COVID-19 advocacy bias in the BMJ: meta-research evaluation.

Objectives: During the COVID-19 pandemic, BMJ, a leading journal on evidence-based medicine worldwide, published many views by advocates of specific COVID-19 policies. We aimed to evaluate the presence and potential bias of this advocacy.

Design and methods: Scopus was searched for items published until 13 April 2024 on 'COVID-19 OR SARS-CoV-2'. BMJ publication numbers and types before (2016-2019) and during (2020-2023) the pandemic were compared for a group of advocates favouring aggressive measures (leaders of both indieSAGE and the Vaccines-Plus initiative) and four control groups: leading members of the governmental SAGE, UK-based key signatories of the Great Barrington Declaration (GBD) (favouring more restricted measures), highly cited UK scientists and UK scientists who published the highest number of COVID-19-related papers across science (n=16 in each group).

Results: 122 authors published >5 COVID-19-related items each in BMJ: 18 were leading members/signatories of aggressive measures advocacy groups publishing 231 COVID-19-related BMJ documents, 53 were editors, journalists or regular columnists and 51 scientists were not identified as associated with any advocacy. Of 41 authors with >10 publications in BMJ, 8 were scientists advocating for aggressive measures, 7 were editors, 23 were journalists or regular columnists and only 3 were non-advocate scientists. Some aggressive measures advocates already had strong BMJ presence prepandemic. During pandemic years, the studied indieSAGE/Vaccines-Plus advocates outperformed in BMJ presence leading SAGE members by 16.0-fold, UK-based GBD advocates by 64.2-fold, the most-cited scientists by 16.0-fold and the authors who published most COVID-19 papers overall by 10.7-fold. The difference was driven mainly by short opinion pieces and analyses.

Conclusions: BMJ had a strong bias in favour of authors advocating an aggressive approach to COVID-19 mitigation. Advocacy bias may influence public opinion and policy decisions and should be mitigated in future health crises in favour of open and balanced debate of different policy options.