Manipulation of language in reporting of nonsignificant results for primary outcomes in CV RCTs

Introduction and methods

Results of randomized clinical trials (RCTs) form the basis of clinical guidelines [1] and therefore accurate presentation of trial results is very important. However, language in scientific articles to describe results can be used in a way to shape the impression of results for readers. ‘Spin is defined as the manipulation of language to potentially mislead readers from the likely truth of results’ [2-4].

In CV literature, there is no reporting of systematically identified and evaluated spin. Therefore, in this systematic review the prevalence and extent of spin was examined in a cohort of CV RCTs.

MEDLINE was searched for CV RCTs published from January 1, 2015, to December 31, 2017, in 6 high-impact journals: New England Journal of Medicine, The Lancet, JAMA, European Heart Journal, Circulation, and Journal of the American College of Cardiology. Only RCTs with parallel groups, clearly identified primary outcomes and statistically non-significant results for primary outcomes were included, resulting in 93 studies.

Two reviewers assessed spin type, severity and extent; classifications were based on the article by Boutron et al. [5]. In this study, spin was defined as the “use of specific reporting strategies, from whatever motive, to highlight that the experimental treatment is beneficial, despite a statistically nonsignificant difference for the primary outcome [ie, inappropriate use of causal language], or to distract the reader from statistically nonsignificant results [ie, to focus on a statistically significant secondary result]”[5]. Spin was classified as follows: (1) authors focused on statistically significant secondary results, (2) authors interpreted nonsignificant results to show treatment equivalence or to rule out an adverse event; and (3) authors focused on treatment effectiveness with or without acknowledging the statistically nonsignificant primary outcome. Level of spin was classified as high, moderate, or low. There were 4 categories to classify the extent of spin : (1) spin in 1 section other than the conclusions section, (2) spin in the conclusions section only,(3) spin in 2 sections, not including the abstract, (4) spin in all sections.

Main results

• Spin was observed in titles of 10 articles (11%, 95% CI:6-19%). In abstracts, spin was identified in 38 result sections (41%; 95%CI: 31-51%) and in 45 conclusions sections (48%, 95%CI: 38-58%). In the main text of articles, spin was observed in 35 results sections (38%, 95%CI: 28-48%), in 35 discussion sections (38%; 95%CI: 28-48%) and 50 conclusions sections (54%, 95%CI: 44-64%).
• In the abstract results sections, spin in 13 articles (14%, 95%CI: 8-22%) focused on statistically significant secondary outcomes, in 10 articles (11%, 95%CI: 6-19%) focused on statistically significant subgroup analyses. In conclusions section of abstract, spin in 17 articles (18%, 95%CI: 12-27%) focused on reporting treatment effectiveness.
• In the main texts, spin in result section was classified as presenting statistically significant secondary outcomes in 11 reports (12%, 95%CI: 7-20%) and in the conclusions section in 16 articles (17%, 95%CI: 11-26%) classified as focusing on treatment effectiveness.
• Spin in at least 1 section occurred in 53 abstracts (57%, 95%CI: 47-67%) and 62 main texts (67%, 95%CI: 57-76%). Spin in all sections was identified in 26 abstracts (28%, 95%CI; 20-38%) and 18 main texts (19%, 95%CI: 12-29%).
• If spin was identified, level of spin was mostly low.
• Conflict of interest disclosures of first and last authors was not correlated with spin. Citations per year and level of spin in abstract and main text was negatively correlated. There was no association between CV subspecialty, industry funding, or journal type with spin.

Conclusion

References

Find this article online at JAMA Network Open