115 Quality of conflicts of interest declarations in a conference setting – are audiences given a chance to adequately assess bias?. (May 2019)

• Record Type: Journal Article
• Authors: Crawford, William; Camm, C. Fielder; Prachee, Ishika; Ginks, Matthew
• Is Part Of: Heart. Volume 105(2019)Supplement 6
• Page Start: A96
• Page End: A97
• Publication Date: 2019-05
• Subjects: Conflict -- Declaration -- Bias; Heart -- Diseases -- Treatment -- Periodicals; Cardiology -- Periodicals; 616.12
• DOI: 10.1136/heartjnl-2019-BCS.112
• Languages: English
• ISSNs: 1355-6037

Abstract

Abstract : Introduction: Potential conflicts of interest (COIs) are common. Interaction with industry is often required to develop cardiology research. However, it has been suggested that COIs may affect professional opinion objectivity. As a result, appropriate COI declarations are required when presenting research. Methods: Recorded presentations from the British Cardiovascular Society annual conference 2018 were reviewed. Presentations were excluded if they were panel discussions, chairperson remarks or motivational talks, or slides were not available. A 3-point quality score was developed to assess declarations; 1) was a COI slide displayed for a minimum sufficient time (≥250ms/conflict), 2) the presence of a verbal COI explanation, and 3) use of a font size comparable (<10% smaller) to the BCS default. Presentations were given a score of 0–3 based on these criteria and compared against the presenter's H-index as a surrogate marker of seniority. Results: A total of 84 presentations were available for review; 70 met the inclusion criteria, and 48 (69%) contained a COI statement. Of those with a statement, the median number declared was 0 (IQR 0 to 1); while 12 (25%) declared at least one COI. Of those presenters with at least one COI, 5 (41%) did not offer a verbal explanation; 5 (41%) used a font size smaller than the default; and conflicts were displayed for a median time of 1s per conflict (IQR 0.8–3.0s). Three (25%) presenters with a potential COI declared it with a quality score of 3, while 1 (8%) presenter scored 1. No presenters scored 0 (see figure). Linear regression models showed a presenter's H-index was predictive of the number of conflicts reports (OR 1.1 95% CI [1.05 to 1.15], p<0.001, n=45), but not time spent per COI (OR 0.89, 95% [0.68 to 1.16], p=0.38, n=11). An ordinal regression model did not show an association between presenter's H-index and total quality score (OR 0.94, 95% CI [0.87 to 1.01], p=0.11, n=11) Conclusion: A large minority of presenters did not display a COI statement. Those COI statements that were present were of variable quality. Many were displayed for insufficient time, lacked any verbal explanation or used a small font. These factors may limit the ability of audience members to understand the potential conflicts. Seniority of presenter correlated with number of conflicts but not quality of the COI statement. Strategies should be implemented to ensure all presenters outline whether they have any potential COI in a standardised manner. Conflict of Interest: None