Uncategorized

QUALITY/TECH: A nice conversation with Brent James

This is possibly the most interesting podcast yet on THCB. And it’s certainly the longest. if you didn’t have the time to listen to the interview with Brent James, here’s the transcript. I really recommend this one–there are so many amazing nuggets that if you care about health care in the US at all you owe it to yourself to read!

Matthew Holt: This is Matthew Holt with The Health Care Blog, and I’m back with yet another podcast and this time it’s really very exciting for me that we have one of the pioneers of the entire medical safety and industrial process of medicine movement in the U.S., Dr. Brent James, with us this afternoon. Brent, good afternoon. How are you?

Dr. Brent James: Good afternoon. It’s a delight to be here.

Matthew: Great, great. Just by way of introduction for those who don’t know, and I’m sure most of my readers will know—I hope they do—given that it comes up enough in the blog. Brent, your official title is VP of Medical Research for Intermountain Healthcare? Is that correct?

Brent: That’s correct, and I’m the Executive Director of the Intermountain Institute for Healthcare Delivery Research.

Matthew: Great. And I would say that that sounds very well and good, but in fact that is really understating Brent’s impact. He’s both at the regional level in Utah with Intermountain been largely responsible with his team for some really dramatic change in the entire way  clinical care is being delivered on the in-patient side, and has had a lot of great information published and distributed out of that. On the national level, Brent, you’ve been involved in both the Institute of Medicine and the more recent Citizens Working Group in Healthcare, and there’s probably some other things you’ve been involved with. I don’t have them on the tip of my tongue, but certainly you’ve been a very visible player on the national level. In addition, and we’ll touch on this at some point in the conversation, you’re currently involved with the Institute for Healthcare Improvement’s — Don Berwick’s organization — new campaign which was announced last week for the Five Million Lives. Is there anything else big and important I’m missing from what you do? [laughs]

Brent: I think that’s definitely enough. It certainly keeps me busy night and day.

Matthew:
Let me start with something that was said last week at the IHI meeting,
which I had put in the blog just yesterday. Larry Weed, who is also
another great pioneer of patient safety and medical semantics movement
said, something along the lines, that “you know that the rate of
prostate cancer surgery is four times in Denver than what it is in Salt
Lake City, but you don’t know whether you should move to Denver to get
your prostate cancer taken care of properly or if you should move to
Salt Lake City to avoid unnecessary surgery.” I said that it seems to
me the data coming out of that is not quite true, and that we do know
there’s a lot of stuff that goes on that perhaps is unnecessary, and
it’s more likely you’re better off in Salt Lake City than you are in
Denver. Without castigating your colleagues over on the other side of
the Rockies, do you think that’s roughly true?

Brent:
That is roughly true. The person who did the main work was Jack
Wennberg. At least we give him credit for a massive body of work. It
had to do first with variation analysis. Jack carefully analyzed
delivery patterns across the United States, now he’s extended it to
Canada and has done some very nice studies in Europe. What you see is
massive variation. Now for some clinical conditions, there’s almost no
variation. Total hip arthroplast, artificial hip joints for hip
fracture: almost no variation. Inguinal herniorrhaphy, hernia repairs:
almost no variation. But on the other hand, for some other medical
conditions: as high as thirty fold variations from one community to
another.

He then asked the question, what benefit is associated with this?

He divided the world into two big chunks. One he called ‘Supply
Induced Demand’. It corresponds to thirteen medical conditions: top of
the list is hospitalization for congestive heart failure, while the
bottom of the list is hospitalization for AIDS, and lots of things in
between. In those circumstances he showed massive variations in
resource utilization for care delivery. It was the first thing. He
showed that it was tightly tied to the supply of specialists and
hospital beds. That’s why he calls it ‘Supply Induced Demand’. It was
really Dr. Elliot Fisher, at Dartmouth, who showed that the high
utilization communities were getting worse medical outcomes than the
low utilization communities. So high resource consumption directly
connected to a worse clinical result.

Now the second big category that he
covered was called ‘Preference Induced Demand’. That corresponded to
eight major surgical conditions. There are people like Dr. Al Mulley at
Massachusetts General Hospital, David Wennberg up in Maine, who
developed formal decision analysis tools so that they could measure and
validate the quality of patient decision making, and more important,
how comfortable a patient was with their decision in the long term,
after they received the care, when they had the retrospective scope to
give them perfect hindsight.

You see? They discovered that for
these preference sensitive conditions, if the patients were given a
fair choice, as validated by these formal decision analysis tools, and
hindsight, the utilization rates dropped by about forty to sixty
percent, suggesting that the physicians involved were much less risk
adverse than the patients that they treated, and again, this concept of
over-utilization. That being said, the other thing you need to know,
Matthew, being familiar with Utah and our utilization data, it just
depends on which condition you look at. Yeah, we look pretty good on
prostate data. On the other hand, we have some very high utilization
rates in some of our hospitals in some of our communities, around
things like spinal fusion surgery for mechanical low back pain. So I
wouldn’t want you to think we’re uniformly good across the board. We’re
not. We’ve got lots of room to improve.

Matthew:
Now, let’s talk a bit about improvement. There are two aspects to this.
One is, I’ve kind of joked around on the blog about some of the
statements you’ve made about the financial consequences of the types of
changes you’ve made. Could you, first off, describe a little bit about
the kinds of changes within Intermountain that you’ve made? How they
happened, how they came about and what kind of procedures, process, and
technology you’ve had to put in place to make those changes?

Brent: We really went through two phases.

The
first phase I called our Project Phase, and it ran from about 1987,
when I met Dr. W. Edwards Deming and it kicked us off on this track,
through about 1995 or ’96. We were running a big training program
called the Advanced Training Program in Clinical Practice Improvement.
We were using it to change the culture of our administrative and
clinical staff. It was a much better understanding of Quality Theory
and what it could mean, and a big debate about the nature of the
healing professions and what our core professional commitments really
mean. Well, you have to do a project to graduate, and we had some truly
outstanding projects, led by a physician or a nurse, coming out of the
course. It was clear that it worked.

Dr. Deming had made the
case that in many areas, not all, but in many, that as you improved
your clinical outcomes it will cause your cost of operations to drop.
We fairly quickly proved he was right in trials, where we demonstrated
that. We just started to track cost outcomes, side by side and in
parallel with our clinical outcomes. Our problem was that we weren’t
able to deploy. In 1995 Bill Nelson, he was then our Executive
Vice-President he’s an accountant by training, our CFO technically, asked me for a summary. I could identify about 65 major
clinical improvement projects that had happened inside Intermountain
with proven results. It was Bill asking, so I tended to focus on the
cost outcome results, and it was impressive. It showed a massive return
on investment. The problem was
those projects happened in a single location, around the physician or
nurse who came through the course. They implemented it in their
practice, but it stopped dead there and went no further.

