Note to readers
The main draft of this case-study was finished in January 2014 as part of my unpublished (unpublishable?) book about error and fraud in scientific and medical research. Since that time Chandra has fought and lost a major libel action against the Canadian Broadcasting Corporation (CBC) who had openly accused him multiple acts of research fraud in TV programmes and in print. Chandra was ordered to pay $1.6million to cover the CBC’s legal fees. The jury concluded that whilst the CBC broadcasts were defamatory they were also true. They dismissed his claims for invasion of privacy. The judge is quoted as saying:
“In the end he failed miserably”
This court case prompted at least two journals the Lancet and BMJ to retract faked papers that had been polluting the scientific record for more than a quarter of a century. In January 2016, he was stripped of his Order of Canada award. I have made amendments to my 2014 draft to try to reflect these latest developments in this long running saga.
From 1974- 2003 Chandra was based at Memorial University in the small city (population c100000) of St John’s, Newfoundland, Canada. From this relatively small town he produced a huge output of research papers, review articles and books and became acknowledged as one of the world’s leading experts on the relationship between nutrition and immunity. He practised paediatrics at the Janeway Child Health Center and was director of the World Health Organisation Center for Nutritional Immunology both also based in St John’s. He has been dubbed the father of nutritional immunology. He retired from Memorial University in 2003 and returned to India via Switzerland, he remained professionally active and in 2015 was managing director of an Indian based company selling dietary supplements. His sudden retirement was prompted by allegations initiated by the BMJ (formerly the British Medical Journal) that he had fabricated data for a clinical trial which suggested that a micronutrient supplement could improve cognitive function (memory) in elderly people. This paper was retracted by the editor of the journal Nutrition in 2005 with open accusations that these data and probably those in other papers had been fabricated. His work on the effect of this same supplement on immune function and infection is now also generally disregarded. The publicity surrounding this retracted paper led to revelations that in 1994 Chandra had been accused of fabricating data which reported that a certain type of infant formula reduced the risk of babies developing atopic (allergic) diseases like asthma and dermatitis. This work is now also generally disregarded. Marilyn Harvey was the whistleblower who first raised suspicions about his infant formula work in 1994. She was threatened with legal action by Chandra and clearly felt that her career was damaged by her whistleblowing and that she was badly treated by the university. In 2014, Memorial University instituted The Marilyn Harvey Award for research ethics and she was its first recipient.
My personal involvement with this case
This case was my first direct contact with allegedly fraudulent research and it sparked my interest in the topic. In my first book (1995), I cited a 1993 review by Chandra that was published in The Proceedings of the Nutrition Society. Much of the general information in this review is still considered to be essentially correct but I would no longer have confidence in using anything written by Chandra as a source material for my own writing. His reviews frequently cite his own work; no fewer than 18 of the 26 citations in this review list him as an author or editor. In my 1996 book on The Nutrition of Older Adults, I cited three of Chandra’s papers and reviews and he undoubtedly influenced my views on the relationship between nutrition, aging and immune function. I summarised one of Chandra’s papers (Chandra, 1992), published in the Lancet which purports to show that a micronutrient supplement improved immune function and reduced infections in older people. In the 2005 retraction notice for Chandra’s Nutrition paper, the editor, Michael Meguid suggested that the data in this Lancet paper was also fabricated and the trial was probably never conducted. This paper was eventually retracted by The Lancet in 2016, almost a quarter of a century after publication.
I continued to cite Chandra in various editions of my books including one on Dietary Supplements and Functional Foods. These Chandra papers and reviews remained in the literature and even the online versions had no “statements of concern” so I saw no reason to doubt their veracity. His only retracted paper up until 2015 was the one published in Nutrition in 2001 whose 2005 retraction notice contains detailed and wide-ranging accusations of research fraud against Chandra.
