Faking it

Can you really trust science? Perhaps not. In a worrying report published in June, a third of medical and scientific researchers admitted to committing scientific fraud, and nearly three-quarters said that they have witnessed deliberate warping of data to show desired results. Only a very small proportion of these cases are known about fully—thanks to the lack of effective policing of the drugs industry—but medical fraud could affect anything from mild asthma medication to treatments for serious cancers. (See Jim Giles's account of the Vioxx scandal, Prospect, November 2008.) Yet a crucial chance to tackle the problem is being squandered—thanks to resistance from the heart of academia.

When research fraud hits the headlines, it typically involves outrageous cases, like the South Korean scientist Hwang Woo-Suk's fabrication of stem-cell research that purported to treat diseases like Parkinson's, or Norwegian researcher Jon Sudbø's faked cancer trials in the Lancet in 2005. These, however, are the tip of the iceberg. The real damage to scientific integrity is wreaked by the everyday, routine distortion of results by researchers anxious to bolster their careers, appease academic bullies or justify research grants.

The true extent of this abuse was glimpsed in Edinburgh University researcher Daniele Fanelli's report in June, which analysed all the existing anonymised surveys questioning scientists about misconduct. The results, published in the journal PLoS ONE, suggest that altering or making up data is more frequent than anyone had previously estimated. Misconduct was reported most frequently by medical and pharmacological researchers, and involved things like deliberately omitting test results, adding non-existent "positive results," altering basic test data or cherry-picking volunteers who would respond to tests in the desired manner. This supports growing fears that industry sponsorship is distorting scientific evidence to promote commercial drugs. Huge sums are at stake here: in the US alone, an estimated £18bn a year is spent by drug companies on marketing their products. The pressure on those who develop them, then, is intense.



Fanelli was driven by his own experience as a behavioural ecology researcher. "I had naively assumed that scientists would be principled, but I ended up not believing most of the research that I read," he says. "Scientists are human beings driven by their interests, hopes and beliefs. Given an opportunity to cut corners by falsifying data, they may well do so." It seems criminally easy to get away with. In the US, about 1,000 incidents of suspected fabrication, falsification or plagiarism in scientific research go unreported every year, according to a 2008 study in Nature. Meanwhile, Britain's Committee on Publication Ethics (Cope) estimates that every year there are about 50 cases of seriously fraudulent research, primarily in biomedicine, in major British institutions. Many involve falsification or fabrication of laboratory results, undeclared conflicts of interest, authors putting their names to research they had not performed, and clinical malpractice. But hardly any cases are picked up officially, says Cope.

To tackle the problem, the UK Research Integrity Office (UKRIO) and Research Councils UK have proposed national bodies with powers to investigate misconduct allegations independently. But objections from academic institutions meant both bodies capitulated. Universities claimed that the plans trespassed on their "employer responsibility" to ensure good conduct. Critics believe that they simply prefer to deal with misconduct cases covertly, lest their reputations become expensively sullied.

Aubrey Blumsohn was a Sheffield University lecturer who in 2006 raised the alarm over a study that backed a new osteoporosis drug despite not having had open access to the drugmaker's research. (He was threatened with disciplinary charges, but eventually left with a six-figure settlement.) "Since my case became public, I have had hundreds of emails from people with similar problems," he says. "There's a huge undercurrent of things below the radar."

Blumsohn believes that the problem will only be tackled effectively by a fundamental change in academic research culture. An independent investigatory body would be a good first step, but even this may not be enough. According to UKRIO's Harvey Marcovitch, the job is not as simple as introducing strong regulation. "The Norwegians were quick off the mark to produce an inquiry into Jon Sudbø. But since then, Sudbø has been allowed to return to practice." We may have to wait for a fresh scandal before anyone acts. Until then, patients will remain in real danger of taking expensive drugs whose risk of harm or inability to cure have been fraudulently suppressed.