Spin in the Titles and Abstracts of Allergy and Immunology Randomized Controlled Trials With Nonsignificant Outcomes

TL;DR

This study finds that over half of allergy and immunology trial abstracts with nonsignificant results use misleading language that suggests positive outcomes.

Contribution

The study quantifies the prevalence of spin in allergy and immunology RCT abstracts with nonsignificant outcomes, revealing high rates of misleading presentation.

Findings

53% of 66 eligible abstracts contained spin in their title, results, or conclusion.

85.7% of spinned abstracts had misleading language in the conclusion section.

No significant difference in spin prevalence was found between industry-sponsored and non-industry-sponsored trials.

Abstract

In scientific reporting, ‘spin’ refers to presenting neutral or negative outcomes in a manner that infers favorable results. Spin can be especially problematic in the abstracts of randomized control trials, leading to an incorrect interpretation of trial outcomes, and potentially impacting subsequent studies, patient care, and policy decisions. The present study aims to evaluate the presence of spin in the abstracts of registered randomized control trials published in allergy and immunology journals. A systematic search of the PubMed database was performed to identify randomized controlled trials of human subjects registered in a clinical trial registry with a test and control group and a nonsignificant primary endpoint. A total of 1,248 articles were screened and 66 abstracts met full inclusion criteria. Thirty-five of the 66 (53%) abstracts were found to contain one or more elements of spin, among which 11 (31.4%) had spin in the title, 29 (82.9%) in the abstract results, and 30 (85.7%) in the abstract conclusion. Industry-sponsored trials did not contain more spin compared to other funding sources (p=0.62). High rates of spin may adversely affect the interpretation and integration of new research. Careful evaluation is recommended when reviewing abstracts lacking statistically significant primary outcomes.

Figures