The BMJ played a pivotal role (see later) in raising suspicions that the data in this paper on supplements and cognitive function had been fabricated. Despite this, in April 2005 the BMJ published a meta-analysis by El-Kadiki and Sutton which amalgamated all controlled trials testing the effects of micronutrient supplements on immune function in the elderly. This amalgamation included data from Chandra’s group. Chandra’s results certainly affected the conclusions of this paper and the authors had difficulty reconciling the very positive benefits and very low variability reported from the Chandra group with the absence of benefits and wide variability seen in other papers. The review’s authors would have seen no reason to exclude Chandra’s results on immune function from their analysis. I felt less embarrassed about my own continued use of Chandra’s work when I realised that the referees and editors who had assessed this meta-analysis had accepted the paper despite the BMJ’s earlier key role in making accusations of fraud against Chandra. Publication of this meta-analysis prompted Richard Smith to write a 2005 BMJ editorial focusing on the problem of dealing with the previous output of an author who had committed research fraud. This example illustrates the importance of disinfecting the literature of suspect work after a fraudulent author has been unmasked. In a correction published in July 2005, El Kadiki and Sutton presented a re-analysis of their data to show it when results from Chandra’s group were excluded, supplements then made no difference to infection risk. It was not until a couple of years ago that any serious disinfection of the literature to remove Chandra’s tainted papers began.
Chandra was born in 1938 in the city of Multan which is now part of Pakistan. As Hindus the family felt forced to flee from Pakistan to India in 1947 when India became independent and “British India” was partitioned into “Muslim” Pakistan and “Hindu” India. According to a story recounted at Chandra’s 60th birthday celebrations in 1998, the Chandra family were stopped on their trek to the Indian border by “Muslim marauders” but were allowed to proceed unharmed because Chandra’s father had saved the life of the son of the leader of this band of marauders (a great Bollywood scenario?). This anecdote is attributed to a speech made by Professor John Soothill an Emeritus professor at the Great Ormond Street Hospital for Sick Children. Another anecdote attributed to Professor Soothill concerned his dealings with a Canadian immigration officer. When Professor Soothill told the immigration officer that the purpose of his visit was to celebrate Chandra’s birthday, she told him that Chandra had successfully treated her sick daughter and that he need pay no import duty on the present that he was bringing for Dr Chandra. Reading through the other speeches at this celebration, on Chandra’s web-site, one was left in no doubt that this is the birthday of someone seen as a great scientist, physician and human being who is highly respected and revered by his patients and within the scientific and medical community.
Chandra studied medicine in the Indian city of Amritsar in the Punjab and graduated from the Government Medical College there. He gained his MD from the All India Institute of Medical Sciences in New Delhi and taught at this institution for a while where one of his students was his future second wife Shakti.
After leaving India, Chandra spent some time in London as a postdoctoral researcher at the Institute of Child Health, in Great Ormond Street Hospital. He then became a professor at Memorial University and director of the World Health Organisation Center for Nutritional Immunology based in St John’s. He remained in St John’s until his retirement from the university was precipitated in 2003 by accusations of research fraud made by the BMJ (details in the Richard Smith editorial). Chandra had apparently held visiting professorships at John Hopkins University, the New York Medical College, Massachusetts Institute of Technology, Harvard University and at medical schools in China, Chile, Italy, India, Argentina and Spain. He is said to have received 11 honorary degrees and many other honours including a D.Sc. from Punjab University. In 1989 he was awarded the Order of Canada which is awarded to living Canadians for:
“The highest degree of merit, an outstanding level of talent and service, or an exceptional contribution to Canada and humanity”
From 1981 until 2003 he was founding editor-in-chief of the journal Nutrition Research. During his tenure as editor he accepted for publication several dubious papers supporting his own research.
He was listed on Macleans’ magazine’s 1995 annual honour roll of 12 Canadians who had “made a difference”. He was said at this time to have twice been nominated for the Nobel Prize in medicine. He was also “awarded” (i.e. purchased) other very grand sounding titles such as Honorable Baron de Tronville from France and Honorable Baron of Blackburn from Scotland. After leaving Canada under a cloud of suspicion he seems to have spent a period of time in Switzerland where he styled himself vice-chancellor and Proviseur Designate of the Universite International des Sciences de la Sante in Crans-sur-Sierre in Switzerland . According to the Canadian Broadcasting Corporation this Swiss university is just a box number. An online search for this institution only generates hits related to articles published by Chandra; there is no university home page or details of this institution. He then returned to India and was associated with various institutions in Gurgaon with limited internet footprints including President of the Nutritional Immunology and Allergy Centre .
