 |
Dr.
John Wennberg
Director, Center for Evaluative Clinical
Sciences at Dartmouth |
|
| Dr.
John Wennberg is the Director of the Center for
the Evaluative Clinical Sciences at the Dartmouth
Medical School. As an expert in the application
of epidemiological principles to the healthcare
system, he co-founded the Foundation for Informed
Medical Decision Making in Hanover, N. H. Through
various studies, Dr. Wennberg has proven the importance
of patient preference in the rational choice of
treatment. |
|
|
|
Dr.
John Wennberg
Taming
Unwarranted Variations in Health Care Delivery
The importance of patient preference in the rational choice
of treatment.
There
was a time when most people assumed a physician was acting
not only in the best interests of the patient, but also of
society according to Dr. John Wennberg, Director of the Center
for the Evaluative Clinical Sciences. In other words, if the
supply of hospitals, testing centers or other resources available
exceeded patient needs, some would eventually close, "like
having too many shoe stores or too many Wal-Marts in a particular
market." But Wennberg's life's work has shown that too much
medical care is as dangerous to patients as too little.
The
Vermont Experience
During
the 1960s, Dr. Wennberg began examining the rates at which
different hospitals in Vermont were performing tonsillectomies.
With four young children, he had a natural interest in this
topic. What he found surprised him.
In one
area, the rate was so high that about 70% of the children
in a particular school system were having their tonsils removed
by the time they were 12 years old. In the neighboring town
(where Wennberg's children attended school), the rate was
about 20%. "By 1973", Wennberg says, "there
was a tightening of the general variation throughout the state
. . . without any intervention on our part other than simply
feeding back the information." Wennberg's team discovered
that two physicians in one high-rate area Morrisville had
initiated a second opinion procedure along with an extensive
review of the relevant literature.
This
led to a central question: which rate is right? In many respects
Wennberg has been studying the answer to that question ever
since. One thing is certain: proper measurement of outcomes
is key.
Outcomes
Research for Prostate Disease
 During
the 1980s, with clinicians in Maine, Wennberg studied why
prostatectomy rates for benign prosthetic hyperplasia (BPH)
varied significantly. In some parts of the state only 15%
of men were having that operation by the time they reached
75 years of age. In other parts of the state the rate was
up to 65%.
Wennberg
began discussing his findings with well known leaders in what
was then the relatively new field of outcomes research: Fred
Mosteller, John Bunker and Benjamin Barnes, and together they
identified two sets of issues. First, the probability side:
what happens when you treat condition Y with treatment X?
Second, the value side: given the outcomes associated with
a particular treatment, are the risks and benefits worth it
to the patient, compared to the risks and benefits of an alternative
treatment?
What
they found was that the conventional theory for doing prostate
surgery was not factually correct. Most surgeons assumed operating
early reduced the risk of early mortality. But outcomes research
indicated otherwise. Mortality rates were actually slightly
better for those who did not have the surgery. While early
operations improved urinary function it was accompanied by
an increased risk of incontinence and/or reduced sexual function.
"Through
outcomes research we were able to clarify that this is a quality
of life choice, not a preventive choice," said Dr. Wennberg.
"You're not preventing something bad from happening. What
you're doing is you're improving the quality of life but there
are tradeoffs. That led us to the next phase-to begin to develop
methods of informing patients about treatment options through
shared decision-making tools."
When
patients are actively engaged in the decision processes, they
typically make choices that are more in line with their own
values. In the example above, having prostrate surgery was
no longer a decision based on urine flow or other biomedical
properties, rather it depended on understanding what the patient
valued more-relief from symptoms or avoiding the risk of side
effects.
Which
Rate Is Right?
Generally
speaking, the inclusion of the patient in decision-making,
when it's a risky situation and there are real tradeoffs,
will lead to more conservative treatments than when one delegates
the decision to physicians. This is supported by a study done
in two staff-model HMOs, looking at the rate of BPH surgery
prior to shared decision-making and after. It showed about
a 40% drop in patients choosing surgery. In that context,
the "right rate" would seem to be the rate that's found in
a conservatively oriented practice.
The
Dartmouth Atlas
During the
1990s, using a large database of Medicare and other claims data,
Wennberg and his colleagues representing diverse disciplines
including epidemiology, economics, and statistics, studied variations
medical care use and distribution across the United States.
