Covid Bias at the BMJ

This week, Ioannidis and his colleagues published a paper on COVID-19 advocacy bias in the BMJ, concluding that the “BMJ had a strong bias in favour of authors advocating an aggressive approach to COVID-19 mitigation”.

The authors don’t hold back, saying the “BMJ had massive bias towards specific COVID-19-related advocacy favouring aggressive measures”. The BMJ became an outlet for independent SAGE/Vaccines-Plus advocates who outperformed SAGE members 16-fold and Great Barrington Declaration (GBD) advocates 64-fold. Short opinion pieces and analyses drove the majority of these differences.

Advocates of restricted, focused measures were virtually extinct from the BMJ pages: “BMJ editors, staff and apparently advocate contributors developed a massive literature, comprised mostly of opinion pieces that in general (as acknowledged by the BMJ) underwent no external review in the BMJ.”

If the BMJ were a broadcaster, it would have been reported to Ofcom, the UK’s communications regulator, because the news should be reported with due impartiality.

The BMJ’s approach is the exact opposite to its response to the Swine Flu Pandemic. Back then, it joined our Tamiflu team to publish our reviews.

It also created the Tamiflu campaign: The BMJ’s first open data campaign aimed to pressure companies to release the underlying clinical trial data for two globally stockpiled anti-influenza drugs, Tamiflu and Relenza. With Deb Cohen as its Investigation Editor, it helped track down the data.

At the outset of the pandemic, the BMJ was on a similar track. On March 2nd 2020, Tom published ‘COVID-19 — many questions, no clear answers’ in BMJ ‘Opinion’. “Jokers and spoofers are doing overtime on the web. The authorities cried wolf in 2005 and 2009 with influenza and see what you get now,” he wrote. On March 20th Tom published his last BMJ post on supermarket wisdom. In October 2020, Carl stood down as BMJ Evidence Based Medicine Editor in Chief.

So, what changed?

In Spring Tom submitted the Cochrane review on non-pharmaceutical interventions to the BMJ at the request of one of the journal editors. The two previous review updates were published in the BMJ (2008 and 2009) in response to the Swine Flu pandemic, and there was a need for an update. The review – last updated in 2011 – had grown in size and was submitted in two parts. The first part included the mask and distancing evidence, which was rejected after the committee meeting on April 10th and the second part was rejected without review.

Editors expressed “worries” because the confidence intervals did not exclude a huge protective effect of face masks for healthcare providers and a moderate (and potentially important) effect for the general public. They also wanted to lower the evidence bar: “Most editors felt that it was important to integrate RCT evidence with observational evidence.” Apparently, “case control studies can be quite good for looking at the effects of preventive interventions”.

Ultimately, the review did not find convincing evidence from randomised trials for the effectiveness of face masks, eye protection or person distancing. Because the results didn’t fit with the editor’s preconceptions, it was rejected.

The final straw wasn’t when we submitted a paper on transmission that led to abusive anonymous review comments. It was the publication of a character assassination that opined “How best can scientists push back against science denialist campaigns?”

The authors Gavin Yamey and David Gorski didn’t fact-check their article, there was no right of reply or communication with Sunetra Gupta or Carl, the BMJ thought it was okay to slander those mentioned as “merchants of doubt”.

In the Real Clear Investigations, Paul Thacker reported: “While Gorski and Yamey provided no evidence that Koch money funded the GBD signatories, the BMJ still published their piece. … The BMJ article is full of errors that ought to have never found their way into any publication,” wrote Martin Kulldorff in the Spectator.

We have published many times in the BMJ since 1995, including several articles with the Editor in Chief (e.g. 2017, 2019), and we have worked together on the Tamiflu campaign and the ALLTrials effort. Yet, Ioannidis’s team has shown what everyone in academia could increasingly observe in the Covid pandemic – the BMJ lacked impartiality, and chose to favour one side during the pandemic.

Medical journals aim to share the latest medical knowledge, including research findings. However, with the rise of the internet, they have started to include more news, opinions and articles that are better suited for a magazine format.

Journals that exhibit polarisation and lack impartiality during pandemics fail to represent the available evidence accurately. Despite this, their established reputations grant them significant sway, allowing them to shape doctors’ perspectives, influence academic discourse and play a crucial role in public policy formulation. This can lead to widespread acceptance of biased viewpoints, ultimately impacting healthcare decisions and responses to health crises.

Open dialogue and exploring diverse perspectives are essential for making informed, impactful decisions. The review of COVID-19 advocacy bias in the BMJ concludes: “The BMJ undermined the ability to navigate the complexities of the pandemic issues we faced and chose to champion opinion over evidence” by side-lining vital discussions.

Once a bastion of an evidence-based approach, the BMJ journal lost its way. History will judge that the lack of debate was a notable misjudgement.

Dr Carl Heneghan is the Oxford Professor of Evidence Based Medicine and Dr Tom Jefferson is an epidemiologist based in Rome who works with Professor Heneghan on the Cochrane Collaboration. This article was first published on their Substack, Trust the Evidence, which you can subscribe to here.