The financial rewards of academic success
As well as generating considerable professional and academic prestige, Chandra also seems to have generated a considerable fortune. His second wife Shakti Chandra left the marital home in 1993 and divorced him. The legal attempts to resolve disputes between Chandra and his ex-wife about the family financial assets are detailed in judgements of the Newfoundland United Family Court, for example in December 2000 . It is clear from reading these judgements that Chandra had amassed a considerable fortune during his time in Canada. Chandra’s failure to co-operate and provide full disclosure of information about these assets meant that an independent chartered accountant appointed by the court could not properly assess the extent of his financial assets. His finances were described in these judgements as a labyrinth.
His wife provided the court with a list of 120 bank accounts scattered across the world that had been in existence up to 1993. The cumulative total of monies held in these accounts amounted to over $2.5 million ($4.2million at 2017 prices). Numerous other assets and saving instruments were also found. According to his ex-wife, much of this money was generated by her husband’s consultancy and speaking fees and honoraria paid to him as a result of a large number of speaking engagements worldwide. He maintained to the court that he only received expenses for these engagements. Shakti provided a summary to the court of the extent and duration of his travels since 1980. In several years he spent more than 100 days annually away from Newfoundland. For example in 1991 he spent a total of 126 days on his travels and according to her received over $120,000 in fees and honoraria.
The Canadian court considered it unlikely that Chandra would spend more than 100 days a year away from this successful medical practice without receiving compensation for loss of income from his practice. Although the court could not precisely estimate the extent of his wealth because of his lack co-operation, it was concluded that substantial amounts of money had been accumulated and that these could not be explained by the couple’s joint incomes from their teaching and medical practices in Newfoundland.
Chandra patented his specific formulations of vitamins and minerals and results from his research and writings were used to promote their sale. A 2003 fact sheet for the Javaan Corporation promotes a formulation called “Javaan 50” as:
“The first daily multivitamin formulated to support the immune system of persons over 50”
The founder of Javaan Corporation was Chandra’s daughter Amrita Chandra and Ranjit Chandra was acknowledged as the inventor of this formulation. Chandra’s long experience of studying the effects of nutrition upon the immune system is noted as is his fabricated data relating to cognitive function in the elderly:
“Dr Chandra believes that the cognitive benefits from improved nutrition could dramatically improve the lives of the elderly”.
A product profile (March 2003) gives the composition of Javaan 50 and also reviews the evidence, mainly the discredited clinical trials of Chandra, that support the efficacy of this product. There is no other substantial evidence that supplements improve immune or cognitive function in normal, apparently well-nourished, elderly people. Chandra and the Javaan Corporation have maintained that he never received any personal financial benefit from sales of this supplement.
The rise to scientific stardom and influence
Prior to his fall from grace, Chandra was widely regarded widely as one of the world’s leading authorities on nutrition and immunity and has even been called the father of nutritional immunology.
During his early professional life in India, he became editor of the Indian Journal of Paediatrics and published prolifically in this journal from the mid-1960s right through to 2002 and contributed scores of articles, editorials, case reports, commentaries, and book reviews (PubMed lists 114 articles but there were many more). This prolific output of written material during his early career in India set the tone for the rest of his career and he claimed, probably correctly, to have published over 200 peer reviewed papers, 190 academic reviews and book chapters and 22 books.
Several of Chandra’s papers concern the benefits of taking micronutrient supplements on immune function. The now finally retracted 1992 Lancet paper, has been cited more than 300 times; it purported to show that his patented supplement greatly reduced time spent with an infection in elderly Canadians. Aging is associated with a decline in the efficiency of the immune system. Malnutrition and micronutrient deficiencies also have deleterious effects upon immune function. These papers by Chandra seemed to indicate that micronutrient deficiencies may help to cause or exacerbate this age-related decline in immune function or as Chandra put it:
“A causal relationship does exist between under-nutrition and impaired immunity and that this is a correctable abnormality in the majority”
Given his striking output of review articles and book contributions, this was a view that he had many opportunities to promote to a wide scientific and medical audience. His writing certainly influenced my views on the usefulness of micronutrient supplements for older people in early editions of my books. Several of Chandra’s papers have been cited more than 200 times so his work was influential.
Earlier in his career, Chandra published a series of papers on the effects of different types of infant feed upon the risk of babies developing allergic diseases like eczema. His work suggested that infant formulae made using hydrolysed (partially digested) whey milk protein had advantages over conventional formulas for babies at high risk of developing allergic diseases.