Among their findings:
- Eye
exams for diabetics (1999- 2000) varied between 30% and
70% depending on where the patient lived.
- Major
leg amputations (1998-2001) for non-black patients were
1 to 3 per 1000, but the rate for black patients was as
high as 9 per 1000 in some parts of the country.
- Coronary
bypass surgery rates (1999-2000) varied from 3 or 4 per
1000, to as high as 11 per 1000, again depending on where
the patient was treated.
These
findings seem to indicate that region and the supply of medical
resources is a much more important predictor of probability
of treatment than either sex or race. They also suggest there
may be extensive underuse of effective care in the Medicare
population.
Supply-sensitive
Care
"All
the evidence is that provider opinion, not patient opinion,
is the important driver of the decision to have procedures,"
notes Dr. Wennberg. " Many people believe that medical behavior
is driven by theory, or by evidence. But the data doesn't
support that. What happens depends on where one is practicing
and the capacity of the system in which one is practicing.
Behavior is dictated simply by availability. It's very hard
for physicians and administrators and, probably, patients
to accept that."
"Wouldn't
it be a shame if the available supply of doctors wasn't being
used?" Wennberg asks. Actually, outcomes are slightly worse
in regions with more intensity of care. One reason for this
is medical error. "If you hospitalize people twice as often
in region A than in region B, and medical error is a constant,
you would expect twice as many problems in region A," he says.
Wennberg
points to an important link between supply-sensitive care
and the cost of Medicare. In 2000-2001 some regions were spending
less than $4,500 per person, while others spent almost $10,000
per person, overall. The difference is not attributed to surgery,
says Wennberg, but in the way chronic disease is being managed.
The remedy
for supply-sensitive care is to rid the market of excess capacity.
This will be difficult to do given the wide-spread "more is
better" belief. And yet, Wennberg suggests, we need to do
this and more: reform requires population-based, provider-specific
information. His latest research, Care of dying patients in
the best U.S. hospitals, British Medical Journal, 03/13/04,
has converted data into provider-specific information. It
identifies all the people who died within a given year in
the Medicare program and follows them back, for the last two
years of life, to the hospitals that they most frequently
used. Wennberg noted striking differences.
 There
were huge variations among several top academic medical centers
in the way they treated chronically ill patients. The number
of days patients were hospitalized ranged from 10 to 27; physician
visits from 22 to 76. Some facilities were reimbursed a lot
less in the 18- to 24-month period before death than others.
Medicare reimbursements ranged from $2,000 to $8,000. A trend
among the lower-reimbursement facilities was the use of more
primary care physicians versus specialists.
Wennberg
notes, "The real driver of cost in fee-for-service medicine
has been the internal needs of institutions. Once the patient
is hospitalized, doctor visits are much more frequent. It's
very efficient for the hospitals in terms of their resource
allocation for doctors. Yet there are still large differences
within a single state; within a given market. This opens up
some fairly interesting prospects for selective contracting,
and other ways of identifying efficient providers, and rewarding
them for their efficiency."
Allocating
Resources Effectively
Wennberg
acknowledges that any hospital today attempting to balance
supply with demand by downsizing faces revenue implications
that are catastrophic under the current fee-for-service system.
"We need a solution that works for providers, works for the
government, works for private insurers, and then we could
begin to get the incentives of the healthcare organizations
aligned with the incentives of the payors to improve quality,"
he says. In a reform proposal that addresses some underlying
causes of Medicare funding problems Wennberg suggests there
should be experiments or demonstration projects in Medicare
that would allow provider groups who agreed to deal with unwarranted
variation in medical care to petition for revisions of the
fee structure.
Centers
of Excellence that remain in the supply chain must focus on
providing effective, patient-sensitive care by
- Eliminating
under service of effective care
- Reducing
medical mistakes
- Learning
what works, outcomes research
- Assuring
informed patient choice or shared decision making
- Promoting
conservative practice patterns when more isn't better, and
achieve efficient allocation of resources, which means become
accountable for capacity
The
views expressed by the speakers, which are reflected in these
articles,
are those of the presenters themselves and not necessarily
those of American Re.
|