I
could, in a reasonable way, point to $30 million in structural cost
savings in those 65 projects that our finance people would agree were
real. The trouble is if we were able to deploy, it wouldn’t have been
$30 million, it would have been somewhere in the range of $100-$150
million a year, in structural cost change. The data were compelling.
And Bill wanted to achieve those savings on the behalf of our patients.

What
we were forced to examine as we moved into the second phase how do you
build a quality control infrastructure? Joseph Juran described it best
when he contrasted Quality Improvement versus Quality Control or
Quality Management. Of course, they interdigitate. The Quality Control
is that infrastructure that enables you to identify and consistently
manage routine operations on a quality foundation. So in 1996 we
decided to build that in-site at Intermountain. That was the second
phase. We first did something called a Key Process Analyses. A crazy
thing: we found that for about ten percent of the clinical conditions
we treated they corresponded to work processes, I called them Condition
Based Work Processes. They accounted for well over ninety percent of
all of our care delivery. So, it wasn’t the 80/20 rule. It was the
90/10 rule. Concentrated like crazy. Well, first was to figure out what
those work processes were. Second thing, we needed a data system to
figure out what was going on. You know, crazy thing, we tried to start
clinical management inside Intermountain twice and had it fail twice.
Each time it failed it cost us five or ten million dollars in sum
costs, and it took out a senior vice president for medical affairs, the
people I worked for actually. Might be a message in that!When they asked us to come back and try it a third time, Dave Burton
and I, tasked with this little project, the first thing we thought to
do, we went back and did a careful autopsy on the first two failures,
given that we were asked to walk down that same scary road one more
time. What we discovered was, was in a neat instance, with our clinical
leaders who were willing to step up and try to manage sharp people, but
we had unthinkingly assumed that they could use the same data systems
that we’d traditionally used to manage the facilities to all of those
financial administrative data. We’d assumed those data were sufficient
to manage critical care delivery. And, on careful examination, it turns
out that just wasn’t true. It was a very bad assumption.

So, our
second task was figure out a way to build an outcomes tracking system
to give those frontline clinicians the data they need to manage the
clinical process. And, we did that.

The third element is we
built an organizational structure. We used those data to hold people
accountable for their work, and, more important, to enable us to set
and achieve clinical outcomes improvement goals. That was about 1998, I
think.

And then the final piece, the last piece, we called it
aligning financial incentives. We discovered that most of the time you
actually achieve cost savings, the money all went back to purchasers as
windfall savings. I mean, your costs drop, but your revenues drop as
far or further.

Don Barwick hates it when I say this, but
clinical quality improvement is a fast way to the poor house if you
haven’t figured out a structured way to harvest back some of those
savings in administrative tasks. So we figured out how we could, at
least for commercial purchasers, begin to contract on the basis of our
quality improvement efforts and get some of the savings back to make
this thing a viable enterprise. Those are the four big steps, though.
Matthew.

Matthew:
Right, right. And that last part, I think you were quoted in the New
York Times, I want to say, a couple years back, or maybe more recently
looking at the correlation between moving
patients between different DRGs in Medicare, and the cost savings that
come back…I think you’ve had a big push on when patients aquire
infection after surgery , and you’ve actually saved a lot of money, but
you end up losing money on the end because Medicare reimburses at a
lower rate.

And that’s kind of one of the many data points, one
of the many sentinel events that has helped move the needle towards pay
for performance, or quality based reimbursement, or whatever you want
to call it. Do you get the impression that that train is now leaving
the station? Or, are we still a long way from actually changing the
overall sentiment to making people want to put in place the kind of
things that you’ve done, given that they are hard work, and if you
don’t have that change, you’re going to lose money?

Brent:
Yeah, it’s very clear that that train is leaving the station. There’s
no question that we’re going to see a lot more pay for performance in
the future. The trouble is, it’s a very complex topic, and I think
there’s a good chance that some of the pay for performance efforts
won’t bear the desired fruit.

One way of looking at it, there’s
two ways of implementing shared savings models. Probably more than
that, but two that I know off the top of my head though. One is called
a premium model. Usually it involves getting a bunch of providers and
ranking their performance. And, then, on the basis of their
performance, giving a premium to those who do particularly well. And,
perhaps, in the long haul, some sort of a financial distancing of those
who do poorly to counterbalance those who get the extra payment. It’s
the Premier demonstration project that’s running right now for
Medicare. For the conditions they’ve identified, if you’re in the top
twenty per cent, you get a one per cent premium. If you’re in the top
ten per cent, you get a two per cent premium.

Now, stand that in
contrast to models that instead of trying to rank and then give a
premium, to models that are called shared saving models.

Imagine
that I run a big improvement project and I have good enough data
systems that I can actually measure the cost savings. And then what you
do is say, "Well, that’s money on the table. Let’s figure out how we’re
going to split the savings between healthcare purchasers, hopefully
going back to their patients, and the physicians and hospitals on the
care delivery side that are making that possible."

I much prefer
shared savings models. They’re much more attractive from a pure
technical standpoint, but they require some very sophisticated data
systems. That’s the downside. Most places couldn’t run them.

Matthew:
And also, to a certain extent, you’re talking about a shared savings
model is such that you capture enough of the care in some well
integrated way in such a way that you figure out who is responsible for
what savings. At some point, the end purchaser – it probably makes
sense for self insured employers and for some commercial insurers –
that at some point, the end purchasers have got enough of their arms
around it that they can actually see the benefits over a population.

Because
a lot of what you’re doing is talking about changing from dealing with
the acute issues to avoiding that at all by doing something better than
that earlier on. I’m wondering even within Medicare, there is a lot of
discussion around pay for  performances; I see a lot of concern around the
ranking systems and the AMA and others are already getting involved
into that. But is it your sense that we can get to a rational shared
savings model without larger reforms on the insurance side? Where do
you think that needs to go?

Brent:
I believe that we can get there. I believe that it will happen first in
a few leading big systems with good data systems, but then we’ll be
able to learn from those and generalize the models. For example, if you
look at the current efforts in California, I think 14 different health
plans– I’d have to review Steve Shortell’s data — leading that project
but there are some very, very positive things in there. A number of
other efforts across the nation as well, participating broadly, so
yeah, we’ll get there.

The thing that health care providers need
to watch out for is that as that ground shifts underneath their feet
from the current fee-for-service payment system, some sort of the cost
plus system into this new quality-based system, it could profoundly
change their financial models and if you’re not on top of your care
processes, if you’re not organized to manage the care, and measure it,
you could get hurt very, very badly in that new, evolving environment.

