Advertisement

Retracted studies may have damaged public trust in science, top researchers fear

<span>Photograph: George Frey/Reuters</span>
Photograph: George Frey/Reuters

Public trust in science may have been shaken by the publication of academic papers based on false data in leading medical journals, according to world-renowned infectious disease doctors and former advisers to the World Health Organization.

The director of Australia’s Peter Doherty Institute for Infection and Immunity, Professor Sharon Lewin, said she and her colleagues were “gobsmacked” by the saga and said it should be “a wake-up call” in a global rush to publish studies about Covid-19.

Lewin said when the paper about the impact of anti-malaria drug hydroxychloroquine was published in the Lancet in May, she thought “on the face of it, it looked very impressive”. The study involved 96,000 patients across six continents, using a database called Surgisphere owned by a co-author of the paper, Dr Sapan Desai, which purported to collect anonymised patient information. It found that the drug was associated with heart problems and a higher risk of death in Covid-19 patients.

The Peter Doherty Institute is leading a trial of hydroxychloroquine’s effect on Covid-19, known as the Australasian COVID-19 trial (Ascot). “What we did of course is take the Lancet study seriously,” Lewin said. “The lead investigator of Ascot assembled his leadership group and a number of governance committees which oversee the study, and given the possibility of harm the usual response is to pause and review the trial, which is what we did.”

Related: Lancet had to do one of the biggest retractions in modern history. How could this happen? | James Heathers

But higher numbers of Australian patients were reported in the paper than Lewin knew had been in hospital at the time, something that made Lewin highly suspicious of the findings. It was an error also identified by the Guardian, and prompted The Lancet to issue a correction.

The World Health Organization, which halted its hydroxychloroquine trial due to the Lancet paper, reversed that decision on Wednesday. Ascot followed suit on Thursday. Surgisphere itself has come under greater scrutiny, culminating in the Lancet and the New England Journal of Medicine retracting two studies between them based on Surgisphere data by Friday.

“Once you get a paper like that in the Lancet or New England Journal of Medicine, because those journals command so much respect, what they print becomes gospel,” Lewin said.

“For this to happen in the midst of a pandemic it’s a wake-up call. Once a paper gets through peer-review there are standards we expect. I don’t think it was intentional by the journals, but the speed by which these journals are publishing Covid-19 research and the pressure they’re under means it’s a good, very clear wake-up call that standards should not be compromised.

“Not being able to answer an important scientific question about where this data came from raises doubt amongst community members about the value of studies, or they may make up their minds based on misinformation. If a finding is in a journal like the Lancet it can also affect clinicians and their biases. These journals change clinical practice. Science is powerful. [R]esearch is really important to advance the field and to us finding treatments that work and save lives. We need to bring the community along with us and that requires trust.”

Related: Unreliable data: how doubt snowballed over Covid-19 drug research that swept the world

In 2011, Desai led a paper that was published in the Journal of Vascular Surgery titled: Conflicts of Interest for Medical Publishers and Editors: Protecting the Integrity of Scientific Scholarship. In the piece, Desai wrote: “It is incumbent upon the publisher, editors, authors, and readers to ensure that the highest standards of scientific scholarship are upheld. Doing so will help reduce fraud and misrepresentation in medical research and increase the trustworthiness of landmark findings in science.”

Guardian Australia has contacted journals that have published Desai’s work to ask if they will be reviewing the dozens of papers involving Desai published since then, but is yet to receive a reply.

A professor of infectious diseases and former advisor to the World Health Organization, Prof Peter Collignon, said he believed the publication of Desai’s papers would have lasting implications. He said there was an assumption that the leading journals had rigorous standards, and that in his own experience, it was extremely difficult to get papers through the peer-review process of the journals.

A leading researcher in antibiotic resistance, Collignon regularly reviews academic papers before publication and said he always looks at the raw data. If he doesn’t understand statistical models used, the paper will usually get passed on to a statistician for review.

“It is unclear how something that at least looks suspect and at worst was fraudulent got through that process, and that does undermine trust,” he said.

“That’s a real problem when you already have people who don’t believe in science, and now they will say ‘how do we know everything isn’t made up?’”

On Friday Australia’s health minister, Greg Hunt, said concerns about the Lancet paper had been raised with him by a professor at the prestigious Walter and Eliza Hall Institute.

“He was right,” Hunt said. “What we’ve seen is that Australia is taking cautious, careful steps, following the highest medical standards to look at a range of different therapies and treatments, as well as the developments of vaccines. We’ll continue with our programs.” The chief medical officer, Prof Brendan Murphy, said “that paper has no status”.

“The situation with hydroxychloroquine is it is still an investigational drug, it is still a drug being studied in trials in Australia and in other countries, and we still await evidence about what its place may be, if any.”