A study published in the BMJ in 1989 reported data for over 200 babies who had a family history of allergic disease. This report (eventually retracted in 2015) suggested that amongst breastfed babies 33% developed eczema; for those women in this group who avoided common allergens during lactation, rates of allergy in these was half that in those who did not (i.e. 22% and 44% respectively). In the babies who were formula fed, the rates of eczema were 70% for those fed on normal cows’ milk formula, 63% for those fed on a soy-based formula but only 21% for those fed on a new hydrolysed formula. Babies who fared best were those who were breastfed by a mother who avoided the common allergens during pregnancy and babies fed the new hydrolysed formula; the results for these two groups were very similar. Around this time, Chandra published other similar studies using a total of 700 babies perceived to be at high risk of eczema and asthma. These babies were all supposedly recruited from the area around the small town of St John’s, Newfoundland. These were double-blind studies where many thousands of tins of formula with coded labels would have been needed to be supplied so that neither parents nor physicians knew what formula was being used until the study had ended. His work indicated that new hydrolysed formula was superior to standard formula in reducing the risk of allergic disease and as good as breast milk in this regard.
Acknowledged experts like Chandra exert further influence by being given opportunities to influence scientific opinion by invitations to write reviews, speak at conferences and sit on advisory bodies. Chandra wrote many review articles including some in prestigious journals and some based upon invited keynote lectures given at major international conferences. Any scientist writing about or researching the influence of nutrition and supplements upon immune function and allergy would have almost certainly been influenced by Chandra in the 1990s. Review articles purport to review all available literature but, like many other review writers, Chandra quotes extensively from his own work in his reviews.
Manufacturers of hydrolysed infant formulas and manufacturers of dietary supplements used Chandra’s publications to support the benefits of using their products. Take for example, a 2002 Fact Sheet published by an American trade association representing dietary supplement manufacturers the Council for Responsible Nutrition entitled Benefits of nutritional supplements: immune function in the elderly by Dr Annette Dickinson. This fact sheet, which is no longer accessible online, overviewed the size and health problems of the elderly and highlights their increased risk of fatal infections. Dickinson cited several studies which all suggested that supplements can reduce infection risk in the elderly and argues that all people in nursing homes should be given supplements including the following quote:
“It would appear prudent to place all nursing home residents on an inexpensive multiple vitamin containing the RDA and to consider placing the more debilitated residents on generic supplements containing several time the RDA of water-soluble vitamins”.
Chandra’s publications and views are heavily featured in this article especially his 1992 Lancet, paper and a 2002 paper with similar findings in a slightly younger (50-65 years) age group. A paper attributed to Amrit L Jain but published in a journal edited by Chandra and generally thought to be written by Chandra is also used. The papers by Chandra and Jain provide most of the support for supplement use in Dickinson’s fact sheet. Of course she would have seen no reason not to rely on the writings of this leading authority on nutrition and immune function in producing her fact sheet.
Chandra’s research output has been largely discredited in recent years especially his work in two key areas:
- The influence of feed choice on allergy risk in infants
- The effects of vitamin and mineral supplements on immune function, infection risk and cognitive function in the elderly.
In 2000, Chandra submitted a paper to the BMJ which suggested that his micronutrient supplement markedly improved cognitive function (i.e. memory) in elderly people. This data was supposedly collected at the same time and using the same subjects as his infamous 1992 Lancet paper which suggested it improved immune function and halved infection time in this group. There thus appears to be an eight year gap between data collection and submission of the second paper. This raises suspicions because most researchers who had generated such seemingly important data would want to publish it immediately. There is no mention in the Lancet paper that data about cognitive function was also collected and would be published in a later paper.
This paper on cognitive function was rejected by the BMJ and one of their reviewers said that it “had all the hallmarks of having been entirely invented” (Smith, 2005). This paper was subsequently published (Chandra, 2001) in the journal Nutrition. The BMJ notified Chandra’s employers and the other journal about their concerns. In August 2002 Chandra resigned his post at the Memorial University of Newfoundland, Canada and in February 2005 the paper was eventually retracted by the editor of the journal (Meguid, 2005). Until two or three years ago this was the only paper of Chandra’s that had been retracted so its faults are discussed in some detail.