Now is the day to prepare. You know the old joke: It wasn’t raining when Noah built the ark.

Matthew:
Right, right. That’s actually a very interesting piece, because if I go
back to the classic book on this whole issue, in which I know you
featured, which is Michael Millenson’s "Demanding Medical Excellence",
he brought up a number of instances where people had changed too early.

Brent: Yeah.

Matthew:
You could argue that with a certain cases, Intermountain was almost one
of those. Now you guys have deep resources, strong culture and your religious
heritage, some others don’t, but as you said, there are also some
change management processes that cost some people their jobs there and
you had to have some tough questions from the CFOs about saying, "Why
are we doing this thing? Just because it’s the right thing to do? But
it’s going to lose us money."

So I think there’s always this
tension and the majority of American providers are still, I think,
waiting to see which way and how fast the train is leaving the station.
Although I would agree with you, it does seem to me that it is
certainly getting decent levels of interest now from the government and
private players. But it’s still only getting a small fraction of the
overall reimbursement that is going into some kind of shared savings or
pay performance model.

So give me a sense of where you think in
Utah we are in that process or within Intermountain or within Utah.
Where are we nation-wide? What’s your sense of the timing of this?

Brent:
Let me just give you a few examples. We now routinely use these data as
we contract with our major purchasers. And we have a real advantage out
here that really helped ease the transition. We have a health plan that
belongs to Intermountain.

Matthew: Right. [laughter]

Brent:
It accounts for about 25 percent of our total care delivery, liberally.
I mean, it’s probably a little bit less than that. But at least for
that 25 percent, we can see the money coming back. Well more
accurately, the right way to think about this is Bill Nelson, who is
now our CEO took Dave Burton, my chief compatriot and doing this whole
thing.

See, Dave’s the physician that started the whole health
plan and he knows it cold. And he got Syd Pullson, an old childhood
friend, as it turns out, who’s now the CEO of our health plan and the
way I metaphorically describe it is that Bill stuck them in a room and
locked the door and told them they could come out when they had some
sort of a way of working together. And that’s about what it took.
Despite the fact that they were lifetime friends, just because of the
different business models, it required someone to say you got to do it.

Well,
out of that, we got some very good models that we then perfected
internally. It is amazing, Matthew, you talked about it conceptually
and most people it just leaves them cold. But when we got it to the
point where it was actually running where you can hold it up and show
it to people, show them the little gears turning inside of it and they
could see it running, boy, that makes a lot of difference and so we
started to take that and then share it with other commercial payers in
Utah. They found it quite attractive. They were willing to base
contracting decisions around it. It’s not like the negotiations go
away. We can still play hardball pretty hard with each other but it
just changes the foundations from which you were arguing and it was
enough to make this viable, you see.

Now, on the other hand, we
are in the midst of a partnership with Mayo Clinic and Dartmouth, same
exact aim. One of the goals that we really had was to generate enough
national data that other groups could see the gears turning,
particularly CMS, in terms of their payment system, and see a few big good systems that can show them how much it should cost
to deliver ideal care for congestive heart failure or the true
necessary cost for treating lung cancer or for doing that full hip
arthroplasty. People don’t have those data out there. We hope
that it will generate those kinds of data. That will be an accelerant.
That will really start things moving. It will also teach us what kinds
of data you need to collect. It is not just us. There are many other
groups who are moving in the same direction, you see. I hesitate to put
a timeline on it. I think this will continue to develop over the next
decade. Well, wait a minute, hopefully, not to change the subject, the
best I can tell listening to David Walker, the Comptroller General of
the United States, Medicare moves into financial crisis in about six
years and I think that’s the trigger right there. As Medicare gets into
deeper and deeper financial trouble, it will accelerate movement on
this massively.

Matthew:
Now, I think there are several interesting places we can take the
conversation. I think one of them is the issue of, is Medicare is going
to be a driver to change the behaviors outside of those players you are
talking about. I was very struck by the fact that in the recent Elliot
Fisher piece in Health Affairs which looks at ICU care, sorry end of
life care. In various big academic centers here, the Mayo Clinic came in at very
much the conservative somewhere in the lower end of the numbers, the
same way that InterMountain and some other academic centers did. Obviously
there were players on the other end who didn’t look so pretty on the
resource use side. I was thinking back in the early 90s, the Mayo
Clinic was cited by Bill Clinton and others as part of his health plan as being
the ideal way to be cared for. It has been obviously a great
institution probably one of the best known in America. Yet after 120
years of it being around not much of the rest of America looks like it.
So a lot of this is how do you get the data to prove it but it is also
how do you politically get this through and passed and I agree t hat this concept of Medicare going to insolvency should be a big
driver—but on the other hand, the political will to do stuff just
because bad consequences are facing us, isn’t always very apparent so I
don’t know how confident I am about how fast this is going to happen.

Brent: Yeah, Medicare has been in financial crisis for about the last 30 years, hasn’t it?

Matthew: Right.

Brent:
Not on the way out exactly. One way of looking at it, Matthew, if you
carefully examine group practice models they tend to do better than
other care delivery models in a pretty general way. One of the
questions you can ask is why do they do better? What’s going on here? I
think we are finally beginning to understand that. I think we are
understanding what made the Mayo Clinic successful and that’s one of
the key understandings of this whole debate right there. In
conversations with Mayo, well, I cross checked this to see if they
agree and it has to do with the fact, this is a fundamental thing that
the practice of medicine is shifting. Back in the old days, a single
physician expert could pretty much take care of a whole patient one
physician, one patient, physician-patient relationship, let’s keep it
uninterrupted, absolutely sacrosanct, but as medicine has become more
complex and more effective, it is no longer one physician, one patient.
If you look at our hospital here at Intermountain I think the minimum
number is four for hospitalized patients. I mean and that’s without
even counting all the support staff that gets involved. It has changed
into a team-based sport and I think that when you start to study the
team concepts that are involved, it starts to really explain why a Mayo
Clinic was successful. As you understand why they were successful, what
team factors were making them successful, you can leverage them, You
can start to say how can I make that little bit work better even than
involved at Mayo, you see. I think that potentially is a very rich area
of investigation moving ahead. I don’t know, what are your thoughts on
that?

Matthew:
Well, my thought is that’s dead right because there are a couple of
things that if you look at very basic levels. What did you have that
a collection of community hospitals with independent physicians don’t
have? You had the ability to put aside capital resources to study the
problem. Everyone agrees that there is a problem but somebody actually
has to put aside those resources to study it. You have to have a little
insulation from the day-to-day financial drama or the administration
of running an organization to be able to put these kind of studies
together. Then you have to have significant leadership and buy-in to do
the hard work and change it throughout the organization.

