Trials were halted and hydroxychloroquine abandoned as a potential COVID-19 treatment after The Lancet and The New England Journal of Medicine published papers linking it to harmful side effects.
But the research has since been accused of being junk science: the papers were retracted last week and the datasets questioned.
Experts warn this may be the tip of the iceberg in a slew of poor research practices as journals rush to publish during the pandemic.
In Australia retractions might not be so swift as it is one of the few countries without an independent body to investigate dodgy research and regulatory standards going backwards.
Little oversight and accountability
Since the pandemic began, 15 papers have been retracted and two temporarily retracted. Four were published by The Lancet and two by The New England Journal of Medicine. Eight covered hydroxychloroquine.
Retractions and junk science are not new, scientist Professor David Vaux at Walter and Eliza Hall Institute told Inq. But they are growing.
The number of published retractions increased tenfold between 2001 and 2011. One study found 67.4% of retracted papers were pulled due to misconduct and fraud; 3.8% of papers in scientific journals contained duplicated images; researchers were unable to reproduce the findings of 47 of 53 “landmark” cancer papers.
Australia doesn’t have an independent research integrity office to investigate claims of academic misconduct. Instead institutions rely on the Australian code for the responsible conduct of research, although changes to the code in 2018 have made it easier for them to cover up bad science.
“The new code removed any requirement for independence, strengthened the secrecy provisions, and made it so institutions don’t have to use the term misconduct,” Vaux said.
The term is reserved for serious, repeated and intentional cases, but even then it’s up to an institution to decide whether to call it misconduct.
The new code was co-written by Universities Australia.
“Like the church, universities deny there’s anything bad going on,” Vaux said. “They want to cover it up to protect their reputations.”
Publish or perish
Dr Ascelin Gordon, research fellow in the interdisciplinary conservation science research group at RMIT University in Melbourne, told Inq the business model of journals encouraged novel, headline-making research.
“There’s an incentive to churn out these novel results rather than what’s best for science as a whole which would be replication of research,” he said.
Journals and researchers look to publish headline-grabbing papers which will receive a large number of citations, he said. Researchers’ grants, promotions and salaries are tied to their published works.
“It makes it easy to focus on questions that will give you something novel and interesting and do it in a quick way … Journals have to accept papers they think will be influential,” Gordon said.
Some journals, including Nature and Science, have implemented pre-registration of planned research, meaning a paper will be published even if its results are negative or boring to ensure topics are addressed and replicated.
Associate Professor of Law at Bond University Dr Wendy Bonython was surprised flaws in the hydroxychloroquine studies had not been found earlier.
“There was no data governance documentation framework,” she said. “It wasn’t clear where the data came from.
“This isn’t the first instance of questions raised on COVID-19 research … There’s a risk some standards of practices are being glossed over.”
Concerns over data’s validity
A Health Department spokeswoman said the department had concerns over the validity of the data.
The Medical Journal of Australia’s editor-in-chief Nick Talley says there had been an influx of submissions during COVID-19.
“We’ve gone up nearly double in submissions, which is a lot, and of those about a third or more or [are] on COVID-19, so it’s a very big spike,” he said.
“There is pressure [to publish]. Obviously if you wait months to review papers, the risk is people may die. You want to get information out that’s accurate and allows policy and public health action.”
The journal has introduced a rapid review process — and pre-printing draft research papers — to get through all the submissions, but hasn’t hired extra staff.
“Speed like this means we can make errors,” Talley said.
As for the recent retractions, Talley found them “disturbing. What a disaster for everybody.”
Junk science can be dangerous. In 2010 The Lancet retracted a paper linking the measles, mumps and rubella vaccine to autism, which is still widely cited by anti-vaxxers.
Deputy director at the Australian National Centre for the Public Awareness of Science Rod Lamberts says the fact the results had been questioned showed the system was working: “Mechanisms are in place to keep an eye on things.”
He encouraged everyone to scrutinise research carefully.