In this Nutrition paper, the sample of 96 elderly people received either a supplement for a year or a placebo. All subjects were middle class Canadians who were said to be psychiatrically healthy and cognitively normal at the start of the study with no signs of dementia. Blood samples were taken from the subjects and levels of vitamins, trace minerals and iron measured. The subjects’ memory was assessed at the start and end of the year using a battery of seven cognitive function tests. The control subjects performed similarly on all tests at the start and end of the year. However for the supplemented group there was a significant improvement over the year on six tests that measured short-term memory but not the one that assessed long-term memory recall. Long term memory is known to be less affected during early dementia than short term memory. He claimed that those with one or more blood nutrient levels classified as deficient also had significantly lower scores on the cognitive function tests. If this study was conducted as described then it represents a very formidable undertaking yet it is a single author paper and there is no acknowledgement of any assistance in carrying out the study. Chandra’s own specialities were paediatrics and immunology and yet there is no acknowledgement of anyone with psychiatric/psychological testing expertise.
In a letter to the editor of Nutrition, Shenkin et al (2002) indicated their scepticism of the dramatic effects of the supplements upon cognitive function. The supplemented group improved their scores on one test, the Mini-Mental State Examination (MMSE) from a mean of 18 to 28 i.e. six times the improvement registered in trials using a drug designed to treat Alzheimer’s disease. Despite Chandra’s claim that the subjects were cognitively normal, they pointed out that score of less than 22 would indicate significant cognitive impairment and to be classified as “normal” subjects would need a score of over 26. Thus many patients went from being highly cognitively impaired to normal i.e. the supplement appears to have cured dementia in many subjects! They also pointed out that the standard errors quoted by Chandra meant that some subjects would have had to register minus scores and some more than the maximum on the MMSE test! In his reply Chandra said that there was a typographical error and the standard errors were actually standard deviations. Whilst this would have corrected the impossible range, it created other statistical impossibilities that are discussed later. In his reply Chandra suggests that he used other information to exclude dementia at the outset but no one with any psychiatric expertise was acknowledged. Chandra describes three precedents for nutritional interventions being more effective than drugs; two of these relate to his own papers on the effects of supplements on immune function and the effect of hydrolysed infant formulas in reducing allergy incidence (both now finally retracted)!
Chandra’s paper caught the attention of a reporter, Jane Brody, who summarised the claimed findings in the New York Times. This newspaper report in turn attracted the attention of psychologists Drs Seth Roberts and Saul Sternberg. They reiterated the point that the standard errors reported by Chandra would have indicated ranges from less than zero to more than the maximum score and which he had previously explained by claiming that the term standard error rather than standard deviation had mistakenly been used. Roberts and Sternberg found that substituting standard deviation for standard error opened up a whole series of massive new problems. It meant, for example, that for all seven of the tests of cognitive function the placebo and supplement groups were statistically significantly different at the start of the study; an impossible finding if subjects had been randomly assigned to the two groups as claimed. Using standard deviations to estimate statistical significance between the control and test groups at the end of the trial they found probabilities that were unlike any I have seen in any clinical trial of this type (P<0.0000000000001!). If these were real results one could say without any remote possibility of doubt that the supplements caused huge improvements in cognitive function amounting to a cure for dementia. They also found other statistical improbabilities in this paper and were highly critical of the statistical data in the Lancet paper on immune function that had used this same sample group. Chandra’s published results suggested many of the subjects had dementia at the start of the study despite Chandra’s assertion that they were all cognitively normal. The numerical values for some of the other tests were out of line with those recorded elsewhere.
Roberts and Sternberg note that unlike the short-term memory tests no name or reference was given for the long the long-term memory test. Chandra gave just a vague description of long term autobiographical memory test i.e. involving recall of details about the person’s own life history. Expert psychologists Roberts and Sternberg state that there was no well-established test of this type and development of such a test would have been a substantial achievement in its own right (especially for a non-psychologist!). In a second letter to Nutrition, Susan Shenkin and colleagues suggested that if his results were genuine, Chandra:
“Had reported on of the most efficacious treatment effects in medicine since the introduction of penicillin”
These closely and logically argued criticisms of this paper indicate that Chandra probably went a step too far in attempting to fabricate results for tests outside his area of experience and expertise. He invented results that can be shown by experts in the field to be ludicrously improbable. Chandra’s responses did not address the specific criticisms and amount to little more than vague bluster. He suggested that because his work has been peer-reviewed and some of his findings praised by others then they must be sound. He suggested that Roberts and Sternberg might be trying to discredit his work to serve the interests of companies making anti-dementia drugs.