I
understand when you don’t really have an umbrella organization like a
Mayo Clinic or Intermountain or Kaiser Permanente or anywhere else,
making those kinds of changes amongst the majority of American
providers is very, very difficult because there is no either cultural
or financial relationship amongst them to make that kind of choice.
Then if you go back and you talk about Elliott Fisher’s work it’s
exactly true. He is quoted in recent Health Affairs studies about this,
as is Jeff Goldsmith, that in fact the people who do that kind of work
and practice more conservative medicine actually make less money and
don’t do as well financially.

So I think that although you can
show it’s true, it is very hard to figure out if it’s the right thing
to do. That’s where I think the real struggle is but I do think there
is one way out. This is a segue to the next question I want to ask you
which is when you go outside the world of health care what we are
seeing increasingly in other businesses is that smaller organizations
are able to adopt the tools and technologies that the larger
organizations have put in and used basically the Internet, information
technology and other skills and outsourcing and have been able to show
quality control and divergent ways of working together that had been
very successful. What we are now seeing when you talk a bit about the
project you have underway with GE, is we are now seeing that kind of
thing is starting to be developed in health care. I am wondering do you
think there is a solution for the optimization of medical care in the
US to be solved by information technology. Tell me a little bit more
about what you are doing with GE and the types of work you are doing on
the IT side.

Brent:
A little bit of background, as you know I am actually a member of
Department of Biomedical Informatics at the University of Utah, very
heavy informatics background. I worked in the fields many, many years.
I guess the short version, I believe six or seven years ago we had a
shift in Informatics where suddenly medical records started to take off
and a very interesting question is what happened, why that shift. Well,
having lived through it, it happened. It happened first at the level of
clinical decision support. We published that article in 98 about
antibiotic resistance at LDS hospital which showed a much better level
of antibiotic use and infection control associated with this decision
support. What lies at the heart of every one of those tools is
technically, it’s protocol, an evidence based best practice protocol.
So we started to play with those a bit and understand them a little bit
better.

We have come to call them shared baselines, the same
idea we were just talking about, the idea of clinicians working as a
team. So imagine that I have demonstrated as we have, that you can’t
write an evidence base best practice guideline that perfectly fits
every patient. Just take that as a given, the human beings coming to us
for care are too genetically different, response to disease is
different, response to treatment are different. They have different
resources, different preferences, different values.

On the other
hand, if I can get together all of the physicians who manage a common
high priority condition like congestive heart failure, lets say, or
recently the more common febrile infant strokes and putting them all in
teams. I get them all together-or at least their representatives with a
good communications chain. I get the nurses there too, the pharmacists,
therapists, technicians and even the administrators. We hammer out an
evidenced-based best practice guideline, but we decide to use it as
what is called a shared baseline.

This is a concept from Lee. It
was first invented in 1991 by a guy named Alan Morris, an internist at
RLDS Hospital here while he was doing a randomized controlled trial. He
was struggling to figure out how to run this trial when he came up with
this crazy idea of a shared baseline.

The way it works is that
you put it in place. It’s more than that you allow or even
encourage–you kind of demand that your clinicians adapt this shared
baseline to an individual patient’s need-given that we know they never
perfectly fit any patient. It’s kind of the opposite of cookbook
medicine.

What we have discovered, though—we have a bunch of
these running now—if the hospital administration will design around
this shared baseline, you can staff to it. You can train to it. You can
supply to it. You can lay out your physical plan.

If you are
willing to do that, it means that the process become reliable, so a
practicing physician, then, can focus their time and attention on a
relatively narrow band of the patient’s entire care—the part where they
can really add value, where they really make a difference, the most
productive part, the value-added part.

We’ve discovered in
serving a whole bunch of these at 95 percent confidence interval that
you will end up modifying about 5-15 percent of the shared baseline.
That makes the physicians massively more productive in terms of value
added. It has also been associated with much better clinical outcomes.

The
basic idea is that you try to take evidence-based best practice and
that shared baseline and you tend to make it the lowest energy
state—the default—if everything is left alone. The only reason as a
physician that you can get away with this is because the rest of the
process is reliable. You don’t have to bird dog every little step to
make sure that it happened correctly. So we went around building these
things and having some pretty good results with our physicians. It’s a
way of protecting your income if you do it well in these tough days.

Then
we discovered that that approach. It came out of the idea of how we did
decisions supporting the electronic medical records. You see, it’s a
protocol. Then we realized that at the heart of all our decision
support tools on the EMR you have the same sorts of protocols that we
treat as shared baselines. So we started to structure the electronic
medical record around them. All of a sudden the data automation
techniques that had worked so well in other fields began to work in
health care. So this huge leverage is complimentary, I guess. I have
come to believe very strongly that if you don’t understand shared
baselines you cannot effectively implement an electronic medical
record. Similarly, you really can’t fully implement shared baselines
unless you have that tool. They go hand-in-glove. It’s kind of the key
to using them effectively. You start to use it to work as a team and to
document around that.

So here we are. We’re figuring this out.
We’ve got it tied into our key process analysis. We knew that it was
going to require the third complete rewrite of our electronic medical
record system. We have built two so far – Build Help System and now
Help 2. It’s such a huge investment. It’s just a daunting task. We were
looking at what it was going to take. We had one false start with a
collaborator that didn’t work out. Eventually, it looked so big that we
went out and did what is called a build vs. buy analysis. We took a massive
effort for nine months to take all major computer system vendors in the
United States—well internationally as far as that goes, but most of
them in the United States—to say how well would they run in this
environment for clinical process management. The disappointing result:
We concluded that none of them could do it currently, and more
importantly, that none of them had the core infrastructure within them
to allow them to transition easily. We had made a decision to build our
own from scratch one more time. We were talking about budgeting on the
order of 50-70 million dollars a year to make this happen. It was a
huge investment. Right at that point, GE showed up just out the blue
sky. I mean, we’d already decide to go with the build. GE had their Logician
product. Pretty good product actually. Especially for scheduling and
billing services. They wanted to jump ahead. They were looking for a
clinical partner with help them do that. Frankly for us, they have
enough to be a reliable partner on the IT side. It meant, we got to
share development and maintenance cost across the groups instead doing
it all by our self. The partnership eventually came together but that
is the goal. Frankly, as I watched development, it is a generation of
electronic-medical record that simply does not exist on this planet
today.

Matthew:
Could you give a practical example of it, when we
talk about this baseline and altering from that, what does that mean in
terms of an example of physiology in patients or we are also talking
about patients in more general care management?

