News from
PRINCETON UNIVERSITY
Communications and Publications
Stanhope Hall
Princeton, New Jersey 08544-5264
TEL 609/258-3600 FAX 609/258-1301

Release: January 25, 1995
Contact: Tom Krattenmaker (609/258-5748)


Researchers: Increasing Prescriptive Choices
Can Lead Doctors to Make Less Rational Decisions

Researchers at Toronto and Princeton Find 'Cognitive Bias' Affects
Doctors' -- As Well As Legislators' -- Decisions

PRINCETON, N.J. -- In a new study with important implications for
medical decision-making as well as public policy-making,
researchers from the University of Toronto and Princeton
University have found that increasing the range of prescriptive
options reduces the likelihood that a physician will make a purely
rational decision.

In an article appearing in today's issue of the Journal of the
American Medical Association, Toronto's Donald Redelmeier and
Princeton's Eldar Shafir present the findings of their survey of
several hundred Canadian and American physicians, and of a smaller
sample of Ontario legislators.

In a scenario involving a patient with osteoarthritis, for
example, the researchers found that family physicians were more
likely to prescribe a medication (72 percent) when there was just
one medication under consideration. In the alternative scenario --
in which the original medication and an additional medication were
considered -- only 53 percent prescribed either. "Apparently, the
difficulty in deciding between the two medications led some
physicians to recommend not starting either," according to
Redelmeier and Shafir.

The researchers observed the same phenomenon in the policy-making
realm. A sample of 41 Ontario legislators grappled with the
following two scenarios involving a decision on closing a hospital
in an over-served community: The first scenario presented them
with a straightforward decision whether to close a redundant and
inferior hospital, which 74 percent of the respondents chose to
do. In the second scenario, that same redundant hospital was
paired with another obvious candidate for closing. But in that
instance, 64 percent of the legislators, apparently unable to
decide between the two, chose to render no judgment and, thus, to
keep them both open. "This pattern reflects a general tendency to
resort to the status quo when facing difficult decisions," the
authors write, adding that it has "important policy implications."

"This kind of cognitive bias is nothing new," says Redelmeier, who
is affiliated with the University of Toronto's Department of
Medicine. "Real estate agents have long known that if they want to
sell a house, they should show the customers only a few. They know
that if they show too many the chance of selling anything
paradoxically goes down. What we've found is that the same pattern
can be found in decisions made by expert physicians in their field
of practice, we well as by legislators."

Another pair of scenarios in the Redelmeier-Shafir study deals
with the difficulty in choosing whom to operate on when multiple
patients are waiting but operating room space is limited.
According to the scenario presented to half of the physicians in
the sample, two patients -- a 72-year-old man and a 52-year-old
woman -- are waiting for carotid artery surgery, and only one can
be immediately accommodated. Most respondents chose to operate on
the woman, apparently swayed by her longer life-expectancy. The
other half of the sample had three patients from which to choose:
the same man and woman, and a second woman similar in age and
background to the first woman. Faced with that choice, most of the
physicians chose the man. "Apparently, the difficulty in deciding
between two similar patients ... led many physicians to avoid this
decision and recommend operating on (the man) instead," the
researchers write.

Lest people become unduly alarmed, Shafir cautions that the
physicians' apparent irrationality is not likely to be at play
when the "right" answer is overwhelmingly preferable. "These are
not giant errors we're finding," he says. "It's when a decision is
subtle that we see this effect, when it's a very close call."

Nor, he says, should the public want or expect physicians to be
perfectly rational "automatons." Says Shafir: "At the end of the
day, these physicians are going to have to rely on their
intuition, and that's usually pretty good. But it's bound to lead
in some cases to systematic biases and errors. An awareness of
these potential errors may be the first step in helping improve
medical decision-making."

---