What are the rates and validity of French authors' conflicts of interest disclosures in Orthopaedics & Traumatology: Surgery & Research? Analysis of the year 2017 with comparison versus the Transparence.sante.gouv.fr database.

INTRODUCTION

Authors are responsible for their own Disclosure of Interests in submitting to Orthopaedics & Traumatology: Surgery & Research (OTSR). Disclosure may be incomplete or false, by deliberate or unintended omission. Omission may impact the analysis of results or at least the reader's perception of it. The rate of validity of disclosure in OTSR is not known, and we therefore conducted a bibliometric study on articles published in 2017, in order to: 1) determine the proportion of articles in which authors disclose interests; and 2), when interests are disclosed, to assess their validity by comparison against the Transparence.gov.fr database.

HYPOTHESIS

Disclosure rates exceed 75%.

MATERIAL AND METHOD

We analyzed all articles published in OTSR in 2017, and extracted those with at least one French author. We also analyzed data from the Transparence.gouv.fr corporate declaration database, for comparison with the authors' own disclosures in each article. Significant interest was defined by a €1000 threshold, although higher thresholds (>€5000, >€10,000 and >€100,000) were also investigated.

RESULTS

In 2017, 337 articles were published in OTSR, 210 of which had at least 1 French author. Of these, 201 (95.7%) had at least 1 author with significant interest (>€1000) according to the Transparence data. In 189 of these 201 articles (94%), authors had failed to disclose at least 1 direct or indirect conflict of interest. This omission rate fell to 22/45 (48.9%) for more substantial interests, which were more faithfully disclosed. At the €1000 threshold, in only 8/201 articles (4%) did authors disclose all their interests with perfect validity, but this rate increased up to 25/45 (55.5%) at higher thresholds. At the €1000 threshold, 66 of the 201 articles cited the trade-name of interest, resulting in 54/66 (82%) correct disclosure of direct interest; this rate increased up to 25/26 (96%) at higher thresholds.

DISCUSSION

At a threshold of €1000, the rate of complete and valid disclosure was 4% and the rate of omission 94%. At higher thresholds, rates were better, with just 48.9% omission and, above all, 55.5% validity, even if these rates were lower than hypothesized (75%). Authors and editors need to take greater care. Disclosures were often made, but incompletely, and authors need reminding that they must disclose not only interests related to the article in question but also all interests that might directly or indirectly influence their interpretation of the results reported, allowing readers to make up their own minds.

LEVEL OF EVIDENCE

IV; systematic retrospective study.