data. Risk of bias is assessed for selected outcome results using the Cochrane Risk of Bias tool. For statistically significant dichotomous outcomes, we calculate the fragility index (FI) using Fisher’s exact test. For continuous outcomes, we estimate the continuous fragility index (CFI) from reported group sample sizes, means, and standard deviations using simulated individual-level data and Welch’s t-test. FI and CFI are summarized separately using medians, interquartile ranges, and ranges. Where more than 20 studies are available for an index, linear regression is used to explore associations with study characteristics, first in univariable models and then in models adjusted for total sample size. The expected outputs are a description of the distribution of FI and CFI values, their relationships with study characteristics where analysis is feasible, and an assessment of how these findings should be interpreted alongside risk of bias.
Language: English
Field: Medicine
Open Access status: green
OpenAlex ID: W7219810179
Cites:
Cited by:
Related to: 10