Abstract 12410: “Positive Spin” in Cardiovascular Research Publications: How Are Trials Without Statistically Significant Primary Outcomes Presented?

• Published in: Circulation (New York, N.Y.) N.Y.), 2018-11, Vol.138 (Suppl_1 Suppl 1), p.A12410-A12410
• Main Authors: Khan, Muhammad S, Rehman, Karim A, Lateef, Noman, Riaz, Haris, Khan, Safi U, Siddiqi, Tariq J, Fatima, Kaneez, Doukky, Rami, Krasuski, Richard
• Format: Article
• Language: English
• ISSN: 0009-7322; 1524-4539

BackgroundClinical researchers are obligated to present results both accurately and objectively. In case of studies where primary results are not statistically significant, placing a “positive spin” (hereafter “spin”) can distract the reader and lead to misinterpretation and misapplication of the findings. We sought to determine the frequency and extent of spin in cardiovascular research trials.MethodsWe searched Medline from January 2015 to December 2017 for all cardiovascular randomized controlled trials (RCTs) published in six major journals (New England Journal of Medicine, The Lancet, JAMA, European Heart Journal, Circulation and Journal of American College of Cardiology). Only parallel group RCTs with primary outcomes that were not statistically significant were included. Spin was defined a priori according to established objective criteria (Figure). Data were extracted and verified by 3 independent investigators using a pre-tested standardized collection form.ResultsOf the 1166 articles reviewed, 93 were eligible for inclusion. Spin was identified in 53 (57%) of the full texts and 62 (66.7%) of abstracts. The title included spin in 10 trials (10.8%). Spin was observed in 35 (37.6%) of the results and 47 (50.5%) of the conclusions of full texts, and in 38 (40.9%) of the results and 41 (44.1%) of the conclusions of abstracts. 26 (28%) trials had some spin in all sections of the abstract and 18 (19.4%) in all sections of the full text.ConclusionAmong cardiovascular publications reporting and interpretation of findings is frequently inconsistent with actual results. This makes it critical for readers to be educated in properly interpreting clinical study results.