Ranjit Chandra: how reputation bamboozled the scientific community

First doubts at The BMJ

Among them was the former editor in chief of The BMJ, Richard Smith, whose suspicions were aroused in 2000 when Chandra submitted a follow-up study of a major trial published in the Lancet in 1992.

Chandra was the sole author of the study, which showed that his patented multivitamin and mineral supplement boosted the memory and thinking in people aged over 65. But peer review raised several concerns, including the inability of one person to carry out such a large trial and the limited expertise of an immunologist to perform the psychometric tests used.

When the journal’s statistical reviewer concluded that “the data had all the hallmarks of being entirely invented,” Smith contacted Chandra’s employer, Memorial University of Newfoundland, requesting an investigation.

The university replied that there was no case to answer. But it didn’t mention that it had already investigated Chandra for suspected data fabrication, prompted by his research nurse Marilyn Harvey, whom Chandra later tried to sue. She was responsible for recruitment to the infant formula studies Chandra had been asked to carry out for Ross Laboratories, Nestlé, and Mead Johnson in the 1980s.

She told internal university investigators in 1994 that the numbers of infants recruited and reported on for the Nestlé study didn’t match and that neither a three year follow-up study of these infants nor the Mead Johnson study had ever been carried out. She had also seen the proofs of a five year follow-up to the Nestlé study, she said.

These allegations were investigated in some detail by Memorial in an internally published report in 1995 by H Kiefte, chair of the investigating committee. This report drew on expert opinion and the testimony of many witness statements but not on Chandra’s raw data because he threw these away after publication.

The report concluded that “scientific misconduct has been committed by Dr Chandra in this matter” and highlighted the “remarkable lack of communication and openness” in Chandra’s research environment, including with his coauthors.

“Dr Chandra’s research activity was very much operated as a pyramid system where only one person at the top had access to all the final raw data,” said the report. “The research personnel functioned mainly as physician-technicians and had, or were shown, little insight into experimental objectives, design, and procedures.”

But Memorial took no further action, and the report wasn’t made public until it became part of the evidence for the libel case that Chandra brought against the broadcasting company CBC, after its three part investigation into his research activities and financial affairs in 2006. CBC alleged that it had uncovered a pattern of scientific fraud and financial deception dating back to the 1980s. The jury ruled in 2015 that the CBC’s broadcasts were “true” and rejected an additional invasion of privacy claim, brought against the broadcaster in 2011.

University failures

In one of these programmes, Jack Strawbridge, director of faculty relations at Memorial at the time, defended the university’s failure to act on the report’s findings. He explained that Memorial was publicly funded and Chandra had threatened to sue for defamation.

He played down the potential impact of Chandra’s research: it was not a life and death matter, he reasoned. “If he [Chandra] were claiming a cure for cancer … that was fraudulent … it’s possible the university might have acted differently.”

In correspondence with Memorial this summer, The BMJ asked for a copy of the Kiefte report. Gary Kachanoski, Memorial’s president and vice chancellor, replied that it would be “inappropriate” to release it, because the “investigation report in question could not be relied upon on the basis of significant flaws in the investigation process.”

Memorial has so far declined to explain what these flaws were, or to answer many of the questions The BMJ has put to it. And in response to a request to share details of the legal settlement reached with Chandra in 1995, Richard Marceau, the university’s vice president of research, simply said that Chandra had dismissed his claim “without payment or any acknowledgement of wrongdoing by Memorial or Dr Strawbridge.”

In the wake of the CBC broadcasts Memorial commissioned Paul Pencharz of Toronto’s Hospital for Sick Children to carry out an independent review of its research integrity procedures and policies, dating back to the 1990s.

This concluded in 2007 that Memorial’s processes were “sound” and in keeping with those of other research institutions and national guidance of the time. Pencharz said that he was particularly impressed by the 1992 collective agreement Memorial had signed with its faculty association, which detailed how concerns about research integrity should be raised, investigated, and acted on if proved true.

He recommended that Memorial advocate for a national research integrity agency and re-open the Chandra case from 2000 onwards. “If it is found that there was scientific misconduct, then the other publications of the accused must be reviewed to determine whether or not any other publications contain false information,” he wrote.

In response, Memorial’s then vice president of research, Chris Loomis, doubted the feasibility of revisiting the 2000 allegations in the absence of the raw data but said that a list of all Chandra’s publications had been placed on its website, “with a note that the research data on which some of the articles are based cannot be located or verified.”The BMJ has been unable to find either this list or the Pencharz report on Memorial’s website.

Conflicted interests

Saul Sternberg, professor of psychology at the University of Pennsylvania, along with Seth Roberts, emeritus professor of psychology at the University of California at Berkeley, raised the alarm in 2003 with both Nutrition and the Lancet about Chandra’s research.

Sternberg doesn’t think Memorial did everything it could have. “But Chandra was famous and brought in research funds, and universities really do have a conflict of interest when it comes to investigations into their own personnel, because it makes them look bad, especially if they don’t expect justice to be served,” he suggests.

But The BMJ’s former editor Smith believes that Memorial was and is duty bound to pursue the matter. “Universities are about integrity and truth, and for Memorial not to pursue it fundamentally undermines what a university is about,” he contends.

Marc Masor, who managed and monitored Chandra’s infant formula study for Ross, thinks Memorial is “ultimately culpable” for letting Chandra off the hook. But others also had their part to play, he suggests.

Various anomalies and inconsistencies, including Chandra’s insistence that he had not been paid enough to design the study correctly, fuelled mounting doubts about the validity of the research and convinced Ross not to publish.