Brent:
Well let me give you first one from the out patient setting. We did our
key process analysis. We use that to start to build data systems around
key processes. We also use this to framework organization structures.
We call it "Clinical Programs." We sort of hire these medical directors
quarter time. We did it around families of related care processes defined by
groups of clinicians to work together.  So you
get these structure, you get these data systems,

For our community based medicine. We have about hundred out
patient clinics.We have something called "Primary Care Clinical
Program." set  up by Dr. Wayne Cannon, a pediatrician and Eileen Tippits on the nursing side. All of the
programs are position leader and what we call clinical ops administrator
because so much of the work happens at a support  staff self level. One of
the great big thing that we are doing is the  management of diabetes mellitus and we put together what we call a "Care Process" model.  It is an evidence based practice test guideline step one. Number
two, you build re-plot arguments or tools to build it in the workflow.
The primary rule of that level is thous shalt not destroy
clinical  productity.  This is designed to make work easier not
harder. As a third element, there is a formal method originally
published by The National Quality Forum by which you could use evidence
based practice guidelines to figure out the data system to manage the
process. It is called outcome tracking system; a dashboard, a balance
measure. In theory, If CMS for example build their national measures
off set foundation; it would find fully integrated data systems working
much better.

Fourth step, you blend it into electronic medical
report for decision support of clinical content. Fifth step, we
currently have 10 people whose only job in life is to develop education
materials for the professional teams and the patient around that best practice. We did that for diabetics. That means
if you are practicing medicine inside Intermountain or  in association
with us, about a third of  the docs involved are employed in our
medical group and about two thirds are community based independent
positions. a group of 1250 co-physicians. Whether you want it or not, every quarter you
get the patients’ reports.

First among them, we call them action
lists. This system maintains a registry of all of your diabetic
patients in your practice. You will see every patient on the list but it
will flag any patient who is not in ideal care in terms of frequency of
testing following the measures of elements; for blood sugar level, for
lipids, for blood pressure, for other things on the list. The way
people tend to use this is they know the patients they’re  caring for, they just read
the list and if they see a patient who’s not at ideal care in terms of
testing rates or level controls, they’ll flag them, mark them, either
add them to a care management nurse and say "get on it" if they have a
care management nurse in practice with them, or give it to the
receptionist and say "schedule it." Turns out to be really popular. It
means you’re managing the diabetes as a chronic disease, instead of
episodically as the patients choose to show up. And they really like
it.

Oh, yeah, and you can generate it on demand, we’ll push that every quarter, but most of the guys generate it monthly.

Second
piece of the care process model: initially this just ran on our
internal electronic medical records, but now it runs through our
results-review web-based system, you blend it into the workflow at the
level of your prep-care nurses, when you’re getting ready for a patient
visit. You just call the patient up on the computer. If the patient’s
diabetic, the computer will spit out what we call a patient care
worksheet. It highlights a patient’s hits relative to that chronic
disease… well, to any chronic diseases they have these days. So for
diabetes it will show basic demographics, it’ll show a full list of all
their active medications, it’ll show summary lab and history relative
to the disease, so it’s going to show you their "A1C" levels going back
(I think) seven measures. It’s going to show you their LDLs. It’s going
to show you their blood pressure levels — all the measures you need,
and then down at the bottom, in a passive voice, it’s going to
recommend ideal care, what you need to do during this visit. In fact,
that’s an easy form of computerized physician ordering. Because it’s
showing you the orders you might think about,That’s very popular.

Oh, the third element: every quarter we’ll show
you a comparative report about how you stack up against your peers
relative to performance in diabetes management.

Fourth element,
the same copy of that report goes to the medical director within the
clinical program; let’s just say that if you’re dragging the group
down, if you’re a little bit behind, very likely you’re going to get
invited out to lunch…

Matthew: [laughter]

Brent:
And the lunch topic’s going to be, "Here’s what Dr. X is doing to
really nail this one, can we work with your staff to get that set up
for you? Just so you’re state of the art in terms of performance that’s
going out to the community here. This is
what you use for contracting, you see?

Matthew: Right, right, of course.

Brent:
And lately… you know, for the first five or six years of this we
based it purely around professional values, and in the last couple of
years we’ve thrown in some aligned financial incentives back to the
physicians. I have to tell you that for our physicians, the vast
majority were right there, just on the professional values alone. I
don’t know; the financial incentives are important, but they’re not
essential for a real physician.

Now, we manage about 30,000
diabetics with this system on that registry, about 30,000 unique
individual patients with diabetes. Ten years ago we were typical for
the United States. About 35 percent of them had hemoglobin A1Cs greater
than nine, which means you’re at very significant risk for major
complications. That has fallen to under seven percent. Now the
proportion of patients with hemoglobin AC1s under seven, where you’re
really getting the patients down to levels where the diabetes is not
life threatening, they’ll have a fairly normal lifestyle, that’s
increased to almost 70 percent. I could show similar statistics for how
we’re doing on lipid controls, for blood pressure controls, for urinary
proteins, for diabetic foot exams, pedosensory exams, for diabetic
retinal exams, all the elements of good diabetes control. It came by
making it easy to do it right, and that’s how you think about it.
You’re just trying to get the elements in place that make it really
easy for your clinical teams to nail it. You see what I mean?

Matthew: Right.

Brent: I could actually give you about 30 of those, right at the moment.

Matthew:
Well, let’s not take the whole study right now, then. But I pretty much
get the picture. So you’re talking here about combining a process and a
program that in some cases people aren’t using the EMR, but you’re
making it easy for them, using the tools of data collection and data
dissemination.

Brent:
Yeah, they’ll actually want to be on the EMR because they’re just so
much more efficient. Right now, a big part of it’s paper based.

Matthew:
Right, of course. Now, there’s one piece that’s worth picking up,
something that’s getting a lot of interest in the web world (and in
general here) is the connection back going out to the patient. You’ve
basically been talking so far about activities internal to the clinical
professional but of course we know that for say a diabetic, a lot of
what’s happening that’s good, bad, or indifferent happens out with the
patient in the home and the community.

Brent: Yeah.

Matthew:
Some of that can be influenced by the physician but also some of that
has to do with the patient. I have had a couple of people talk to me
about developing software programs which either on a health plan basis
or provider basis encapsulate the entire team including the patient
and their caregivers. What’s happening at Intermountain in that regard?

Brent:
We found this really little handy little trick, I guess. We ought to
write it up as a model. It was first developed by a guy named Larry
Staker working on diabetes. We call them treatment cascades and what
you do is define a startup care that works for most people and if you
can’t get them under control with that you drop to the next level, drop
to the next level, drop to the next level coming on down. In diabetes,
just to pick up the point that you were talking about, the first step
in the treatment cascade is very thorough patient education. You want
to train them to manage their own condition.

Managing diabetes
is based around three factors: diet and exercise are the most important
followed by some sort of a medication and as you probably know so much
of diabetes is associated with obesity. For some of these patients,
maybe even many, if you can get their weight under control, diabetes
would not be an issue, you see.

