On 16 June 2020, the life-saving benefit of dexamethasone in severely ill patients with COVID-19 was announced at a public briefing at the home of the UK prime minister, 10 Downing Street. The results, from the RECOVERY trial in the UK, changed practice in the UK immediately, with a letter sent later that evening from the UK chief medical officers and the chief executive of National Health Service (NHS) England to all NHS trusts that stated: “Normally we would advise waiting for the full paper before changing practice, to ensure final analysis and peer review do not lead to different conclusions. However, given this clear mortality advantage, with good significance, and with a well-known medicine which is safe under these circumstances we consider it is reasonable for practice to change in advance of the final paper.”

At that time, and repeatedly throughout the pandemic, there has been criticism that science should not be communicated by press release. Some doctors demanded a preprint manuscript, others the full peer-reviewed paper, before any decisions about the use of dexamethasone in patients with COVID-19 were made. We released full results of the dexamethasone comparison in a preprint manuscript just 6 days later, on 22 June 2020. We took to heart the criticisms of publication by press release and subsequently, for all RECOVERY results that have a potential effect on current COVID-19 care, have endeavored to put out a preprint simultaneously with the press release.

Credit: OU Images_John_Cairns 2021

I am not in favor of publication by press release alone. The practice of some pharmaceutical companies of publicly announcing preliminary results, either from a legal obligation to disclose findings that may affect the company’s market valuation, or as a marketing ploy, may have negative consequences for patients, as well as damaging the reputation of science communication.

But during health emergencies, there are many tensions, one of which is the mismatch between the urgent need for information and evidence and the much longer time frames of scientific peer review and publication. The COVID-19 pandemic is the first global health emergency of the new information age, with data and results widely shared via social media. This has resulted in very real difficulties in distinguishing important information from noise, and real news from fake news. How should the research and medical community best manage this new reality?

The results of clinical trials certainly require careful validation and critical appraisal before patient care is changed. But the business-as-usual biomedical peer-review system falls short during a rapidly evolving global health emergency. It is simply too slow. The full peer-reviewed paper for the dexamethasone result was published online in the New England Journal of Medicine on 17 July 2020. In the month between the public announcement of the result and the publication of the peer-reviewed manuscript, over 700,000 newly confirmed COVID-19 cases were reported globally. The interval between the RECOVERY preprint for the antibody therapeutic tocilizumab and the peer-reviewed paper was even longer, at 11 weeks (11 February 2021 to 1 May 2021). Waiting for peer review would undoubtedly have cost lives.

Some may argue that the speed advantage of preprints does not outweigh the risks of poor-quality, misleading or even fraudulent research being published and acted upon. I would counter that clinicians should not rely solely on peer review to assess the validity and meaningfulness of research findings. This is because dubious, perhaps fraudulent data can still get through peer review, as was seen with early COVID papers published and then retracted from two of the most prestigious medical journals. In addition, even valid data can be misleading. There has been an avalanche of observational data that passed peer review and was then used to justify treatments, most notably with hydroxychloroquine, but the susceptibility of observational methodology to moderate biases means that such data should not be the basis of patient care.

I take two lessons from our experience running the largest COVID-19 clinical trial over the last two years. The first is that that the preprint system has come of age, demonstrating huge value in rapidly communicating important research findings. Almost daily I am alerted through social media alerts from trusted sources and colleagues of important new findings published as preprints. A degree of immediate peer review is also available by means of the preprint comments section and from colleagues via social media. The full peer-reviewed manuscripts usually appear many weeks or even months later. I cannot envisage a future without such rapid dissemination of new evidence.

Given this new reality, the second lesson is that we must ensure that the medical community and policy makers are sufficiently skilled in critical thinking and scientific methods that they can make sensible decisions, regardless of whether an article is peer reviewed or not.