|
Everyone has
heard of urban legends, stories -- like alligators in the sewers
or LSD on stickers handed out to schoolchildren -- that spread around
the country, repeated so often that people swear they are true.
Skeptics who try to track such tales to their sources
find that the grapevine that transmitted them trails off into mist.
It is a phenomenon
that seems totally alien to the world of science. In all its branches,
science is proudly defensive of its rigorous, disciplined empiricism,
methodically excising unproven beliefs from its discourse and brushing
off pseudosciences that try to attach themselves to its validating
coattails.
But two Danish
researchers offered science a humbling lesson last week, saying
in a paper published in the New England Journal of Medicine that
the placebo effect does not exist.
Now the placebo
effect -- which is supposed to account for the fact that about a
third of patients get better when given a dummy pill or a sham treatment
-- has been cited in textbooks and journal articles for decades.
It is part of the medical mainstream and the popular culture. So
established is the placebo effect that some scientists focus their
careers on exploring how it works, looking for changes in the immune
system or in hormones that allow the mind to affect the body.
How could the
idea of the placebo effect be so prevalent if it does not exist?
For the same reason, it seems, that those alligator stories spread.
The Danish researchers
happened to notice that the placebo effect had a sort of hearsay
quality in medical papers. As they grew more intrigued, they found
layer after layer of cross-references in scientific publications.
Finally, the found the source: a 1955 paper, "The Powerful
Placebo," by an anesthesiologist, Henry Beecher, at Massachusetts
General Hospital in Boston.
Dr. Beecher
examined 15 studies that compared a placebo to an active drug. Using
a method of analysis that would not be accepted today, he chose
the subset of patients who improved with a placebo, and disregarded
those who got worse. From that, he concluded that about a third
of patients get better simply from taking a dummy pill. "He
came up with the magical 35 percent figure that has entered placebo
mythology," said Dr. Asbjorn Hrobjartsson, one of the Danish
scientists.
He and his colleague,
Dr. Peter C. Gotzsche, decided to revisit the phenomenon, using
more up-to-date research methods. They searched the world's publications
for well-designed studies that included not just a placebo group
but a group that got no treatment. They found 114, which involved
about 7,500 patients with 40 different conditions. When they analyzed
the data, they concluded that patients given nothing improved just
as much as those given placebos.
Their conclusion,
that the placebo effect is nothing more than a medical legend, has
presented doctors, scientists and the public with sort of a Rorschach
test on their faith in a mind-body connection. (Come to think of
it, that may not be the right analogy -- psychologists have
increasingly said that the notion that the inkblots reveal character
traits is, ahem, a myth.)
The idea that
the mind can control symptoms and disease can be immensely appealing,
said Dr. Clement McDonald, a professor of medicine at Indiana University.
"Everyone wants it to be so," he said. "It gives
them control." Dr. John C. Bailar III, an emeritus professor
at the University of Chicago, said: "It's a secular religion.
And as a religion, no kind of evidence is going to get believers
to change their minds."
Added to that,
said Dr. Donald Berry, a statistician at the M. D. Anderson Cancer
Center in Houston, is the desire to attribute effects to causes.
For instance, he said, he read an article on the effects of biofeedback
on blood pressure. The blood pressure of a group that did
not have biofeedback fell just as much as it did in the group that
did. The researchers decided that they were seeing a powerful placebo
effect just from entering the study. "They said, `This is amazing.
It shows the benefit of being in a clinical trial,' " Dr. Berry
said.
But there is
a simpler interpretation, he said. Among the study participants
were people whose blood pressure, by chance, had just edged up enough
so that they qualified. According to the statistical rule called
regression to the mean, it is almost certain to be lower the next
time it is measured. The high reading was an anomaly. The biofeedback
-- and the placebo effect -- may well have done nothing.
Few expect that
the paper published last week will automatically convert true believers.
It may take more research, and more skepticism even of the skeptics,
before the dust finally settles.
Dr. McDonald
said he wrote a paper 18 years ago that concluded that the placebo
effect did not exist. But, he said, the New England Journal of Medicine
rejected the manuscript, saying that everyone knew the effect existed.
The paper was eventually published, in Statistics in Medicine. But
he met with such disbelief that he gave up even talking about his
findings.
"It wasn't
the right time," he said. "But the good thing about science
is that sooner or later the truth comes out."
|