What Larry did was create tools
to make the patients much more effective in managing their own
condition. He started them to chart technically their specification
charts, their blood sugar levels with home blood glucose check each
morning, finger stick check and we showed major improvements in patient
control of their own disease. We’ve done similar sorts of things with
congestive heart failure. It’s funny I tend to not emphasize the last
five because it’s always the first major step in a treatment cascade is
putting the patient in control. You can come back and buff it up. How
interestingly, the last step in that treatment cascade for diabetes is
that you refer patent to a diabetologist. That means when I talk about
30,000 diabetics in the Intermountain system, 88-90 percent of them,
somewhere along in that range are managed through ideal care by a
primary care physician in a very inexpensive setting. The crazy thing
is, is that the handful of diabetologists who back these guys up in our
system, they actually have worst medical outcome statistics than do the
primary care frontline but the reason is, is that we are getting the
right patients to them, the really tough ones to the specialists, you
see and then into that you build a very, very careful education network
at a care management first level for general education but then at a
specialty education level and your aim as part of the treatment cascade
is to get patients absolute state of the art tools.

Now
sometimes those tools are Web based. So, for example, with congestive
heart failure we try to get the patients to hop on the Web and enter
their weight, some symptoms relative to how easily they breathe, some
other factors each day on the computer so that the care team, well the
computer monitors it and flags them for people who need a call or need
some sort of intervention along the way. You build into that, of
course, full support from patient education across the Web. Those kinds
of tools, I think, have a real future. Always your first line of
defense in managing a disease is the patient themselves.

Matthew:
Right, of course, another great pioneer of this whole thing is Don
Kemper talking about patient self-management and information therapy in
that connection. I think it is clear there is a lot of opportunity to
bridge that gap but it is also clear that a lot of the requirements to
bridge the gap between patient management, patient self-management and
patients keeping to treatment protocols and the rest of that doesn’t
have to be with advanced technology, although the technology makes it
easier and makes it more possible.

Brent: Yeah, that’s right.

Matthew:
That’s a clear part to this. We have a few minutes left and we haven’t
got into the original reason we were going to talk which is the IHI and
the five million lives. We are having a little discussion about what’s
happening at Intermountain and more generally the pay-for-performance
world and the whole world around this developing and spreading
evidence-based health care and evidence-based medicine.

Clearly this is the one area that the quality debate in healthcare has actually
broken through to the public debate. You go around the country, you
hear politically and in policy forums you hear a lot of people
complaining about cost. You hear a lot of people complaining about the
promise of the uninsured and access to insurance and all that kind of
stuff. You don’t really hear too much about the industrial quality problem that healthcare clearly has, the over use and under use. It
hasn’t really resonated but since the IOM report back in 1999, there’s
clearly been a lot of emphasis on patient safety. Don Berwick is kind
of becoming more than just a health care figure now in this whole
initiative.

Give us an update of where you think IHI, the work
it’s been doing, the 100,000 Lives, and now the Five Million Lives, is?
Where we are in that whole space, and what do you think the exciting
developments around the Five Million Lives campaign are going to
produce?

Brent:
You know Don and I served together on the Institute of Medicine’s
Committee on Quality of Health Care in America, which back in ’99
published "To Err is Human." We were part of the original evidence
review. In fact, in that original publication all of the major examples
came from inside InterMountain. We have a fellow named Scott Evans,
Dave Class, and John Burke who led one of the leading research teams on
this topic for many, many years. We had pretty good background on this
one. Don and I spoke a lot about the proposed campaign. Here’s why, in
fact, many of the figures Don used in his kickoff at the National
Quality Forum last week, again, came from research that was done here.

If
you carefully look at the field every time we’ve reexamined the issue
of patient injuries we’ve been able to increase the hit rate. Another
way of saying that is the tools that were used to initially estimate
injury rates were insensitive. We’re discovering…

Matthew: It was actually worse than we thought?

Brent:
Yeah. You’ve got it exactly right. It was worse than we thought. That’s
precisely right. This is one of the slides in Don’s presentation. Back
in the Committee on Quality of Health Care in America, Gidden Evans
from the UN found about 60 million articles talking about
care-associated injuries.

Those estimates of 44,000 or 98,000
preventable deaths per year came from two studies. The first was the
Harvard Medical Practice Study, so that’s the study that Lucien Leap
and Troy Brennan led. 1984 data, hospital discharges from New York
State, they drew a valid random sample where the sample contained just
over 34,000 charts.

They used a technique called implicit review
where they asked trained nurses to read through the charts and see if
they saw any care-associated events. If they did, they’d send it out to
two independent physician reviewers who had to agree that there, first
of all, was a care-associated injury. They’d then judge whether it was
preventable. If it was preventable, whether it was negligent, and then
Troy tracked it through the legal system. All right?

Well, if
you took the New York State figures for 1984 using that implicit review
method and extrapolated them to the country as whole, 98,000
preventable deaths per year. As I recall, Harvard Medical Practice
Study had a 3.7 percent injury rate. 3.7 percent of the patients
suffered some sort of care-associated event, and about 53 percent of
those were judged by two independent physician reviewers to be
preventable.

Well, a few years passed. A guy named Eliot
Williams here in Salt Lake City, a prominent local attorney, became
quite disenchanted with his colleagues in medical malpractice. He was
troubled that most of the money in medical malpractice goes to the
attorneys, not to the patients. He put together a proposal for no-fault
malpractice insurance, speaking of tilting at windmills here.

Matthew: Right.

Brent:
He got a big Hartford Foundation grant, and we went to work on this
thing just to see if it was viable. As part of it, we replicated the
Harvard Medical Practice Study here in Utah. To build up our sample
size we included Colorado as well, hospitals from Colorado. It’s called
Utah-Colorado. Dr. Eric Thomas, part of the Harvard Medical Practice
Study research group, came out and helped us do that.

In that
study, Utah-Colorado, we found a — I’ll have to do this by memory —
2.9 percent injury rate. 58 percent were judged to be preventable. That
was the source of the 44,000 preventable deaths per year in the United
States, was Utah-Colorado. Here’s the crazy thing, Matthew. We picked
those two because they were among the most conservative in the
literature.

Matthew: Right.

Brent:
We were purposely trying to be conservative. Well, the next step that
happens, a guy named Ross Wilson, a physician at Royal Northshore
Hospital in Sydney, Australia and the head of the Quality Assurance
Department at Royal Northshore, Bernadette Harrison, Bernie Harrison, a
nurse, decide to replicate Harvard Medical Practice Study in Australia.

