|
By
Susan Rodsjo
If a Nobel Prize
were given for healthcare informatics and health system reform,
fans of Lawrence Weed, M.D., believe this provocative one-man
think tank would be writing his acceptance speech on a plane
headed for Stockholm. Known as the father of the problem-oriented
medical record, which was adopted throughout the country in
the 1970s, Weed has spent the last 50 years contemplating
the current state of chaos in medicine and formulating a vision
for reform.
The sweeping changes
he proposes are based on the notion that no physician, regardless
of intelligence, can remember every possible reason for a
given symptom and adequately process diagnostic information
in his or her head. Weed criticizes
medical schools for perpetuating a system that relies on the
unaided mind.
"He has called
medical education an effort at stuffing information into the
brains of young people," says John Eisenberg, M.D., director
of the Agency for Health Care Research and Quality, which
is charged with reducing medical errors. "Instead, we
need to teach them how to find the information and use it.
It's a simple and elegant concept," says this
longtime supporter of Weed.
A
Logical Foundation
As Weed illustrates,
"Patients have no obligation to develop problems that
fit with what you learned 10 years ago and within the 85 percent
you got correct when you passed your exam. This memory-based
system is fundamentally flawed." His criticisms are solidly
based on more than 40 years' experience working at medical
schools such as Columbia, Johns Hopkins, Case Western Reserve,
Yale and the University of Vermont.
Weed's vision led
him to develop the idea for "problem-knowledge couplers,"
decision support software that couples patients' symptoms
with diagnoses and management options culled from medical
literature. For example, if a patient presents with chest
pain, the computer asks a series of relevant questions, then
requests pertinent physical findings and lab results. Each
answer leads the computer to choose one of 80 causes of chest
pain, process the information and organize it into a list
of possible diagnoses. It also points out the particular characteristics
relative to what the patient is experiencing and recommends
various treatments.
Weed says that
such tools, which virtually eliminate the need to rely on
physicians' unaided minds, are the foundation for the sweeping
reforms he proposes. His concept culminated in the formation
of PKC Corp., a small software company in Burlington, Vt.,
that has developed more than 75 problem-knowledge couplers.
Although PKC, started
in 1982, may not be well-known today, that could change considering
its current client base. The U.S. Department of Defense and
the U.S. Department of Veterans Affairs are already using
problem-knowledge couplers, and both plan to incorporate them
into the Government Computer-Based Patient Record (G-CPR)
under development. Informatics experts have speculated that
this multibillion-dollar information system could serve as
a model for other healthcare organizations.
A
higher purpose
Whereas most entrepreneurs
start businesses with profit as the No. 1 goal, Weed's writings
from the last 30 years leave the impression that he holds
more idealistic goals. His product is just one step in realizing
his vision.
In the 1970s, Weed
developed the first computerized problem-oriented medical
record as principal investigator for a National Center for
Health Services Research and Development-funded project at
the University of Vermont, Burlington, called the Problem-Oriented
Medical Information System (PROMIS). At the time, computers
were so new to medicine that he described them in his writings
as "television-like, touch-sensitive screens." He
could have run with this product, but he realized it wouldn't
change the practice of medicine, but rather, only automate
it. "There's no point in automating a record until you
can control the inputs. Otherwise you're automating chaos,"
says Weed.
The
long haul
Asked why he continues
to work full time at age 76, Weed explains that changing a
paradigm is a long-term project. "You can't just quit
when you reach retirement age, although many days I wonder
if I have rocks in my head."
Weed is often referred
to as a visionary genius. "He's brilliant," says
Bob Boysen, vice president of information systems for Sisters
of Charity, Leavenworth, Kan. Boysen also calls Weed a rebel:
"He has challenged the medical profession's sacred oaths
and egos, and he's had the strength of character to withstand
the criticisms. It's unfortunate he may not receive the recognition
he deserves during his lifetime."
To truly understand
Weed's vision requires time spent reading his ample writings.
In fact, Weed expressed concern that a short article will
provide only a fragmented understanding of his ideas culled
from 50 years of work.
"Until people
take the time to fully understand the vision and start implementing
it, this chaos can go on indefinitely. You can't go drifting
through a few notes of a symphony and get a sense of the composer's
body of work," he says.
In the following
text, Weed discusses his vision and the events that led him
to formulate his ideas. If your curiosity is piqued, you may
wish to peruse his insightful, albeit lengthy, works (see
sidebar below).
S.R.: In the 1970s,
you led the development of a computerized medical record (PROMIS).
How did you come to the realization that automation wouldn't
fix medicine's problems?
L.W.: We had a
computerized ward at the University of Vermont in the '70s
with a better automated record than most people have now.
We had 55,000 displays of information with a branch to all
the symptoms of each diagnosis. I could sit in my office and
say, "Bring up problem such and such with all the plans
and progress notes." Once I could do that, I always found
mistakes. That's when I realized if you solve the mind's memory
problem with a computer, you uncover a processing limitation
with the human mind that's worse than the memory limitation.
The human mind
can't do it alone. We've got to build tools that will tell
physicians what to ask, then tell them what it means. If you're
going into astronomy, you better have a telescope. You might
see better than I do, but compared with the telescope, we're
both blind. Until the profession
accepts that computers process knowledge better than the unaided
mind, we're going to have a mess.
