Evidence - Based Medicine Challenges
Accepted Medical Practices
<<< Back to main document
Straight Goods - Canada's independent on-line source of news you can
useIf forwarding , pls edit out reference to myself...thanks Michael
Evidence-based medicine challenges accepted medical practices
Doctors have not been taught how to evaluate original research or help
patients weigh up the benefits, and risks, of alternative therapies
Dateline: Monday, June 30, 2003 by Gordon Guyatt, MD "After a decade of encouraging women who have gone through menopause to
take hormone replacement therapy, it�s unpleasant telling them that they
�ve been taking a dangerous drug."
The speaker, a family doctor I met at a medical conference, was
uncomfortable. She had followed expert advice, and told her patients
that they could reduce their risk of heart attacks by using the hormone
replacement therapy (HRT).
A new study had shown that HRT does not reduce risks of heart attack,
and may even increase the risk. Furthermore, the study found that HRT
increases breast cancer risk.
"If the experts making the recommendations had understood the principles
of evidence-based medicine," says Dr. Brian Haynes, Professor of
Medicine at McMaster University, "the family doctor would not have got
herself, and her patients, in so much trouble."
Traditionally, doctors have not been taught how to understand original
research articles. That means they could not independently decide
whether evidence was strong or weak. They were at the mercy of experts,
or pharmaceutical representatives.
Perhaps even worse, doctors had limited training in helping patients
weigh up the benefits, and risks, of alternative therapies.
Many research studies had suggested that HRT could lower cardiovascular
risk. But the research had used weak study designs.
So what is a "weak study design"? Consider a study of whether hospitals
keep people alive, or kill them. The study shows that people are more
likely to die in hospital than in the community. So, hospitals are
hazardous, right?
We would laugh at that conclusion. More people die in hospitals not
because they are risky places, but because people in hospital are sicker
than people in the community.
This problem plagues all "observational" studies. If people taking a
treatment are healthier than people who don�t, we may falsely attribute
life-saving properties to that treatment.
How do researchers solve this problem? They decide who gets treatment
by, in effect, flipping a coin. This "random allocation" makes sure that
people who get treatment are, at the start of a study, no healthier or
sicker than those who don�t.
This explains the HRT results. In earlier, observational studies, women
taking HRT were at lower risk of cardiovascular disease than women who
didn�t take HRT. They may have exercised more, had less stress, or been
wealthier, all factors associated with better health outcomes.
So, it looked as if HRT reduced heart attacks when, in reality, women
taking HRT were destined to have fewer cardiovascular events whether or
not they took HRT. The randomized trials revealed the real situation.
There are other examples of doctors going wrong because they didn�t
respect the principles of evidence-based medicine (EBM).
Othopaedic surgeons were sure that they could help patients with painful
osteoarthritis of the knee by, in effect, washing the knee out. They
inserted a surgical instrument, an arthroscope, into the knee and
"washed out" chemicals they believed caused the pain.
In 2002, researchers reported the results of a randomized trial in which
patients received the real surgery or "mock" surgery in which surgeons
made a cut in the skin, but never used the arthroscope. The result? No
difference in pain at any time during two years of follow-up.
How could doctors have gone so wrong? They underestimated "placebo"
effects. We often feel better when we receive a treatment we believe is
helpful, even if there is no real effect.
McMaster has been a world leader in helping understand such issues. In
fact, McMaster may be the single institution that contributed most to
EBM.
Dr. Haynes, for instance, has done more than any other researcher in the
world to get the evidence to doctors in a clear, usable way.
"EBM still faces big challenges," Dr. Haynes says. "We need to do a
better of job of helping doctors understand EBM principles, and making
sure they have the best evidence at their fingertips. Expert
recommendations must reflect the best evidence, and consider patients�
values and preferences."
Other health workers can also benefit from understanding and applying
evidence-based principles. Nursing, for instance, has its own share of
myths. For instance, randomized trials have shown that the widespread
practice of shaving patients before surgery increases, rather than
decreases, wound infections.
McMaster Professor of Nursing Alba DiCenso has played a key role in
establishing evidence-based nursing. Her soon to be released text book
will provide the ideal guide for nurses interested in using evidence to
guide their practice.
McMaster has also been a key participant in an international effort, the
Cochrane Collaboration. The Collaboration�s goal is to bring
high-quality evidence summaries to doctors and patients. You can join
the Cochrane Consumer Network through their website,
http://www.cochraneconsumer.com/
As they better understand EBM, experts and doctors will create fewer
problem stories like HRT and arthroscopic arthritis surgery. And, as a
patient, you will receive more accurate information about the benefits
and risks of the treatments that medicine has to offer.
Gordon Guyatt MD, FRCPC, lives in Dundas, Ontario. He is an academic
physician at McMaster University's Department of Clinical Epidemiology
and Biostatistics and Department of Medicine.
Related addresses: URL 1: www.cochraneconsumer.com
Back to top of Document