They
had to kind of roll their own. They didn’t get a member of the Harvard
team to come out and help, so they put together their own replication
from reading the articles. They ran it at every hospital in Australia.
I was actually down there teaching when the results came out in the
Medical Journal of Australia. 16.6 percent of all hospitalized patients
in that study had some sort of a care-associated injury. And let’s just
say it created a little bit of a furor. Politicians talking to the
press, screaming in public basically, the medical profession pretty
upset; "Is it possible that Australia’s four times worse than the
United States?" which is a hateful thought. Well, that led to some
follow-up studies.

The first thing we discovered was that in
putting together her version, Bernie had not thrown out outpatient
injuries that caused the hospitalization. She left them in. Troy &
Lucien had thrown them out of the Harvard Medical Practice study design
and out of Utah Colorado. Well, when you remove that subset of
patients, who are injured outpatients that resulted in hospitalization,
it dropped the hit rate to about 12.5 percent. Which was still three
times more than the United States.

Well, eventually, to make a
long story short, Bernie got a Fullbright Fellowship to come to the
United States for a year. She spent the first half of that year here at
Inter-Mountain. The second half with Don back in Boston at IHI. What
she did was replicate the Harvard Medical Practice study Australian
version; her study at LDS Hospital. Which has been part of,
essentially, ever major patient safety study ever done, I think.

I
was one of the physician reviewers. I can tell you from having looked
at them personally; those injuries were real. Across the board, a 10.2
percent injury rate. Which was more than four times higher than we got
using Harvard Medical Practice study. Which came as a bit of a shock.

It
turns out the difference when you look underneath the hood. In the
Harvard Medical Practice study they used a review criteria. When Bernie
tried to replicate it in Australia, rather than just sending her nurses
in and saying "look at the charts, look for events", she’d taken a look
at some of the events they’d found in the past and she’s put together a
list of 26 factors and asked those nurses to specifically check for
those 26.

It’s the difference between implicit criteria and
explicit criteria. And it bumped the hit rate by about a factor of
four, by having those explicit criteria. This is a well-established
fact, I would say, in health services research, that in general
explicit criteria outperform implicit criteria. They’re just much more
sensitive.

Well, that was interesting. Next big step, we had
Peter Norton’s study out of Canada which established right up around
eight percent for Canada, using kind of a hybrid approach. We have in
the works studies out of Great Britain and Sweden that have similar
much higher rates than Harvard Medical Practice study or than Utah
Colorado. Just more sensitive instruments probably.

Well, then
Roger Resar at Linthur Middleford Hospital, part of the Mayo
Foundations System up in Wisconsin, takes this idea and expands the
criteria list. Just with better review about the sources of patient
injury. Roger didn’t have 26 triggers for chart review; he had 55. And
he started to run it in the seven Mayo Foundation hospitals and he was
producing injury rates up around, well, 40 percent to 50 percent. He
was saying that 40 percent to 50 percent of hospitalized patients had
some sort of a care associated injury.

Now this was a kind of
roll-your-own in terms of how they implemented. And it was Brian Smith, who took it to IHI and tried to get IHI to
sponser it. That’s where it became the IHI Global Trigger Tool. It’s
the tool Don that proposed in his talk last week.

Roger’s chief
partner in doing this was a graduate of the Inter-Mountain Program:
Dave Classen, infectious disease doc from LDS Hospital, was one of our
stalwarts in patient safety. And before he left Inter-Mountain a few
years ago, he was also very heavily involved in this. Still is today.
Well, I got a call from Roger and Dave one day. They wanted to test the
IHI Global Trigger Tool rigorously at LDS Hospital feeling that so much
else has been done at LDS.

So we ran it here. I drew a valid
random sample containing 325 patients for October 2004. Adult
in-patients. So over 18 years of age, at least a one-day length of
stay. Roger brought out the same team of chart reviewers to use the IHI
Global Trigger Tool methodology on those charts. Six people. We did
inter-reliabilities, the usual thing, that it started with a review,
then went to physician review, just like Harvard Medical Practice did.
Well, long story short, over 30 percent of all patients at LDS Hospital
had some sort of a care-associated injury which is three times higher
than the Australian methodology and in the range that that tool was
showing for the Mayo Foundation Hospitals. We’ve now replicated that at
another major national center and will do a third center in February.
When those data are complete, we’re going to publish it as a major
article comparing performance of the HI Global Trigger Tool to other
major methods of picking up injuries.

This is going to turn out
to have much, much higher sensitivity. I want you to say that when you
start to get your hands on the actual number of events that are
occurring, it profoundly changes your approach. Personally, I think
it’s going to completely restructure the JCAHO approach to
patient safety for example. It changes your focus, profoundly.

Yeah,
they are not all preventable but one of the things we’ve discovered is
just finding every event… sometimes you don’t know how to prevent
them when you first see them! And then people get clever! They come up
with these good ideas.

And the next thing you know, you’ve found another category that you can
prevent. So, we think you ought to detect all of them. You see?

Another
interesting finding, almost 10 percent of all hospitalizations at LDS
Hospital were the direct result of error-situated injuries in an
outpatient setting. I mean bad enough to get you in a hospital but not
bad enough to kill you outright. 10 percent of all hospitalizations,
which well… causes you to kind of pause.

I can also tell you
Matthew, I personally reviewed those cases and they are real. This
isn’t some sort of flim-flam. I mean these are real circumstances where
patients had reactions to the therapy. Is some of it non-preventable…
yeah, I’m guessing as much as half of them, I don’t know how to prevent
them. But on the other hand, it’s sure good to know about them because
maybe we can be better tomorrow. That’s the basis for doing statistics. Don was asking for the country to step up to this problem.

The
truth of the matter is, you find what you look for and we’ve found out
in the last years much better ways of looking. You know, Marie
Bisignola talks about put on your muda glasses, in learning to look for
ways to work on the system. Muda is Japanese for waste, of course.

  Maybe we ought to keep up putting on your patient safety glasses.
Because as you learn to see these things, boy, over the place what a
massive opportunity to improve our care for our patients.

Matthew:
Yeah, it seems to me that  we are really just getting
with the combination of, if you like, cultural understanding of this
both in the profession and there’s still some resistance obviously.
But, as I think you said a couple of years back, talking at San
Francisco, you’re no longer getting people saying the data is lying.

Brent: Yeah.

Matthew:
And people are now accepting this. You’ve both got the cultural
understanding, you’re getting the payers, now starting to focus on this, the
government… the consumers are beginning to get involved in this as
well. And then obviously the issue that, you can now start to use these
electronic medical records and other different tools to be developed to
start measuring this stuff. And there is clearly enormous, enormous
opportunity for massive quality improvement and increased patient
safety across the board.

I feel for Michael Millenson and you and others who have been shouting
loudly about this for many, many years. But it seems to me that we’re
at a point now where we can start to really get at this issue and, of
course it’s going to have a by-product that it should actually, just
like the muda thing, make the "production of health care’, if you want to
call it that, more efficient. And theoretically lower the
cost as well, because as we all know re-work is the most expensive
thing you can do.

I think there’s tremendous, tremendous opportunity here, and the work
that you and your colleagues have done on this, over the years in the
trenches, has been quite wonderful and is just a phenomenal effort. And
again, I said there are issues in spreading this across the country,
and I think the work Don Berwick has done is great, but we’re really at
the beginning, not the end.

Brent: Yeah, you’re dead on track
Matthew. And Don is very sensitive to the rigor behind this one. I
think in the end he attracted some criticism with a hundred thousand
lives from the pure scientists I guess. And he’s trying to structure
this one in terms of formal study tied into helping with that one to
make sure that we have scientifically defensible results this time. And
I think we will.

You know, just to close it up, this stuff almost killed my father,
just to make it real. Back in 1998, when we diagnosed Dad’s congestive
heart failure, he showed up one day, obviously having trouble
oxygenating, a lot of fluid retention, pretty bad edema. We hauled him
into the emergency department at LDS Hospital. On physical exam, you
could hear the water in his lungs. His O2 stats were down below 80, not
oxygenating worth a darn. Well, we started him on some oxygen,
diagnosed the condition, started him on some monitol to get his heart
pumping better. The main therapy of course is you blow up that extra
fluid, which means diuretic. We started him that night on good old
Lasix—the old standard. It had been around forever. It was first coming
out way back when I was first in clinical training just to show how old
it is.Well we put him in hospital for the night because he was
having such trouble oxygenating, just to be safe. The next morning, on
my way to work, I swung by the hospital about seven a.m. to see how he
was doing, thought he would probably be heading home that day. Got up to his
room and he was doing great on a cardiac standpoint. He lost a lot of
fluid. His O2 stats were coming way up. He was breathing a lot easier.
Trouble was that overnight he developed upper left quadrant belly pain,
pretty bad pain. His blood chemistries, his amylase had spiked to 1500.
That means that we were looking at a case of acute pancreatitis. I
remember sitting in his room that morning talking to his attending
physician. We were commiserating about geez, how often does it happen.
You just get on top of the primary disease and out of the clear blue
sky, you get whacked by this stuff just coming out of the clear blue,
acute pancreatitis.We weren’t quite sure what to do. I remember
leaving his room, walking out on the hospital floor very concerned.
This could have killed him. These things are life threatening. When who
should I see coming down the hall the other way. It was a fellow named
Stan Pestonik. Stan was a pharmacist who worked with our adverse drug
effects team, some of the original major researchers on patient safety.
He now runs TheraDoc. Some of you may be familiar with that. Well Stan
is coming the other way with a sheet of paper in his hand and I say,
"Hi" and he says "Hi." He said, "Hey I have a review of a patient named
James." He said, "Is that a relative of yours?" I said, "Yeah my dad."
He said, "Oh, we just picked up his furosemide, his Lasix allergy." I went
"Ha?"

I said, "Stan, Lasix is as safe was tap water. Lasix doesn’t cause
allergies." He said "Oh yeah it does, (he’s a pharmacist). In a very small
proportion of patients it causes a life threatening pancreatitis." Well at
LDS Hospital, Scott Evans, Stan, Dave Classman built a trigger system
where you watched lab values. antidote drugs,
nursing notes, watch you for adverse drug effects for their treatment and it
would produce a list of patients. Stan would come check on them. It took a
third of his time, half of his time to check everybody. Only about one
in five, one in four of the patients he checked would turn out to
actually have something going on.But let me just put it this
way. A few minutes later, my Dad was off Lasix and on
hypochlorothiazide which continued his diuresis just fine. That afternoon
when I checked, his pain was going away. His amylase was dropping. It
probably saved his life. Now frankly, it is not preventable is it? If I
had known—I didn’t, but if I had known—that Lasix caused this problem,
given the incidence rate was so low, I would have used it anyway as
most physicians would have. The difference is at LDS hospital we had a
system that picked up a patient injury while it was still developing so
we could intervene before moderate ADE became a severe ADE, and before a severe ADE became death. See that idea, right there.

Matthew:
Absolutely. That’s the difference between the medical error, the
preventable medical error and the actual catching of it in the system.
I think that’s a great luckily not apocryphal story about your father. So that’s a
fantastic way to end.

Well, Brent, thank you so much for your
time. I have been talking with Dr. Brent James who is the VP of Medical
Research at Intermountain Healthcare, and obviously also on the IOM
committee, working with IHI, a real pioneer in patient safety and
clinical performance improvement in the US and it has been a great
pleasure talking to you Brent. Thank you for your time.

Brent: Thanks. I really appreciate it.

Livongo’s Post Ad Banner 728*90

Categories: Uncategorized

Tagged as: ,

7
Leave a Reply

7 Comment threads
0 Thread replies
0 Followers
 
Most reacted comment
Hottest comment thread
7 Comment authors
abassTDPoor House DocEmily DeVotoJane Hiebert-White Recent comment authors
newest oldest most voted
abass
Guest
abass

Often we forget the little guy, the SMB, in our discussions of the comings and goings of the Internet marketing industry. Sure there are times like this when a report surfaces talking about their issues and concerns but, for the most part, we like to talk about big brands and how they do the Internet marketing thing well or not so well.
http://www.onlineuniversalwork.com

TD
Guest

Amazing interview. Very glad to have found it, and am somewhat pissed I didn’t find this blog a year ago.
Fascinating insights into the world of US health care. Cheers.

Poor House Doc
Guest

So do I have the core concept of implementing a shared baseline in HIT correct? You structure your documentation/tracking system to capture best practices monitoring and treatments for a category of patient. If subsequent outcomes confirm expectations you continue these protocols. But if you see negative deviations from desired then you have (or you build) additonal protocols for this now re-categorized patient (needing addition protocols to achieved desired responses).

Emily DeVoto
Guest

Matthew, thanks so much for interviewing Brent James. You asked great questions, and this is great food for thought on a number of levels. Next, can I interview you for my blog? 🙂

Jane Hiebert-White
Guest

Matthew, interesting interview. Thanks for pointing to the work of Elliott Fisher and Jeff Goldsmith in Health Affairs. Both are continuing the discussion of hospitals and community physicians as “extended medical staff” as a means to control costs.

heromd
Guest

Wow. This was a really wonderful read. I echo R. Oakley request for a better explanation of a “shared baseline.” When will health policy stop interesting me?

R. Oakley
Guest
R. Oakley

Could you please explain better the concept of a “shared baseline?” What does that entail, and what is its function? I can’t find a decent definition of the term.
Thanks. This is a fascinating discussion.