Disclosure of study funding and author conflicts of interest in press releases and the news: a retrospective content analysis with two cohorts

Abstract

Objectives: To examine how often study funding and author conflicts of interest are stated in science and health press releases and in corresponding news; and whether disclosure in press releases is associated with disclosure in news. Second, to specifically examine disclosure rates in industry-funded studies.

Design: Retrospective content analysis with two cohorts.

Setting: Press releases about health, psychology or neuroscience research from research universities and journals from 2011 (n=996) and 2015 (n=254) and their associated news stories (n=1250 and 578).

Primary outcome measure: Mention of study funding and author conflicts of interest.

Results: In our 2011 cohort, funding was reported in 94% (934/996) of journal articles, 29% (284/996) of press releases and 9% (112/1250) of news. The corresponding figures for 2015 were: 84% (214/254), 52% (131/254) and 10% (58/578). A similar pattern was seen for the industry funding subset. If the press release reported study funding, news was more likely to: 22% if in the press release versus 7% if not in the press release (2011), relative risk (RR) 3.1 (95% CI 2.1 to 4.3); for 2015, corresponding figures were 16% versus 2%, RR 6.8 (95% CI 2.2 to 17). In journal articles, 27% and 22% reported a conflict of interest, while less than 2% of press releases or news ever mentioned these.

Conclusions: Press releases and associated news did not frequently state funding sources or conflicts of interest. Funding information in press releases was associated with such information in news. Given converging evidence that news draws on press release content, including statements of funding and conflicts of interest in press releases may lead to increased reporting in news.

Keywords: medical journalism; public health; statistics & research methods.

Figure 1

Flow diagram describing the two datasets and the available numbers for analysis.

Figure 2

Proportion of news stories reporting study funding source or industry funding according to whether the associated press release reported the funding source or industry funding. There were zero mentions of industry funding in news in 2015 dataset, so analysis not performed. Error bars are 95% CI. The relative risks for the three plots are: 3.1 (95% CI 2.1 to 4.3); 6.8 (95% CI 2.2 to 17); 2.1 (95% CI 0.94 to 4.5).