The British Medical Journal retracted a 1989 study about the role of breastfeeding and baby formula in infant eczema this week because of scientific misconduct on the part of the study author, Dr. Ranjit Kumar Chandra. The journal editors called the decades-late retraction "a major failure of scientific governance."
The retracted study claimed that mothers from families with a history of eczema could reduce their babies' risk of developing the disease by following a restrictive diet or feeding their babies formula. An internal investigation by the Canadian university that employed Chandra at the time of his study determined in 1995 that the data to support these claims were entirely fraudulent.
The university, however, didn't release the results of its investigation to the public. It was "The Secret Life of Dr. Chandra," a three-part documentary produced by the Canadian Broadcasting Corporation, that first exposed the researcher's misconduct in 2006. This summer, Chandra, a well-known researcher who'd gained international recognition for his work, lost a $132 million libel suit he filed against the CBC in response to the documentary and has since left the country. Chandra did not respond to the BMJ's request for comment, and HuffPost's requests for comment to Chandra's India-based nutritional supplement company were not immediately returned.
"We had always maintained that our journalism got the story right and the content was true," Emma Bédard, a spokesperson for the CBC, told The Huffington Post. "Clearly, the jury agreed with us."
Chandra came under fire long before the CBC documentary
Perhaps the most distressing aspect of the BMJ retraction is that it took more than two decades for Chandra's falsified data to be exposed publicly.
In 1993, Chandra's colleagues at the Memorial University of Newfoundland suspected that he had falsified some of his study data. The university investigated the claims, and two years later concluded that he had, in fact, committed scientific misconduct. The university did not, however, make its investigation report public.
"This was a bad failure by the Memorial University of Newfoundland," Richard Smith, former editor of the BMJ, told HuffPost. "It has never published the report. It came into the public domain through the court case."
A university official told the CBC on Wednesday that it understands the retraction.
"We at Memorial University have very high standards. We maintain them, we enforce them," said vice president of research Richard Marceau. "We have progressed over the years. We have learned a great deal."
Almost a decade later, a second study came under question. In 2000, Smith, then editor of the BMJ, approached Memorial after Chandra submitted a one-man randomized study about vitamin supplements to the BMJ that seemed off to Smith. When Memorial couldn't produce raw data, the BMJ declined to publish the study.
A different journal, Nutrition, published the suspicious study in 2001. Nutrition retracted the study four years later for "statistically significant errors."
Why did it take so long for the scientific community to act on their suspicions?
"Science seems to have no adequate system for identifying and retracting fraudulent studies," Smith said. "It shows how a fraudulent paper can take 25 years to retract."
Study retractions are relatively rare. Between 2 million and 3 million scientific papers are published each year, while only some 500, or 0.02 percent, are retracted, Ivan Oransky, who runs the transparency blog Retraction Watch, told CBS News in May.
Still, retracted studies can have a wide-reaching impact. British researcher Andrew Wakefield's discredited 1998 study that falsely linked vaccines to autism continues to provide fodder for the anti-vaccine movement in 2015 -- even though the study was retracted by The Lancet in 2010.
Peer review isn't enough to stop fraud
Medical journals have a peer review process -- third-party researchers assess studies to determine if they are publication-worthy -- but the system largely relies on trust.
"Peer review does not protect against fraud," Smith said. "If an author says there were 200 patients, then the reviewer assumes that to be true."
And while some in the scientific community advocate the online publication of all study data to increase transparency, an author could still fabricate or manipulate the data. That was the case in a study published in the journal Science last December that claimed that gay canvassers advocating for same-sex marriage were able to change voters' minds. The study was retracted, citing "statistical irregularities," in May.
The latest retraction raises questions about Chandra's 200 other studies
While the retracted paper is old and isn't likely to influence the behavior of current breastfeeding mothers -- major organizations such the American Academy of Pediatrics don't recommend a restrictive diet for breastfeeding moms to prevent allergies -- the retracted study has been cited more than 100 times by other papers since it was published.
The retraction also casts a shadow on Chandra's 200 other published studies, only two of which have been formally retracted.
"We still don’t know how many of Chandra’s studies are fraudulent," Smith said. "The logical assumption should be that they are all fraudulent until proved otherwise."
CORRECTION: An earlier version of this article stated that the journal Nature retracted a study by Dr. Ranjit Kumar Chandra in 2005. In fact, the journal Nutrition retracted the study.
Also on HuffPost: