[This Transcript is Unedited]




June 24, 2004

National Center for Health Statistics
3311 Toledo Road
Hyattsville , MD

Proceedings by:
CASET Associates, Ltd.
10201 Lee Highway, suite 160
Fairfax , Virginia 22030
(703) 352-0091


P R O C E E D I N G S [9:05 a.m.]

Agenda Item: Call to Order and Introductions – Review Agenda/Intent of Hearing – Mr. Hungate

MR. HUNGATE: Okay, are we all set? Are the speakers working, the sound
system working, etc., all tuned up and ready to go?

I want to start with the mundane things of thanking NCHS for the coffee and
donuts, etc., this is an unaccustomed pleasure —

MS. GREENBERG: The NCVHS team provided —

MR. HUNGATE: Specific? Alright, well we’ll have to advertise that around a
little bit —

MS. GREENBERG: I wouldn’t.


MR. HUNGATE: Okay, welcome all to the hearing that I have to decided to
label as Lurching Toward Measurement. As I look through the record of previous
hearings that led up to this, which started really I think back in 1998 roughly
on quality issues, and one of the first comments that I came across was Simon
Cohn’s comment about we manage what we measure. You really said that, I have
the documentary proof with me.

Well, let’s start with introductions and start over here at my right with
Julia Holmes.

DR. HOLMES: I’m Julia Holmes and I am, I work here at NCHS and staff on the
Quality Workgroup. And my primary task at NCHS and involvement in quality
measurement and disparities measurement have been that I’ve been kind of the
lead person working on the National Health for Quality Report and National
Health Care Disparities Report led by AHRQ, it is an intergovernmental effort
and NCHS and CDC provided a great many of the measures that go into those

MS. GREENBERG: I’m Marjorie Greenberg also here at NCHS, CDC, and I’m the
executive secretary for the committee.

MS. POKER: Hi everybody, my name is Anna Poker and I’m from AHRQ, I’m going
to be the staff lead to the Quality Subcommittee. And my work at AHRQ, I’m
relatively new at AHRQ, I’m working on the National Heath Care Quality Reports,
Disparities and Quality, and I’m also on the patient safety team there.

MR. HUNGATE: Bob Hungate, principal of Physician Patient Partnerships for
Health, which is just me, I’m a Medicare retiree, I also chair the Group
Insurance Commission in Massachusetts. And chair the workgroup, member of

MS. GREENBERG: Bob, we should probably ask if anyone has any conflicts.

MR. HUNGATE: I’m reminded that I should ask that members that have
conflicts should announce them. In that sense as a beneficiary of the Medicare
am I conflicted?

MS. HANDRICH: My name is Peggy Handrich, I am a member of the committee, I
work for the Wisconsin Department of Health and Family Services with
responsibility for administering the Medicaid program and vital statistics
parts of the state.

DR. CARR: I’m Dr. Justine Carr, I am at Beth Israel Deaconess Medical
Center in Boston and I am a member of the committee and member of the Quality

MR. REYNOLDS: Harry Reynolds, Blue Cross and Blue Shield of North Carolina,
member of the Standards Subcommittee sitting in on the Quality Committee today,
and no conflicts.

DR. COHN: Simon Cohn, I am the national director for health information
policy for Kaiser Permanente. Like Harry we’re sort of sitting in as sort of
the implementation arm of everybody’s good ideas and make sure there may be
some way at the end of the day to implement whatever it is that’s decided. I’ve
been told that I need to bite my lip or tongue if CPT comes up since I’m on the
editorial panel.

MS. DILORENZO: My name is Jessica DiLorenzo, I work for General Electric —
really just one of the Bridges to Excellence Products so I’m also the
operations leader.

MR. QUERAM: My name is Chris Queram, I’m the CEO of the Employer Health Car
Alliance Cooperative, a business coalition in Madison, Wisconsin, fellow badger
along with Peggy, and I’m here today on behalf of the Consumer/ Purchaser
Disclosure Group which is an organization and an initiative that I’ll tell you
more about in my comments.

MS. TITLOW: I’m Karen Titlow, I’m the director of operations for the
Leapfrog Group, and my only conflict of interest is that sometimes I need to
get health care —

DR. HAYWOOD: Good morning, I’m Trent Haywood from the Centers for Medicare
and Medicaid Services and I’m all also on staff to the Workgroup on Quality.

MR. HUNGATE: Can we include the audience in the introductions?

MS. COFFEY: I’m Rosanne Coffey, I’m from —

MS. BLOUNT(?): I’m Laura Blount, I’m the associate director of federal
relations —

MS. COLTIN: I’m Cathy Coltin, I’m director of —

MR. HUNGATE: And for everyone’s edification who doesn’t already know it
Cathy was my predecessor as the chair of this workgroup and labored long and
hard in collecting the information.

Much of what we’re trying to do today is if you will improve the
institutional memory and specificity of our understanding of the business case
for increased measurement in some fashion proposed in one way as claims data,
proposed as an alternative maybe as a new quality transaction which would open
a different set of doors so the specifics are not locked down in that sense.
Each of you received a list of if you will questions that relate to trying to
move, and I don’t know whether these two pages are in the packet that everyone
got or not. The speaker questions? The two pages that went out, is that there,

The questions themselves do not take into consideration one other aspect
that’s been talked about within the Quality Workgroup in any very specific way,
this was the question of what, for various measurements, what’s the best
approach vis-à-vis what we can do now and what we should not try to do
now because it will be so much easier once there is electronic health record.
What’s the mix in terms of where we should spend our efforts now versus waiting
for better information to an EHR. So that’s not in the specific questions but
it’s of interest and you should consider it as you speak.

Format wise what I’d like to do is try to keep remarks to 15 minutes, and
just go sequentially with no questions if we can do that because then we can
all as a group talk more collectively afterwards. It’s been my experience that
sometimes one speaker takes a little long and we give up on the discussion
which I think is a lot of where our benefit will be.

The 11:15 to 12 noon time is spoken of as workgroup discussion, I don’t
want to have that suggest that you should leave, I think we’d welcome your
staying for that and participating because I think the reason you’re here is
you’ve got knowledge and will help us. And I would like to just take a couple
of minutes and let other people from the quality workgroup, both members and
staff, add any thoughts that they have in relationship to this hearing and
especially ask Simon and Harry to add thoughts that relate to your
perspectives. The pressures of time have meant that we haven’t had the
collective planning process that we might have liked to have had in preparation
for this so if we can take five minutes and allow for some dialogue there or
issues that aren’t covered that should be I’d welcome it. If no one has
anything to add we’ll just go right on but if there is something I’d like to
catch it now.

MS. HANDRICH: I don’t want to wait, I’ve been looking forward to this.

MR. HUNGATE: Okay, well hearing no other suggestions I suggest we go
forward then.

I want to thank the panel for coming, we very much appreciate your taking
the time to work in our edification and the public’s edification. The other
comment that I would like to make, personal opinion, that one of our
measurement problems is that we haven’t paid enough attention to the importance
of the quality improvement model whereby the best data is taken by people who
take it for their own reasons, who are taking data for their own self
improvement in some way. When you’re taking it for somebody else it doesn’t
have near the same impact on your accuracy or correctness of reporting. That
factor and the need to align the measurement system for payers, purchasers, and
patients as I would describe it I think is an important consideration in
improving quality. So that’s just a generality that I think is important.

With that we start out with Trent Haywood who I had the pleasure of hearing
present to all of Medicare people about the new, public hearing relating to
their initiatives, and I learned a lot more. So Trent, welcome, I look forward
to your comments.

Agenda Item: Purchasers – Panel 1 – Dr. Haywood

DR. HAYWOOD: Good morning, I’m sort of preaching to the choir since I’m a
member of the staff here, nevertheless I will start out with some general
comments. I’m going to be a little less formal then some of the other speakers
for now and if that’s okay I’ll take that prerogative and not stand at the
podium unless I’m demanded to. What you should have is kind of the handout
anyway so we can just move along according to the Power Point handout.

Let me just start out and you’ll see on that first page I have some general
comments that I think we already have or at least allude to in the
recommendations or some of the preamble to the recommendations but I still want
to highlight it in the backdrop of what we’re going to discuss today which are
primarily looking at administrative data to get at some of the quality
information that we’re trying to achieve. And so what I put at the very
beginning, overall in looking at the recommendations I think it’s clearly
stated that CMS is supportive of both the intent and the approach of trying to
actually determine what is the best methodology that we can strive to achieve
as far as doing well now recognizing the limitations that we currently have.

And so particularly what I stated is that we should keep in mind that
administrative data is really more of a bridge to where we want to go then
something we think is kind of the end all be all to some of the problems that
we currently have as far as data gaps. And in other words I wrote here that
there continues to be a constant tension between building upon a foundation
that was never intended to be used in that manner, meaning the administrative
data, versus allocating energy and resources towards quicker adoption of the
idealized design of EHRs.

And I think that because continually as we work, I apologize because
there’s a double page here so skip that second page and go to the third, I
realize that as we continue to work kind of with all the partnerships and
providers that there continues to be that tension where if you talk to
individual clinicians and practitioners they have immediately have some
concerns when you mention that you’re going to do quality measurement from an
administrative source. And particularly as I talk a little bit more about our
physician office quality, once we start moving toward to the ambulatory setting
why physicians in particular have some concerns about that administrative
source, the intent of it, it starts to get into questions of whether or not
you’re really helping me to become a better coder or really helping me become a
better practitioner. And so despite our best efforts I think whatever we end up
concluding with the administrative data you’re still going to have to deal with
that message and the issue around the source of that particular data,
particularly I think for physician offices.

I just quickly highlighted here, I think some of you already know,
particularly my fellow colleagues, that the President, this is how the e-Health
initiative that was announced in April and has the ten year goal of actually
trying to get us to the EHR that the NCVHS has been working quite hard on and
the first prong was that adoption of health information standards that I know
NHII and CHI have been working closely kind of under the umbrella of the NCVHS
to actually move that adoption forward and continue to move forward on that.

The second one listed was kind of increasing funding for demonstration that
are IT related and a lot of that is starting to occur actually at CMS where
we’re starting to have more demonstration and I’ll talk more fully later about
some of those demonstrations where we are starting to incorporate or
incentivize physician offices using information technology. And we kind of
modeled that after what the private sector is already doing so when you hear
some of the discussion around Bridges to Excellence and the physician office
link as well, the CMS activity is primarily based upon the model that was
already out in the private sector.

Then additionally what the President has also asked all the federal
agencies, so the VA, Department of Defense, Medicare is kind of a set large
purchasers but across the federal agency is look at our kind of programs and
policies to see what we may actually be able to alter those that would actually
incentivize adoption of the information technology systems. So we’re starting
to actually do that as well, looking at areas in which we can actually
incentivize adoption, whether that would be through payment or whether that
might actually be through the way that we actually design some of our policies
or regulations.

And then lastly I think all of you kind of know that Dr. David Brailer was
announced as the new National Health Information Technology Coordinator to
actually move us towards that goal in ten years of actually having not only
kind of EHRs in the sense that we may be talking about in physician’s office or
a hospital system but also kind of the personal health record so that as
someone moves across the community that we would actually still have
information flow across that community as well for that individuals. So all
that is occurring, that’s the backdrop in which we’re going to have this
discussion, and the last thing I did want to highlight as far as a general
comment is the tension that always exists around pace, kind of timing and the
pace of activities, because I think one of the things that you had alluded to,
Bob, earlier on is kind of what do we want to actually achieve now versus
waiting for EHR recognizing the current limitations, I think that continues to
be an activity, a tension that we also face throughout all of our different
quality initiatives that you heard me talk about is kind of the pace both from
the side of purchasers and consumers in comparison to maybe where providers
currently are at this particular time or maybe even some of the scientists that
may argue for a little more rigorous methodology and wait for that particular
process or method.

So having said that let me just quickly before I get into the specifics
because I did try to do my homework according to the way you set it out, Bob,
and specifically address those questions but I wanted to give some of the
members that may not be familiar with some of the activities that we have
undertaken at CMS what we are currently doing and that will help you understand
kind of where we are and hwy I’ve come up with certain specific

In 2001 Secretary Thompson announced what we are calling our Quality
Initiatives that have two primary foci which are one, empowering consumers to
make informed health care decisions, and simultaneously to stimulate and
support providers and clinicians to improve the quality of health care. And so
if you look at that middle diagram that’s entitled what we can do to improve
quality, this is a schematic of a process in which we’ve actually tried to
address those two goals. And in particular if you look, once we are working
with our partners to actually get a sense around selecting the priorities as to
where you actually want to move then the two underneath it become kind of the
focus for this discussion which is the measures, which measures, how to adopt
those particular measures, and whether or not you can actually feasibly collect
that data in the way that you actually intend to collect it to actually show
you what you wanted to actually show.

And so that is the methodology and then at the bottom of that what you see
is all the levers if you will, all the levers in which we can actually
influence that process at CMS whether it be through a regulatory method,
whether it be through technical assistance, whether it be providing that
information to consumers, structured payment, and then kind of rewarding that
desired outcome, so pay for performance which is some of the conversation

The bottom slide just highlights the fact that within this particular
administration with Secretary Thompson one of the biggest goals was to really
get more information out to the public at large. And so we continue to do that
and that is really remains kind of the central focus for the quality
initiatives to actually get that information out to consumers.

So on the next slide you see kind of all the different Medicare.gov, which
is our consumer website, the information that is currently available to
consumers, so for our Medicare+Choice plans we have quality measures at the
health plan level, for our dialysis facilities compare, we have that
information as well looking at adequacy of anemia, looking at survival rate,
and looking at anemia management, adequacy of dialysis and anemia management.
In the nursing home compare we have that information, in home health compare we
added in 2003, and then 2005 we’re actually going to transition into our
hospital compare and also have that information available to consumers as well.

So we started off with, when the former administrator Skoal(?) was here we
actually started off with nursing homes and one of the reasons why we started
off with that was because there was the infrastructure already in place in the
sense that there was a common dataset, the MDS dataset that people were
familiar with, whether they agreed or disagreed with the use of that dataset
for quality measurement at least there was some familiarity with the particular
dataset and everyone actually was utilizing in that dataset.

Now we went ahead and looked at that dataset and were able to provide kind
of valid and reliable measures that we actually went out and pilot tested
before we actually implemented on a nation wide basis. So we’re quite
comfortable with the validity and reliability of those measures.

Having said that though we still recognize that not only internally but
even externally stakeholders have concerns about continuing to rely on that
particular dataset, an administrative dataset that really wasn’t designed for
the purposes of actually quality reporting. And so MDS 3.0 is out there where
we’re actually right now apparently talking about how to actually improve the
MDS 3.0 in the revision to MDS 3.0 essentially because there continues to be
even tension among what we actually want to do with MDS 3.0 meaning externally
as well as internally in the sense of do you want to go down the path of
actually the most intuitive, clinically intuitive system that helps kind of
primarily the nurse that’s taking care of that resident and have that type of
system, do you want a system that actually allows researchers to actually be
able to get the data that they need to do to actually assess, do you want to be
kind of a trigger actually with the rest to be kind of a trigger mechanism so
that if you actually follow the actual flow of that actual instrument that it
would trigger for people that may be at risk for anemia, at risk for falls,
etc., do you want it to be that way versus actually you identifying the
diagnosis first and then that you’re going to route.

So there continues to be discussion around even with that MDS 3.0 what the
actual, what the actual primary focus or goal, whether compromises will be
reached to actually achieve all these that everyone wants to place on something
that wasn’t originally designed to do all of this. So that’s just one example
where I think we’ll continue to have discussions around whether tradeoffs will
actually be as far as utilizing that administrative dataset recognizing that I
think, that we’re all recognizing that if we have the EHR pretty much the way
that we would want it that would remedy a lot of these issues.

Similarly with the home health and the oasis, we’ve had similar concerns as
with the MDS is that again you’re relying on a system that was necessarily
designed to do what you actually are required to do in today’s marketplace.
It’s good now as far as kind of where we are currently but there’s a question
as to whether or not you want to continue to go down that particular path. One
of visions at least has been for both of these initiative improvements, all the
initiatives that we’ve done, is that we do work with the National Quality Forum
as our consensus building process so that we’re able to take measurements
regardless of whether it was administratively derived or not, take that through
the consensus building process and make certain that external stakeholders have
input as to whether they’re in agreement or disagreement with that particular
measurement and report it out as a quality metric.

Hospitals now, when we move to hospitals, I’m on the top of the next page
for quality initiatives of hospitals, when we move to hospitals you quickly
move away from kind of a common data infrastructure in which everyone was using
the same data and you move to a place in which you have different vendors for
individual institutions that aren’t necessarily reporting data particularly at
the micro specification level the same way that you may be actually requesting
that data. There’s different I think requests or demands for data at different
levels for the individual institution. And so when we move to the hospitals we
quickly recognized that we don’t have the infrastructure that we had on home
health and nursing homes and so we had to start early on and take a step back
and actually work with some of the stakeholders here and some that aren’t
present to get a sense of where we actually could start off, what could be kind
of a common place to start.

So that’s why we started at looking at kind of the National Quality Forum,
looking at what the Joint Commission was already requiring through some of
their vendors as well. And I know Jerod Loeb is on the schedule later today to
talk but we’re looking at kind of what they were already doing as well so that
we wouldn’t actually increase the burden so that we could take what was already
out there and actually build upon that. So we came up with the ten measures for
three clinical conditions that’s on our professional website that we’ll
transition in February to the consumer website around heart failure, AMI, and
pneumonia. Now we’re starting a phase now where we’re going to actually build
out beyond those three clinical conditions and we’re going to build out to
surgical infection prevention.

On the hospital side most of that is derived, it’s charts abstracted pretty
much, that’s chart abstracted data so you’re moving to really what people are
saying, administrative data is not actually, it doesn’t hold the same amount of
quality of data that you’ve done on nursing home and home health and so you
really get into chart abstracted data and so course that’s resource intensive
and there’s costs that are incurred when you’re doing that. So again, I think
we’re all cognizant of that so we’re trying to move forward in a reasonable
pace recognizing the demands that are out there on the individual providers but
also recognizing the need not only for purchasers but more particularly for
consumers to have that information so that they’re informed about their
particular chart at any of those institutions. So we started out, and so the
next clinical condition will be surgical infection prevention, providers have
been on board with that one as kind of really the true first patient safety one
that we’ll have as a module or a domain.

I did go ahead and also highlight because there was some interest in kind
of pay for performance as part of the backdrop and the questions that we’ll get
into more specifically. I did mention firstly a financial incentive is not
really a pay for performance but a statutory incentive which was the Medicare
Modernization Act section 501 did have a differential market basket update for
those hospitals that would submit to ten measures, so at least they’re a little
more incentivized if they weren’t already doing it voluntarily to actually
provide that information to consumers.

Then at the bottom you’ll see a few slides that talk about the Premier
Hospital Quality Incentive Demonstration and for most of you I think you
already know, when CMS is talking about demonstration we’re talking about
payment demonstration where we’re actually somehow trying to incentivize, or at
least test the marketplace through financial incentives, and so this one links
it to quality and so what you’ll see on the next page, for those providers who
actually score in the top two deciles we’ll have financial incentives linked to
those two percent, two and one percent respectively for those two deciles.

We start to actually put that information on the website, now in the
Premier demo gives some, provides backdrop and instruction on how difficult
some of this activity can be in that this was designed not to test the
financial incentives, not necessarily to tell us what we’re going to go over
all because we take that kind of unsolicited proposal because the Premier
demonstration, Premier was able to come to us and tell us that they could
actually provide measurement for 34 measures so they’re ready to provide this
information on a larger set of measures across, because they have the online
database built in to their particular system so they provide that information.

If we wanted to actually try to model that beyond them as some have come in
and talked to us about doing we quickly recognized that many institutions
aren’t necessarily set up to do it, even internally within CMS we have to build
out our enterprise system to actually accommodate all that activity because
this is just one small demonstration but it wasn’t designed to be built on a
national level. We started to actually build out our own infrastructure to
support that activity as well so we think by next year we can accommodate some
of that and so when I talk about the QIO data warehouse that’s what I’m talking
about, that we need to build out our own enterprise. So in other words if all
the hospitals right now just wanted to submit kind of 34 measures across the
board we wouldn’t be able to actually accommodate that today, we’d be able to
accommodate that next summer but we wouldn’t be able to accommodate that today.
So that’s one thing in the back of your mind and also to keep in your mind that
even if you got what you wished for you also have to make certain it links with
kind of the infrastructure that would actually support it.

And then lastly before I get into kind of the specifics that you asked me
to I just wanted to quickly highlight, one thing, I’m going to skip
physician/physician quality and talk about it later, is ESRD quality
initiative, what the ESRD they already were providing kind of live information
so we have for instance adequacy of dialysis and anemia management where on the
actual administrative claim form they’re providing that information to us. But
they’re financially incentivized to actually provide that information to us and
so we haven’t had much concern at getting that information, we’ve actually been
able to get that information where at least it’s kind of constructed unless you
have that financial incentive, that may be viewed as kind of an additional step
and an additional burden without much return for the individual institutions.

For instance on ESRD one of the things that we’re going to have is a
demonstration and when as you see on that slide that in addition to the quality
measures we already have one of the measurements that we’re talking about out
is bone measurement activity and that takes kind of calcium and phosphorous
intact PTH, which is parathyroid hormone, kind of looking at bone metabolism.
And we quickly realized that actually the administrative data is not that great
for actually once you start trying to get at that particular level and try to
come up with bone metabolism. So we’re thinking that we may actually end up
being forced to kind of go to the chart level to actually get that information
and not go with administrative data or try to bill out that administrative

Now I’m going to skip these slides, if you see the bottom of I think the
next page I actually start off considerations for NCVHS where I kind of just
walk through the specific recommendations. I lumped recommendation number one
and two together for purposes of discussion, which is recommendation number I
believe was live test and recommendation, number two was vital signs. So what I
was reading here is the need to collect selected live results and vitals are
important as they would improve the ability for us to actually evaluate an
immediate outcome. So we have no doubt that we think this information is
definitely important and would be a priority for us. And then I said for
example hypertension control, if we really wanted to be able to get a sense of
how effective someone is actually taking care of a hypertensive patient we
would actually need to be able to get at this level so we definitely think
that’s valuable. And we think it would be valuable for us is to have the lab
results similarly if we want to get control, for instance diabetes control,
things of that nature, LDL control, all that is actually valuable and will be
important for pay for performance as well if we really want to pay for

I did not that I don’t know if we can say all vital signs are equal, I
think hypertension control is like example where I think people would recognize
that. If you start talking about fever or pulse or respirations, particularly
respirations, I don’t know what that gets you, to be honest working in the ER I
never, I discount respiration unless I do them myself so I never, pretty much I
ignore respiration. And oftentimes I ignore fever unless I’ve gotten it twice.
So I think all vitals aren’t necessarily equal but I think they’re important.

Until fully implemented this is probably the best that we’re going to do is
actually as you outlined try to actually get it through some type of
administrative process so I think we’re in full support on that.

Recommendation number three, secondary diagnoses, this does continue to be
an issue that we’re looking, we do agree with you, we think that it is right
now that the best methodology is going again to be some type of administrative
fix for this. And so we’re in full support I think again with that, and we
recognize both from the standpoint of actually trying to do risk adjustment but
also trying to look at complication rates if you’re going to actually try to
look at some of these, complications then it’s important to have that

Recommendation number five, and this is, sorry, recommendation number four,
I skipped one, recommendation number four is I believe this is the operating
physician for the principle of inpatient diagnosis. In talk about this
internally we agree that this is probably an appropriate mode if one is only
interested in that principle inpatient procedure. In other words if the
individual patient depending upon the complication and what procedure they had
that may or may not be of your interest, I can imagine like a lot of patients,
particularly any patient that went to ICU, they’ve gotten a tracheotomy, they
have multiple different procedures and so it really would depend upon what your
point of interest is as to whether or not this is the information that you’re
going to want.

So I think that was the only limitation we had with that particular
recommendation. I think there will be addition burden but I don’t think that
additional burden for the purpose of this would necessarily outweigh it, the
need to have that or the value as the questions had raised it to us.

I’m going to just quickly run through recommendation number five, this
would also again be important I think for part assessment and pay for
performance. There’s not a data element there as to risk adjustment, I think
that was one of the specific questions that was asked was about risk
adjustments for this. There is a concern whether this information is consistent
with kind of what you really want, in other words I’m thinking about timing of
antibiotics or thrombolysis in the ER, whether or not I’ve actually admitted
that patient in the ER before I’ve given thrombolytics, I’m not quite certain
that I should get you the information that you want. So I think there’s some
clinical issues that may not necessarily get you the exact data that you were
actually wanting, this is actually giving you the information that you want.

And then if this is one of your recommendations I think this one may
actually be considered a little burdensome although I know that we purposefully
kind of selected, or selected conditions so that we can actually reduce that
burden and I gave examples of the surgical infection prevention measures or AMI
measures. But then overall if we can actually these, at least I’m presuming
that this is going to be less costly then chart abstraction so it’s just a
matter of if people are actually demanding that information and providers truly
have to provide it then this would be a better route then if they have to
continue to rely on chart abstraction. If there’s not a demand for this
information and we’re forcing them to do it then I think that becomes more of a
burden for them.

And then real quickly, kind of get through the last two recommendations
here. Recommendation number six, in reading this one I think this is definitely
helping us get in the kind of the global procedure codes and somewhat
disaggregate those if you will. This is important definitely actually for
coordination of care and kind of the continuum of care for the individual
patient. I’m not certain as far as the utility for pay for performance, I’d
like to hear from some of the other panelists on that. When I read that I
didn’t get a sense that’s something that we may leverage for pay for
performance to be honest and I didn’t really get a quick sense of any risk
adjustment purposes for that.

And then lastly, that was one of the ones that I think that was more
directed to CMS so I did talk to one person in CMS, I didn’t truly vet this
with Center for Medicare Management that does the payment side of the house but
I did vet it with one of the managers over there and he found that this would
be actually problematic, it would just be problems from the reality of actually
trying to get physicians to go back, disaggregate, bill for something that we
know that we’re not going to pay them any different for because it’s wrapped in
part of the global billing process. So he just thought from his experience this
would probably be kind of problematic to be able to actually take a billing
approach to this.

And then functional status, Karen, I’d be interested to see what you
thought, so if I say wrong I defer to Karen. I did talk to some of my team
about this, in particular some of my team that’s in long term care and had a
lot of experience with functional status. Overall the general sense was that
there was definitely a positive and eager to actually get functional status and
get to his lab, the concern is that there’s not standardization around it and
that continues to be a primary barrier to actually having that functional
status, which is what we’re, what the recommendations I think of trying to
achieve is actually to give that type of standardization.

And so we’re in support of actually trying to explore that possibility, get
a sense of what that would look like so whether that would be ICF or some kind
of ICF with some other lexicon to help us actually be able to get at that
standardization so I think that’s actually where we would come on functional
status that if we could actually get at the standardization around functional
status then that would actually be beneficial across the board in actually
being able to really look at true kind of improvements of care and actually
kind of improvements, to the highest level possible for the independency.

So with that I will stop and look forward to the discussion.

Agenda Item: Purchasers – Panel 1 – Ms. Titlow

MS. TITLOW: Just a little bit about us so that everyone knows who the
Leapfrog Group is, we are a consortium of more then 155 health care purchasers
who purchase on behalf of more then 34 million American. We spend more then $62
billion dollars on health care expenditures every year on employees. It’s also
important that we are a group that’s trying to work as Trent alluded to, it’s
extremely important for us as an organization to work very closely with other
organizations that are trying to standardize measures and to standardize pay
for performance, so that would be CMS, DOD, and OPM are liaison members. And we
want to make this point strongly to say that we think that what you’re trying
to do here is critical so that all of our efforts will have the maximum ability
to move forward quickly. Without this level of data, it has just been a daily
struggle for us to be able to implement what we think is the most important
ways to improve health care in the country without this data, it is the minutia
that we are all being stuck in, all of these groups together. We’re trying to
standardize our message to the providers and to the health plans about what
we’re looking for and we absolutely need help in standardizing the data
collection process. So that’s our global, that’s the biggest statement that I
can make for today is that this is absolutely critical for all of our ability
to succeed.

The Leapfrog Group’s mission, our overall mission, we are mostly known as
being a patient safety organization, that is how we started, but we started
with patient safety because that was one place where there was more data to be
able, that we were able to collect. It was also something that consumers could
pay the most amount of attention to. But our larger goal, and this is how we
are now recasting our goals out to the public, is that our overall goal is to
trigger giant leaps forward in the safety, quality and affordability of overall
health care by supporting informed health care decisions by those who use and
pay for health care and promoting high value health care through incentives and
rewards. And what’s really critical in this mission statement for this group to
see is the word informed health care decisions which right now almost all
health care decisions made by consumers, made by purchasers, are made by
guessing. So we are using financial data almost exclusively to make decisions
and we don’t use that when we buy any other product. If for example I’m buying
car doors and, car door hinges, and the hinges fall off all the time, even if
they were really inexpensive I would be less likely to buy that product. But we
can’t make those same kinds of judgments with health care and the data that we
need to collect that you’re talking about here, the expansion and ability to
get to this information easily, really is going to make the difference for us,
to be able to make informed decisions both on the purchaser side and for the
individual consumer which is half of our mission, or half of our goal is to
really get out to the consumer, it’s important for them to have this data.

What we’re really promoting in order to be able to achieve this mission is
a new purchasing model. We need two strong pillars of this new purchasing
model, the first is transparency, and transparence as I alluded to earlier,
it’s a major goal of Leapfrog, it’s an even stronger goal of the disclosure
group who you’ll hear from in just a moment, which is to have the broadest
possible access to standardized measures and that everybody has access to look
at those measures and be able to compare between individual providers. So we in
order to do that we need easy access to standardized quality health care data
and I have to tell you every time that I work on our survey, we have a survey
that hospitals report to, the most amount of push back that we get from the
hospitals is that everybody and their cousin is asking for something slightly
different and that is where the burden is, it’s the variation in what is being
asked for that they are most resistant to. They are much less resistant now to
the fact of just being transparent, it’s the administrative burden, and so I
think that the standardization that could come about by the efforts that you
are trying to push forward here could make a huge amount of different in
allowing the hospitals, encouraging the hospitals to report what in fact they
are now more willing to report then they’ve ever been willing to report before.

We also believe fundamentally it’s a patient’s right to know about the
quality of care that they can get and this data as I mentioned is a very strong
piece of what they need to be able to make that informed decision making. From
the purchaser’s standpoint we have been buying based primarily on the cheapest
possible deal that we could get and that is an ineffective way to try to
improve the quality of health care. And all the members of the Leapfrog Group
have formally committed to buy based on the highest quality of care rather then
based on the cheapest care that we can get. And in order to do that we need to
a, know what high quality care means, we need to be able to then differentiate
between the high quality providers by either paying them separately, by
shifting more volume to those higher quality providers, or by giving them some
form of public recognition for the fact that they’ve done a fantastic job as

But if we can’t measure it we can’t reward it and if we don’t reward it the
IOM has strongly stated that it really will not happen. And so as you can see,
I reiterated it before but I’m going to say it again, we really have to have
this data or we are going to just, we are not going to be able to achieve our
goals or the goals of the IOM’s To Err is Human are and those are critical
changes in our health care system that need to occur.

I did a very similar thing as what Trent did is that I took your very
specific questions and went through, I talked about what the recommendations
are, what we believe the benefits are, some of that is in general, and then
also to relate it specifically to what Leapfrog is currently doing so that
you’ll have a sense about where it would be able to help our current efforts. I
can also provide further detail later and just send it to you if you need
specific follow-up information about that.

So mechanisms for reporting lab results, that is it’s helpful for case
identification, for identifying appropriateness of care, for tracking
utilization when the multiple times the same labs are ordered over and over
again in difference doctor’s offices, that’s just a waste of a patient’s blood
and everyone’s time and energy. This is specifically useful in Bridges to
Excellence which Jessica is going to speak about, it’s a partner organization
with the Leapfrog Group. Rewarding results, we need this kind of information
and what it results is an RWJ Grant Program for six pay for performance
quality, pay for performance efforts that are being done across the country.
And these would support those efforts, Leapfrog administers the Rewarding
Results program.

E2 is an incentive and rewards program that has been developed by the
Leapfrog Group, it’s going to be going public later on this summer, it’s a
specific program where we identify, it is basically based on the Premier effort
that CMS did and we have developed it in partnership again with CMS and with
Premier to try to remain as standardized as possible. This kind of information
will help to allow us to administer the program, very important for that, and
also for physician decision support which we’re working on again with CMS and
with AHRQ, these are the kinds of information that we would need to have access
to to make this as easy as possible to collect data.

Going forward, I’m not going to go over every single one in each one of my
blocks, I will just call this the incentives and rewards. Basically all four of
these programs are major incentives and rewards programs.

Second is mechanism for reporting vital statistics for standard
transactions, I think that Trent is right in that there are certain, this is
just from my history as a provider, there are certain vital statistics which
are more, vital signs which are more important to others but I will leave those
decisions up to the experts about what those would be. But in terms of is there
a business value to it, it’s an important avenue to help to ease the tracking
of intervention and to reduce the administrative burden on hospitals because we
are asking for this information. If we ask for it in a standardized format over
and over again it’s going to make it much easier for hospitals to provide the
data. It supports two efforts, one is the I&R that I mentioned above and
the overall goal of reducing administrative burden on providers.

Flagging diagnoses that were present on admission, I’d say this was, and
I’ll go through it and what our top priorities are but this is one of our, this
would be one of our top priorities. It’s important for improving risk
adjustment and we started out focused on patient safety and we’re still
extremely concerned about this and this is the perfect way to help to identify
iatrogenic injuries, it’s a very important process that would help us.

Are there any questions? You looked questioningly at me.

MR. HUNGATE: — iatrogenic injuries —

MS. TITLOW: Iatrogenic injuries are ones that are caused by the hospital,
like you didn’t have it when you came in, I think.

DR. EDINGER: I’m curious, one of the problems when you come up with lab
results, standardizing data for reporting like the LOINC code there’s a
different, I went through this about 20 years ago, when you have different
values reported by different methodologies, for example for the same tests done
by five methods, the value you get may be different, that may mean absolutely
nothing because of the way the tests are standardized, normal ranges, etc. So
you can have misinformation by just looking at the actual number without
knowing the methodology and in some cases you can’t compare them, there’s no
linear relationship or simple linear relationship, and sometimes just getting a
number back may actually be more misleading then not even having the number.
Have you ever addressed those kinds of issues?

MS. TITLOW: This is definitely beyond the scope of what the purchasers want
to do, we are hoping that, honestly, that the experts in the field would
determine these issues and that we would then defer to the experts, that’s how
Leapfrog always works is that we convene expert panels and we’re trying less
and less to do that and we’re trying to be able to look towards standardized
panels that would be able to set those practices for us. We make cars and
that’s beyond our capacity.

MR. HUNGATE: We’re going to try and hold discussion until after all the

MS. TITLOW: All right, operating physician identifier code, this is
important, I’m only going to look at this from our perspective and from a
payer’s perspective and a quality perspective, which is we would really like to
be able to pay individual physicians for being able to do the high quality care
that they need to, that’s the same sort of problem that we’re having now is
that we tend to shift towards different health plans who we assume are giving,
are helping to provide the highest quality care. We are now looking at
hospitals but we really need to get down to that granular level of the
individual physician, this is a critical piece of information to be able to do

This is specifically it’s important for, we currently have a standard
evidence based hospital referral physician volumes, this would allow us to
collect that data, we’d like to shift, right now we only have hospital volumes
and new evidence is showing that individual surgeon volumes are probably more
predictive of better outcomes then hospitals volumes. We cannot collect that
data unless we get this sort of information in here, it’s just, we just don’t
have the capacity to do so. And we want to be as accurate and fair as possible
so in order to collect that we would like to have this information.

It’s also related to Bridges to Excellence which Jessica will talk about
more and we would like to be able to, we have a lot of recommendations from
hospitals and from physicians about our payment plan which is we’re only, we
only incent the hospitals for meeting our evidence based hospital referrals and
they said you have to really share that with the physicians because otherwise
the physicians themselves don’t have much incentive to try and improve their
own care, so if we had this level of data we could do a form of gain sharing
with them so we think that would be helpful.

Dates and times for admissions and procedures, this is relevant in
particular for evidence based hospital referral process measures which is for
evidence based hospital referrals which is one of the seven conditions that we
collect individual data on. We collect, we ask hospitals what volume do you do,
do you do specific process measures, and on certain conditions we try and get
outcomes measures and these are related to some of the particular hospital
measures. And this in a broader way support our incentives and rewards but it’s
not a huge issue for us frankly.

Episode start and end date, my director of leaps and measures explained
that it improved risk adjustments for us and therefore we get a lot of push
back about various risk adjustment methodologies since just like there aren’t a
lot of standard measures there are very few standards of risk adjustment
methodology and that would be, if you’d like to pick that after you finish this
entire project that would be good too. But this helps us to reduce the risk
adjustment methodology as that is improved.

And Trent is right, as a former physical therapist I am two thumbs up on
the create functional status code. This is really, when you’re talking about
employee’s functional status code is really where the rubber hits the road,
when can they go back to work, how much can they lift, can they still drive the
cars, this would really be the Holy Grail of being able to relate the care, to
be able to get your employees back to work. So this would be extremely helpful
for us and obviously as Trent said and you have recognized you have to set the
standards first, come to an agreement and then to be able to collect the data.
But I definitely think it’s worth, not just me but the organization thinks it’s
definitely worth investing in the effort to determine what those standardized
measures would be.

In ranking the values of the recommendations, we rank at the top the
mechanism for reporting vital statistics in a standard transmission and
reporting the lab results. And some of this is because we believe this is
really going to force the hospitals to put EMR in a faster rate and because of
that I think that there is more, there’s a wider benefit from, aside from just
the specific of these two different measures, that it would have a cause and
effect to really push EMR faster.

The flag diagnosis that was present on admission, this is something we
could turn around very quickly into incentives and rewards programs and to
really help to identify what things were caused at the hospital level and what
things were present before admission. And then I noticed that I somehow deleted
four which should have come after that. Seven and eight then again, functional
status code, the reason that it’s a little bit lower, even though I think it
should be one of the most important things we do, it’s realistic for us to
realize that it’s going to take longer to put that in place and so for my
understanding of your question was what’s the biggest bank for the buck in the
short term, that’s how I’ve used that to make these rankings.

