May 6, 2004 (New York Times) A Top Scientist's Research Is
Under Attack By JANE E. BRODY
prominent Canadian researcher is facing claims that data in
his widely reported study of a nutritional supplement's
effects on thinking and memory in the elderly are so flawed
as to have no real value.
The scientist, Dr. Ranjit Kumar Chandra, is internationally
known for his many contributions to the field of nutrition,
and his work has been widely cited in professional and lay
publications alike, including The New York Times.
Scientific journals and three independent American
scientists have raised questions about the validity of Dr.
Chandra's findings, saying the study, published in September
2001 in the journal Nutrition, has statistical
irregularities and inconsistencies, and is characterized by
improbable research methods.
The study purported to demonstrate striking cognitive
benefits for people over 65 who took a daily multivitamin-
mineral supplement that Dr. Chandra formulated and has since
patented. He licensed the rights to the supplement to the
Javaan Corporation, founded by his daughter, Amrita Chandra
Gagnon. The company, in Somerville, Mass., sells the
supplement as Javaan 50.
Late last year, in response to objections raised by the
three independent scientists, Nutrition's editor, Dr.
Michael Meguid, published an editorial acknowledging that
Dr. Chandra's paper had serious statistical flaws.
"We regret that our peer review process failed to identify
these problems before publication," Dr. Meguid wrote.
In an interview by e-mail, Dr. Meguid said he was unaware
when Nutrition published the study that it had already been
rejected by the British medical journal BMJ. Dr. Richard
Smith, the editor of BMJ, said scientists who reviewed the
paper had found the methods and statistical findings so
unlikely that they wondered whether the study had actually
been done.
Reached by phone in Gurgaon, India, where he now lives, Dr.
Chandra said, "Anyone with different views should repeat the
study and see for themselves whether my findings can be
confirmed or not."
He added that any royalties he might receive from the
supplement were to be donated to education.
No one has yet established whether Dr. Chandra did in fact
do all the cognitive tests and statistical analyses that he
claims in the journal article to have carried out.
When officials at Memorial University of Newfoundland, where
Dr. Chandra worked for 27 years, asked to examine the
study's raw data, he replied that they had mysteriously
disappeared when the university moved his office. A
university spokesman, Dr. Jack Strawbridge, denied any
mishandling of Dr. Chandra's papers, adding that without the
raw data and with Dr. Chandra now retired and out of the
country, Memorial was unable to investigate the matter.
The three scientists who recalculated Dr. Chandra's
published results, however, said they could make no
sense of them.
One of those scientists, Dr. Saul Sternberg, an experimental
psychologist at the University of Pennsylvania who has
advanced training in statistics, said in a telephone
interview that he had found "statistical anomalies and
inconsistencies, measurements that were impossibly large."
Another experimental psychologist, Dr. Seth Roberts, who
studies learning and memory at the University of California,
Berkeley, said in an interview that he found the reported
effects of the nutritional supplement unbelievable.
"The statistics were not just implausible," Dr. Roberts
said, "they were impossible."
The two psychologists, along with Dr. Kenneth Carpenter, an
emeritus professor of nutrition at Berkeley, said they had
then found similar statistical aberrations in a previous
report by Dr. Chandra, published in The Lancet in 1992 and
based on the same group of elderly participants.
In the Lancet report, which has been lauded as a landmark
contribution to the field of nutrition and immunity, Dr.
Chandra concluded that his supplement greatly increased the
participants' immune responses, halving the number of
infections they suffered.
The question now is whether Dr. Chandra actually did what he
reported to have done in the two studies and, if so, whether
he analyzed the findings correctly. He stands by his
methods, saying in the interview that "there is more than
one way to do statistics."
Dr. Meguid, of Nutrition, pointed out in his editorial that
flawed research could have far-reaching consequences.
"The public uncritically believes the claims emanating from
such studies," he wrote, "and fellow scientists and funding
agencies divert precious resources to attempt to reproduce
or verify published data."