In his retraction notice Michael Meguid, says that Chandra had failed to satisfactorily answer most of the substantial and detailed criticisms of the paper and had failed even to address most of the issues raised. He lists five key unanswered criticisms:
- The initial scores for the MMSE indicate dementia in many participants despite all subjects being claimed to be psychiatrically normal.
- Control and test groups were significantly different at the outset for all seven measures of cognitive function which would be impossible if they had been randomly allocated.
- There was no explanation or citation for the long-term memory test and it is unlikely that this test exists.
- The scores reported for another test (the Wechsler Memory Test) are incompatible with scoring used in the reference cited for this test.
- There was no overlap between the deficient and non-deficient groups for three of the cognitive tests which given the definition used to categorise subjects is statistically improbable.
Meguid goes on to suggest that similar problems with the Lancet paper on immune function, conducted as an arm of this study, call in to question whether this famous study was ever conducted. He also criticises two further papers both published in Nutrition Research (editor R K Chandra!) that have implausible features, Chandra (2002) and Jain (2002) which were mentioned earlier. Chandra failed to declare his financial interest in the Javaan Corporation which marketed the supplement being tested.
Since these concerns were published, others have voiced concerns about Chandra’s work. The veracity of three other supplement papers, two by Chandra and by the untraceable AL Jain. The most significant and influential of these was the highly cited Lancet paper (Chandra 1992) on immune function and infection. The other two papers can be regarded as replication studies that involve partial repetition and apparent substantiation of the Lancet results. Both of these were published in a journal, Nutrition Research, which had Chandra as editor-in chief. Both were accepted within a day of being received which suggests that independent peer review did not take place.
Carpenter et al (2003) wrote a letter to the Lancet’s editor about the paper 1992 Lancet paper. They questioned the feasibility of some of the probability values and discrepancies between the standard deviations quoted in the text and in the histograms. They noted that everyone asked had agreed to take part and that all subjects were middle class despite claims to be a random sample. Sternberg and Roberts (2006) also raised several concerns about the Lancet paper and the two repetition studies as listed below.
- The means and variability measures are inconsistent with the probability values quoted.
- The variability in the number of infected days for the subjects are improbably low in all three papers and much lower than that reported by other authors.
- The results for the number infected days are so similar between the three papers that they considered them too good to be true.
- They were unable to trace Amrit L Jain and could find no record of any publications by him with a PubMed search. They could find no record for his professional address in Jaipur, India and the address for correspondence was a rented mailbox in Canada.
Despite being disregarded by people working in the field, all three papers remained in the literature for many more years. Two meta-analyses of the effects of these supplements in the elderly (El Kadiki and Sutton, 2005; Stephen and Avenell, 2006) both now present results without the inclusion of data from these three Chandra/Jain papers. Both meta-analyses demonstrate just how out of line these studies are with the other findings and how small the variability is compared to that seen in other studies.
These publicly aired doubts about the veracity of Chandra’s supplement trials eventually re-awakened an earlier controversy about his trials of infant formula. These trials were sponsored by formula manufacturers and supposedly involved around 700 babies selected because of a history of allergy in one of the parents’ families. Marilyn Harvey, a nurse working at Memorial, was tasked with recruiting babies for these studies and she clearly felt that to recruit such large samples of willing and eligible families from the local area was going to be a formidable task that would take more than two years. Harvey was very surprised therefore to see that one study was published when she estimated that only around a quarter of the necessary subjects had even been recruited. Mark Masor, a clinical research associate working for Ross pharmaceuticals who had been liasing with Chandra about a study of his company’s product was also surprised because his company’s product had been used as part of this study but had not been supplied by Ross. Around 20,000 tins of formula with specially coded labels would have been necessary for double-blinding the study.
Harvey decided to act when she saw a further paper that Chandra planned to submit for publication; the five-year follow-up of an earlier study (using a Nestle product) that was eventually published in the Journal of Pediatric Gastroenterology and Nutrition (Chandra, 1997). Harvey alerted the university authorities about her concerns and an independent panel was formed to investigate the allegations. This committee sat for 3 months, interviewed witnesses and assessed the veracity of five of Chandra’s papers. The Canadian Broadcasting Corporation (CBC) obtained a copy of this report which effectively finds Chandra guilty and used it in a three-part documentary exposing Chandra’s misdeeds entitled “The secret life of Dr Chandra” broadcast in early 2006 in The National. Below extended quotes from this 1994 Memorial University report as relayed in the paper transcript of the CBC broadcasts in 2006 .