S.R.: You have
been highly critical of medical schools. Why?
L.W.: As Robyn
Dawes has said, "States license psychologists, physicians
and psychiatrists to make (lucrative) global judgments. People
have greatly misplaced confidence in their global judgments."
In other words, if you sleep through lectures for 4 years
but get at least 75 percent on your exams, we'll give you
an "MDity" and a license to go out and say, "In
my opinion. . . ." And you can make a very good living
doing it because the public has no way of knowing.
Medical schools
should never again examine students on what they know. That
is a bottomless pit. Examine them on how well they behave
with the right tools, whether they use the tools available
to them. Schools do just the opposite. By examining what you
know, they inculcate the idea that it's all right for you
to use the limited knowledge you have, throw it at individuals
and then make mistakes. Universities are the root of the problem.
People think I
don't appreciate the greatness of medicine. Physicians are
transplanting hearts, putting in new hips and new knees. It's
incredible what they're doing! But just because they developed
the Concorde and you can get from New York to Paris in 2 hours
doesn't justify letting a few planes crash in Kansas. Most
of the greatness in medicine has to do with hands-on skill.
I'm never critical of doctors. I'm critical of the university
system that produces them.
S.R.: Tell me about
your ideas for creating a patient-centered medical system.
L.W.: You have
to understand that a doctor doesn't take care of patients.
A system takes care of patients. If you want to travel, you
don't say, "I need a travel person to take care of me."
From age 5 on you know there's a system. I know enough to
get myself to the airport, then I turn over responsibility
to the pilot, the radar guy, the air traffic controller. I
assume they've been licensed to do their part well. That's
the way medicine should be.
I might call a
travel agent for some help, but I know the possibilities and
make the choices. Right now the public is not given the means
to make choices in terms of their health. Patients should
all have their own records. They should know how to discover
hypertension and other problems long before they become serious.
We've been led
to believe that doctors went to medical school so they'll
know what to ask. And since they have all the information
and all the authority, they'll tell the patient, "You
ought to have a prostatectomy." But it's the patient
who undergoes the surgery -- who may become incontinent or
impotent -- not the doctor. Doctors had better line up all
the facts and let the patients make their own choices.
S.R.: Would it
be beneficial for patients to have direct access to decision
support software?
L.W.: Yes. We're
building knowledge couplers for the Web. The questions patients
can't answer themselves will be there but will be marked with
a symbol. For example, "Do you have a systolic murmur
in your heart?"
S.R.: You have
criticized medicine for conducting outcomes studies. Why?
L.W.: Medicine
has been doing outcomes studies for years. When I ask why,
they say, "Don't you think it's important to know there's
four times as much prostate surgery in Salt Lake City as in
Denver?" What good is that? I don't know whether to move
to Salt Lake City so they won't miss my prostate cancer or
whether I should move to Denver so I won't have unnecessary
surgery.
Until doctors control
inputs, we shouldn't spend money on outcomes studies. The
Food and Drug Administration (FDA) should release a new drug
and say, "This drug will only be released to those who
use [decision support software], and the computer will tell
you the conditions under which the drug can be used. And since
you're controlling inputs with these tools, we'll know after
10,000 cases whether we should withdraw our approval because
of adverse effects." Right now the FDA releases a drug
and doesn't know what physicians are doing once they get their
hands on it. The whole system has to be brought under control.
S.R.: Recently
there's been much talk about mistakes in medicine. What can
be done?
L.W.: There will
always be errors if you allow people to practice medicine
based on what they remember and process in their heads. Errors
will continue to happen until we control the inputs and get
a corrective feedback loop. Even then there might still be
errors, but we will notice them if we study our data carefully.
Every system without feedback loops runs wild, and medicine
has been running wild.
S.R.: Do you feel
physicians resist the changes you propose?
L.W.: I can't judge
what the profession thinks, because I've already come to the
conclusion that if we use these tools, our picture of the
profession will change.
If you switch from trains to airplanes, you wouldn't say,
"Well, will the people who own Pennsylvania Railroad
accept it?" That's not the question. Naturally, the people
who own the railroad will be threatened by the airplane. It's
a very complex sociological problem.
So now the tools
are getting out. The biggest problem is that doctors control
the system. They're not going to give up authority easily
to these new tools.
S.R.: You propose
huge changes to the practice of medicine. What do you predict
might bring about these changes?
L.W.: There are
four parts to a total system: tools, philosophy, disciplined
users and leadership. Weed and PKC Corp. provide the philosophy
and the tools. What's going to bring about the proper change
is a strong leader who reads my materials and says, "I
will do it." We have a good architect. Now what that
architect needs is a good contractor. And that contractor
needs some good carpenters.
More
about Lawrence Weed
- President
and founder, PKC Corp., Burlington, Vt.
- Professor
Emeritus of Medicine, College of Medicine, University of
Vermont, Burlington
- Medical
degree from Columbia University College of Physicians and
Surgeons, New York, 1947
- Director
of PROMIS Laboratory at the University of Vermont, 1969-1981
- Founding
fellow of the American
College of Medical Informatics
| 