Spin occurs in bariatric surgery randomized controlled trials with a statistically nonsignificant primary outcome: A systematic review

Abstract

Objectives: To systematically identify the strategy and frequency of spin in reports of bariatric surgery randomized controlled trials (RCTs) with statistically nonsignificant primary endpoint.

Study design and setting: The use of specific reporting strategies to highlight the beneficial effect of an experimental treatment can affect the reader interpretation of trial results, particularly when the primary endpoint is not statistically significant. A literature search was performed to identify RCTs publications assessing the impact of bariatric surgery on obesity-related comorbidities published over the past 10 years (from January 2020 till December 2020) in MEDLINE and EMBASE. RCTs publications with statistically non-significant primary outcomes were included.

Results: Of 46 576 reports screened for title and abstract inclusion, 29 RCT reports met the inclusion criteria for spin analysis. In total, 16 abstracts (55%) and 18 main texts (62%) were classified as having a spin. In abstract results and conclusion sections, the spin was identified in 69% of reports. In main text results, discussion, and conclusion sections, the spin was recognized in 37%, 72%, and 76% of reports respectively. The spin consisted mainly of focusing on within-group improvements and the interpretation of statistically nonsignificant results as showing treatment equivalence.

Conclusion: Spin occurred in a high proportion of bariatric surgery RCTs with a statistically nonsignificant primary endpoint.

Keywords: Bariatric surgery; Randomized controlled trial; Spin.