London - The trustworthiness of “open access” science journals claiming they check the reliability of research before publishing it has come into question after a majority of online publications agreed to reproduce bogus papers without spotting obvious errors.
Deeply flawed research apparently demonstrating the anti-cancer properties of a wonder drug extracted from lichen was accepted for publication by more than half of 300 online journals targetted in an investigation.
Open access is a relatively new form of science publishing that does not rely on expensive subscriptions, but asks scientists to pay an administrative fee if their study is accepted for publication on the journal's website.
David Willetts, the science minister, has actively supported the model, on the grounds it enables individuals and small companies to read publicly-funded scientific research that they have helped support through their taxes, without having to pay for the expensive subscriptions of conventional scientific publishing houses.
However, in a “sting operation” carried out by a journalist working for Science - a subscription-only journal - many open-access publications appear to be ready to publish obviously flawed research without the rigorous peer-review normally expected.
More than 300 versions of the “wonder drug” study were sent out to various open-access journals which claim to check each study's credibility before publication. The study claimed that the wonder drug extracted from lichen makes cancer cells more susceptible to radiation.
Any experts reading the methods would have realised the effect could just as well have been caused by the alcohol used to dissolve the substance. Yet such was the rush to publication, and to charge the fictional scientist the administrative fees, more than half of the journals accepted the scientific paper without noticing its fatal errors in scientific methodology, says Science.
John Bohannon, the journalist who organised the sting over a period of 10 months, said that the findings showed just how easy it was to circumvent the usually tough peer-review procedures that science journals should employ.
“The goal was to create a credible but mundane scientific paper, one with such grave errors that a competent peer reviewer should easily identify it as flawed and unpublishable,” Mr Bohannon said.
He said one glance at the paper's “materials and methods” should have explained the outlandish results and if these had been missed then the paper's obvious ethical issues - such as advocating the bypassing of clinical trials - should have raised alarm bells.
FACT AND FICTION: THE PUBLISHED REPORT
Claim: The effect of the wonder drug is “dose dependant”, meaning that the higher the concentration of the drug, the more lethal it is to cancer cells.
Truth: The wonder drug was dissolved in a buffer solution of strong alcohol whereas the “control” was not. So the higher concentrations of alcohol, rather than the drug, could have explained the cell-killing properties.
Claim: Cancer cells exposed to the wonder drug were more likely to be killed by radiation than unexposed cancer cells, indicating a potential use in radiotherapy.
Truth: In fact, a quick look at the methodology would show that the control cells (those that were not exposed to the wonder drug) were never exposed to any radiation at all so the observed effect was nothing more than standard inhibition of cell growth by radiation.
Claim: The paper concludes by stating that the wonder drug is effective against cancer in animals and humans and is promising new treatment for combination radio-chemo therapy.
Truth: The study is only based on cells growing in a test tube with no animal studies and certainly no evidence that could bypass clinical trials in humans. - The Independent