“With the evidence presented, the testimony of many witnesses and the fact that absolutely no raw data (or files) of any kind were exhibited, the committee cannot accept that the (Nestle) study was done anywhere near to the completeness or with the accuracy reported…. For that matter, the same can probably be said for the Mead Johnson work as published in the British Medical Journal.”
“In fact the committee cannot identify anyone who did or remembers (doing) a significant amount of work, and the co-authors of the papers had very little or very likely nothing to do with the work.”
“With respect to the allegations, the committee is, therefore led to conclude that scientific misconduct has been committed by Dr Chandra.”
These damning conclusions might lead one to assume that this would have signalled the end of his academic career at Memorial, however he remained in post and continued to publish prolifically and for example published at least ten review articles between 1995 and 2005. Some of these review articles were in respected academic journals like the American Journal of Clinical Nutrition, the Proceedings of the Nutrition Society in the UK and the European Journal of Clinical Nutrition.
Memorial University took no action because Chandra accused the investigating committee of bias and threatened to sue the university. The university vice-president is quoted by CBC as suggesting that the reputation of the university and its income would have been threatened by a major lawsuit. One rather perverse statement from this interview was that they felt unable to act because the alleged offences were so serious that the only appropriate penalty would have been dismissal because:
“Fabricating research results is usually considered a capital crime in academia.”
Implying that a lesser offence might have been penalised because a lesser penalty could have been imposed! He goes on to suggest that if the research had been about a life and death matter e.g. a false claim for a cancer cure then they might have acted differently. A false report about reducing childhood eczema by using a relatively unknown type of infant formula was not serious enough to demand action! This inaction allowed Chandra to later effectively claim that he had found a very effective treatment for dementia; a claim certainly on a par with claiming a cure for cancer. As noted earlier, Shenkin et al put the therapeutic potential of Chandra’s claims in the same class as the discovery of penicillin. No doubt if confirmed, Chandra’s claimed effects would have meant a third and probably successful nomination for a Nobel Prize.
When Chandra was eventually exposed in 2002, he retired from Memorial and left Canada and so the university appears to have closed the door on the matter. The vice-president suggests that they could not proceed any further because they did not have access to his data. The nutrition community have thus had to act independently to exclude Chandra and Jain’s work from reviews and books. If the summary data used in a paper is fabricated then it follows that the raw data does not exist and has never existed and so lack of access to raw data seems like the perfect alibi for fraudsters.
I pose a few questions about this case below and leave the reader to make their own judgements.
- Should Memorial University have taken action in 1994 to stem the publication of fraudulent data by Chandra? Were they seriously in danger of losing libel suit or were they more concerned about negative publicity?
- Should Memorial have done more in 2002 to investigate Chandra’s misconduct and should they done more to warn the scientific community about his misconduct?
- Should journal editors have retracted more of his discredited papers or at least issued warnings about potential issues with the veracity of the data?
- Is the refusal to supply raw data not a tacit admission of guilt?
- Should the standard of proof required to retract papers be lowered if someone has a past record of misconduct?
- How could he get away with fabricating studies that would have been expected to produce a huge footprint in a community the size of St John’s? For example:
- Many thousands of blood test to measure micronutrient status and immune function tests
- Hundreds of high allergy risk babies being used in feeding trials and monitored for extended periods of up to 5 years
- Tens of thousands of tins of infant formulas with coded labels distributed to hundreds of families
- A small army of assistants collecting and preparing blood samples for analysis
- Hundreds of subject interviews and questionnaires to assess their characteristics, cognitive function and infection incidence.
- How did Chandra’s papers get through the peer review system of some respected scientific and medical journals despite huge and, in hindsight, glaring errors and inconsistencies?