Masor regrets that Chandra’s reputation convinced him to trust more than verify but is clear that the company’s procedures for conducting clinical studies with externally contracted researchers during that period were rigorous. He wonders if the same could be said of his former competitors.

“If they were properly monitored, how could [Mead Johnson] possibly ‘lose track of the study’ [as it had told CBC], and how could Nestlé not be aware that their studies were published before the subjects were even enrolled?” he asks.

A five year follow-up of Nestlé’s Carnation study showed that the drop-out rate among more than 200 infants was zero. “Not only is that unlikely, but you are more likely to win the lottery three days in a row,” he insists.

Nestlé declined to respond directly to The BMJ’s questions about the thoroughness of its monitoring of the Carnation study or whether it had been under pressure in 1988 to provide evidence to the US regulator of the benefits of its infant formula, as CBC had claimed.

But in a statement Catherine O’Brien, vice president of corporate affairs at Nestlé Canada, said that the company had cooperated fully with Memorial’s formal investigation and had ceased referencing Chandra’s study in 2006.

When pressed about the reasons, she said, “It should be noted that any claims we make are based on a wide body of research. We had over 20 additional peer reviewed publications supporting our claim. As a result . . . there was no need or requirement for us to continue to reference this study.”

Asked why Ross Laboratories didn’t make its concerns known at the time, Masor, who no longer works for the drug industry, says, “We realised we couldn’t really trust the data, but we had nothing concrete that we could go to the university with.”

He adds, “Pharmaceutical companies are very averse to getting tangled up in legal wrangling, because regardless of the justification, chances are the public perception of a big company going after a renowned researcher would be very negative.”

Journal failures

In her email statement to The BMJ O’Brien said, “The study Dr Chandra carried out for us was published in a peer review journal,”—actually, more than one—implying that the journal was responsible for ensuring the accuracy of the scientific record.

Masor tends to agree. Chandra cited his previous research in the five year follow-up study. “They can’t have looked at those. He had all these data that weren’t there in the previous papers and no drop-outs after five years. How could a reviewer miss that? Frankly, it’s shocking.”

Michael Meguid, editor of Nutrition, which in 2001 published the paper rejected by The BMJ, accepts blame. “But there’s no question that an aura around an individual as being a good and important scientist and a man of distinction influences one’s perception,” he explains.

He was “bamboozled” by Chandra’s reputation and the lure of a follow-up study to published research in the Lancetthat offered a potential boost to the impact factor of a relatively young journal, he says. Chandra had requested rapid publication to help secure a grant, and when one of the journal’s reviewers took too long, Meguid gave Chandra the benefit of the doubt.

Other researchers soon started pointing out the implausibility of the findings. Chandra at first dismissed the concerns and then when more followed, threatened, “If you publish this, bad things will happen to you,” Meguid recalls. When Nutrition retracted the paper, Chandra wrote to the chancellor of New York State University, accusing Meguid of data fabrication and threatening a $5m lawsuit.

Chandra has published around 200 articles. But only Nutrition, and now The BMJ, have retracted his research since concerns were first raised over 20 years ago. The BMJ has contacted the editors of six journals that featured in the Kiefte report or the CBC programmes: four have agreed to revisit the issues; some said they were not aware of any concerns about Chandra or the outcome of the CBC trial.

“We might need to post expressions of concern about all of [Chandra’s studies], or publish a note or editorial saying that we can’t be sure this research is OK, and retract those looked at in more detail,” says Smith.

The Nutrition paper has been cited more than 130 times since its retraction, and in a bid to curb this tendency Retraction Watch is working with the Center for Open Science at the University of Virginia to create a comprehensive freely accessible retraction database that will include expressions of concern and corrections.

There was no national research integrity body to approach in Canada when Chandra was publishing his suspect studies, and strict privacy laws prevailed. There still isn’t one. But the Panel on Responsible Conduct of Research, which brings together three federal research funding agencies, was set up in 2011. Any research institution that wants to be considered eligible for one of the agencies’ grants must agree to its framework, which includes going public on serious breaches.

As to Chandra, he has not responded to any of The BMJ’s emails but is now the managing director of a company based in India called Peridot Life Sciences, which sells nutritional supplements. Some of these have sparked the interest of the Indian government, says Meguid, who spoke to Chandra during the CBC trial.

The BMJ contacted the Indian Medical Association to find out if it intended to take any action against Chandra in view of the CBC trial result. Its honorary secretary, K K Aggarwal, replied that this would be the responsibility of the state medical council where Chandra is registered and has not responded to any of The BMJ’s other questions on the matter.

The BMJ is still awaiting a formal response from the professional regulator in Newfoundland and Labrador, the province where Chandra worked while at Memorial. But Fleur-Ange Lefebvre, executive director of the overarching Federation of Medical Regulatory Authorities of Canada, told The BMJ that each authority “will be able to act on available external information, like the publicised results of a court case.”

Chandra may also be stripped of his illustrious awards. The Office of the Governor General, which oversees the honours system in Canada, said that it couldn’t comment on specific cases but sent a list of the revocation criteria, some of which Chandra would seem to fulfil.

The Chandra case is an example of how complex the governance of research has become, thinks Chris Graf, vice chair of the Committee on Publication Ethics. And it may yet get worse.

“The whole market is expanding. The amount of research conducted and published is growing rapidly,” he says. “And open access cuts both ways: it has brought with it a whole load of innovation and new ways to communicate, but it has also made it easier for less reputable organisations to speculate and profiteer.”