PATIENT HEAL THYSELF?

Psychiatrists will have to rethink the way they test drugs like Prozac

TENDENTIOUS. Preposterous. A failure of peer review.

These are not words that scientists normally use to describe the work of
other researchers. The recipients are two American psychologists, whose
study has firmly put the cat among the pigeons by claiming that most of the
therapeutic effects of expensive antidepressant pills like Prozac can be
mimicked by dummy pills (see p 13).

In other words, in the real world what matters most about these drugs is
not their ability to alter brain chemistry in specific ways, but their
ability to raise patients' hopes and expectations.

If the new study turns out to be correct, psychiatrists will need to
rethink their views about how antidepressants work. They may also need to
rethink their views about the causes of mental illnesses in the first
place. And, of course, if antidepressants do turn out to be little more
than elaborate placebos, perhaps the same will be true of other psychiatric
drugs. But these are big ifs.

When Prozac shot to fame in the early 1990s, it was touted as one of the
most specific of the psychiatric drugs because it acted only on designated
brain cells. It seemed to confirm what psychiatrists and drugs companies
had been saying for years: that mental illnesses such as depression are the
products of chemical imbalances in the brain that can be fixed---and
safely--- with a cunningly designed drug.

Tarnished by Freudian pseudoscience and controversial therapies such as
ECT, psychiatry has never had the equivalent of a Newton or Lavoisier, but
Prozac seemed to push it into a new era of hope, rationality, and
scientific respectability. In recent years, pharmacists' shelves have
filled up with other so-called serotonin re-uptake inhibitors, and tens of
millions of people around the world have swallowed the drugs on the
understanding not only that they work but that their actions are specific.

The new study seems to throw this perception into confusion. The idea that
patients' hopes and expectations influence the way they respond to pills is
nothing new. This, after all, is why researchers us placebos as control
treatments in clinical trials. But the idea that this "placebo responding"
can account for a staggering 75 per cent of the impact of antidepressants
will amaze doctors and patients alike.

Even more controversially, the study suggests the remaining 25 per cent
might also be a placebo response rather than a consequence of each drug's
biochemistry. The argument is that patients' expectations about a pill will
be greater if it produces noticeable side effects--- something drugs
usually do but not placebos.

It is obvious why drugs companies will find this hard to swallow. To
suggest that dummy pills work almost as well as bona fide antidepressants
is provocative; to suggest that antidepressants only come out ahead in
trials because of the side effects they produce is
to call the emperor naked.

Does it matter how a drug works as long as it gets the job done and doesn't
cause unmanageable side effects? Yes and no.

Drugs companies and most psychiatrists see placebo effects as ill-defined
and unpredictable. But patients suffering from depression may harbour no
such prejudices. They might not mind being given dummy pills that have been
engineered to produce a convincing but harmless array of side effects.

Where the distinction between a placebo response and a pharmacological
response becomes crucial, however, is in shaping ideas about the causes of
mental illnesses. Drugs like Prozac aren't just used to treat patients,
they are supposed to provide psychiatrists with insights into the
underlying brain chemistry of depression. But if placebo psychology not
chemistry is the reason for their effectiveness, those insights become
worthless.

Clearly, every last detail of the study must be thoroughly investigated.
One of these is the mysterious discovery that in trials, placebo responses
tend to vary in line with the impact of real antidepressants (see Diagram).
The researchers believe the simplest explanation for this correlation is
that the drug itself is acting as a placebo. But there are other
interpretations. Perhaps potent drugs enhance the effects of the placebos
they are being compared with by subtly raising the expectations of everyone
involved in the trials.

Unfortunately there can be no final verdict until researchers start to take
the placebo effect seriously. In practice, that means evaluating it instead
of simply controlling for it. Prozac nation deserves nothing less.

MOSTLY IN THE MIND

Antidepressants may be little better than placebos

THE benefits of antidepressant drugs could be almost entirely due to the
psychological boost derived from taking a pill rather than their effects on
brain chemistry, say two researchers in the US.

Irving Kirsch of the University of Connecticut and Guy Sapirstein of
Westwood Lodge Hospital, Needham, analysed 19 studies on selected
antidepressants and sedatives---including tricyclics and the newer
Prozac-type drugs---involving 2318 patients. In each study, the patients
had been given either an active drug or a chemically inactive placebo, and
their psychological conditions had been evaluated at the beginning and end.

Pharmaceuticals companies claim that antidepressants are 40 per cent more
effective than placebos. But Kirsch and Sapirstein found that the drugs
were only 25 per cent more effective. In addition, they suggest that even
that 25 per cent could be due to an additional placebo effect derived from
the side effects caused by the antidepressants, which alerted patients to
the fact that they were receiving an active drug rather than a placebo.
They say that the studies could have wrongly ascribed this additional
effect to a chemical change induced by the drugs.

Their analysis also suggests that antidepressants offer no advantage over
drugs such as anxiolytics and tranquillisers, which adds fuel to the
suspicion that the newer antidepressants are not as specific in their
actions as their manufacturers claim.

Simon Wessely, professor of psychiatry at King's College London, agrees.
"There's tremendous uncertainty about how they work," he says. "The public
thinks the doctors know, but they don't. Any decent psychopharmacologist
will tell you this."

Wessely says Kirsch and Sapirstein are right to point out that side effects
can alert a patient in a trial to the fact they are getting an active drug
rather than a placebo. "If patients know they're getting treatment, their
expectation will be raised and with it their optimism that they will get
better. It's a self-fulfilling prophecy." On the basis of this study and
one that Wessely participated in (British Journal of Psychiatry, vol 172, p
227), he believes that the advantage antidepressants offer over placebos is
just 15 to 20 per cent.

But a psychiatrist commentating on the new analysis in the latest issue of
Prevention S Treatment is fiercely critical of the paper. Donald Klein of
Columbia University, New York, who played a major role in developing
antidepressant treatments, says the work is flawed because the group of
trials chosen was "minuscule and unrepresentative" and amounted to "a
failure of peer review".

Kirsch's and Sapirstein's work does not show that antidepressants have no
pharmacological effect. However, Kirsch says the findings indicate "a
pressing need for new methodologies in clinical trials" to discover the
true extent of the placebo effect. One option might be to give some
patients "active placebos" that cause side effects but have no medical
effect.

The researchers' results also appear in the current issue of Prevention &
Treatment, the American Psychological Association's electronic journal
(http://journals.apa.org/prevention/).

Checked-by: (Joel W. Johnson)