With respect to the apparent failings of Memorial University Newfoundland, the BMJ through its former editor Richard Smith were extremely critical of Memorial in 2015 at the time they retracted Chandra’s 1989 study on infant formula. Its former editor Richard Smith is quoted as saying:
“MUN has failed badly”
“The university should have taken this much more seriously”
“From my point of view, the university’s the real villain of the piece”
MUN was criticised for taking no action after its own 1994 enquiry had found Chandra guilty. They also took no action when notified by the BMJ about concerns regarding the authenticity of data submitted about supplements and cognitive function that eventually resulted in the, now retracted, 2001 paper in the journal Nutrition.
The question about peer is one that I have given a lot of thought to. I was for some years on the editorial board of the British Journal of Nutrition so I have tried to envisage what I would have done if the paper on cognitive function had been allocated to me to arrange for peer review. I would have been aware of the Lancet arm of the study dealing with immune function. I would have been aware of Chandra’s reputation as an expert on the effects of diet on immune function. I would have acted as second referee myself and chosen a psychologist to act as the first referee. I would also have sent the paper to a statistical editor. I would have hoped that the psychologist would have spotted some of the faults in the cognitive function results e.g. ranges that were inconsistent with normal values and the lack of a long-term memory test that met Chandra’s description. I would have expected the statistical editor to spot at least some of the most basic of the statistical flaws with the paper e.g. mean and variability measures that translate into clearly impossible ranges.
The most difficult thing to decide retrospectively is what I, as a nutritionist, would have contributed to this process. At this time, I would not have actively considered the possibility that the data might be fabricated; highlighting the need to increase awareness of scientific misconduct. I would have recognised this as another arm of the highly cited 1992 Lancet study on immune function which had already satisfied the reviewers at the Lancet. I would thus have felt reassured, or maybe complacent, about the basic design of the trial. I would have been surprised by the long delay in publishing the results in view of their apparent importance and also by the lack of any co-authors or acknowledged assistance in what seems like a very labour intensive piece of work. I would have noted the significant improvement in all the measures of short-term memory although I would not have been familiar with these tests and thus not aware of how these numerical values would translate in terms of functional capability of the subjects. I would have seen no obvious problems with the study design as described by Chandra and would have already felt reassured on this point because of the 1992 Lancet paper. I would now expect to spot that some of the means and variability indicators were not feasible although I am not confident that I would have spotted this then. Given the apparent practical significance of the results especially when combined with those already published in the Lancet, I would probably have been minded to accept the paper in principle prior to reading the reports of my statistical and psychological advisers. Given this admission, I feel some sympathy for the original referee(s) especially if their only expertise was in nutrition.
What happened to the whistleblower?
Marilyn Harvey was the nurse employed by Chandra to recruit new parents to take part in his infant formula trials. She “blew the whistle” and reported that he had published before she had recruited the subjects. This prompted the 1994 MUN inquiry that found that “scientific misconduct has been committed by Dr Chandra” but which resulted in no action and the report was not published. In July 2000, Marilyn Harvey was served with notification that she was being sued by Chandra for allegedly stealing data from one of his studies. The lawsuit was dropped a few months later but this must have caused her considerable distress at the time. In 2007, after details of Chandra’s activities were broadcast in Canada, Harvey filed her own lawsuit against the university which claimed that they had failed to properly investigate the allegations that she had made more than a decade earlier. The lawsuit contended that the university had led the medical community to believe that her accusations were unjustified and she thus acquired a reputation as a troublemaker. In response to this lawsuit the university issued a press release in May 2007 making clear that they intended to contest these allegations. Their defence can be summarised as:
- Ms Harvey had never been directly employed by Memorial University
- They had investigated the complaint but found insufficient evidence to sustain a complaint against Chandra
- They had always treated Ms Harvey in a fair and appropriate manner.
Also in 2007, a Toronto-based researcher Paul Pencharz had independently reviewed the Chandra case. He recommended that the university investigate this case further and also recommended instituting whistle-blower protection.
I have not been able to find out how this legal case between Ms Harvey and Memorial University was resolved but during this search, I did find that in 2014 Memorial University had introduced a Marilyn Harvey Award to recognise the importance of research ethics. In the accompanying description of the award it states that the university:
“Has named this award in honour of Marilyn Harvey, BN, a research nurse who brought forward her concerns regarding research ethics to senior administrators at the university”
This seems like a quite a turnaround in attitude and one can only hope that the university regrets its past actions and has tried to learn from any mistakes that were made in its handling of the Chandra case. This seems like a belated recognition of the correctness of Harvey’s actions and something akin to a public apology by the university.