And then further down dates and admissions, dates and times for admissions
and episode start and end dates. Important, would be helpful, but in terms of
overall rankings probably less important.

Correlations between recommendations, every one of them could probably
stand alone but since the question was asked we think seven and eight go
perfectly together, I think that’s fairly obvious. Three, four, and five would
have a bigger bang for the buck if they went together as would one and two.

And that’s it.

MR. HUNGATE: Very good, thank you. Chris, your turn.

Agenda Item: Purchasers – Panel 1 – Mr. Queram

MR. QUERAM: Earlier Trent asked for the committee’s indulgence to allow his
remarks to be made while sitting down, I’d like to ask for that same
indulgence. I would also have to confess though that I am the under achiever in
the group, I don’t have a Power Point presentation nor do I have a handout, so
I feel embarrassed and outdone by my colleagues but I’ll persevere nonetheless.

Much of the material that I’ll be drawing upon, we have developed and
submitted to the National Uniform Billing Committee and so if after I conclude
my remarks if there is an interest in seeing a little bit more of the
documentation behind some of the concepts I can easily make that available to

Let me tell you a little bit about who the Consumer/Purchaser Disclosure
Group is first by way of background. I introduced myself as the chief executive
officer of the Employer Health Care Alliance Cooperation, we’re a business
coalition in Madison, Wisconsin, some refer to it as the People’s Republic of
Madison, 25 square miles of land surrounded by reality. But Madison is a unique
community and one in which the business community and the public sector are
very organized in terms of the effectiveness around health care purchasing. The
Alliance is an employer owned and directed health care coalition, we have about
165 member companies, we pursue a direct purchasing model, direct contracting
model, with health care organizations, not through health plans. And one of the
attributes of our model is that we emphasize the importance of data and we
collect and maintain an administrative data repository that we use for
reporting of cost and utilization information. And increasingly which we’re
using for comparative provider performance reporting recognizing all of the
limitations inherent in administrative data.

The Alliance is one of the founding members of something called the
Consumer/Purchaser Disclosure Group which is an initiative that is supported by
the Robert Wood Johnson Foundation and the primary purpose of the Consumer/
Purchaser Disclosure Group is to advocate for a full dashboard of quality
measures so that by the date of January 1st of 2007 all Americans
will have information with which to make informed decision about hospital care,
about individual physician care, about medical group and integrative delivery
system care, and about treatment options in ways that cross all six domains of
the quality chasm, safety, timeliness, efficiency, equity, effectiveness and
patient centered measures.

A primary focal point for our advocacy is the National Quality Forum
because we believe strongly in the important role and mission of the Quality
Forum as a convener and as consensus based endorser of measurement sets that
will promote greater standardization. And we support it in part because
consumers and purchasers enjoy an institutional advantage in the National
Quality Forum which was a very conscious design feature, as a part of the
consensus based process and more importantly the board approval process,
consumers and purchasers occupy a majority of the board seats, which is a
significant recognition that our needs as the customers of the health care
delivery system have in many ways been unmet and under met by the providers of
health care services and it’s a way to redress the imbalance particularly
around issues of measurement and disclosure of information.

We represent about 60 large purchasers and consumer organizations, some of
the names are certainly far better known then the Alliance, groups like the
Pacific Business Group on Health, 3M and Motorola represent some of the private
sector purchasers that are involved. On the consumer side the National
Partnership for Women and Families, AARP and the AFL-CIO are among the members
in endorsing organizations.

Why are we focused on transparency? Part of the reason is because of the
expanded purchaser fiduciary responsibility and designing and offering a health
insurance benefit that ensures that there is a modicum of assurance of quality
for services that are provided to beneficiaries. And I think that sense of
fiduciary responsibility has been heightened in the wake of the Institute of
Medicine reports beginning with To Err is Human. Certainly we think also with
the movement toward increased consumerism and health care, particularly
consumer directed models of health insurance that place more responsibility on
individual stakeholders for making decisions, there is a need for greater

But perhaps above all is a recognition to play off of the old aperism(?)
that what gets measures gets improved. Our belief is that what gets measured
and reported publicly gets improved faster and we think that there is
increasing evidence to suggest that in the literature, our own experience with
the publication of a comparative hospital performance report in the fall of
2001 was the subject of a research article or research study that was reported
in Health Affairs by Judy Hibberg(?) that clearly demonstrated that hospitals
whose performance is reported publicly catalyzed and invested a greater degree
of resources in improving quality in the areas where their performance was
shown to be deficient then those hospitals that had the benefit of a private
report or had no report whatsoever. So we are firmly of the conviction that
transparency moves markets and transparency changes behavior, especially if we
can begin to link transparent information with rewards and incentives.

We are very interested in the use of billing or administrative data to
foster and support this movement towards increased transparency. We are firmly
of the conviction that there is no other means available near term to move
performance reporting. And while we appreciate and are sympathetic to the
tension that Trent cited between the need to do something now versus the need
to move towards more electronic means of gathering and reporting information,
the simple reality is that our consumers and our beneficiaries and those of us
who work closely with purchasers cannot afford to wait until such time as
electronic means of gathering and reporting information are widely disseminated
and adopted in the industry.

We also think the that the opportunity cost of doing nothing dwarfs the
direct cost of collecting information. We recognize that there is an added
burden associated with collecting and transmitting additional administrative
data elements but we do believe that the opportunity to use that information to
drive improvement, to make better decisions, and to inform quality improvement
efforts within health care organizations outweighs the cost of collecting that

About a year and a half ago we learned that the National Uniform Billing
Committee was beginning its every decade or so process of looking at changes to
the UB-92 form, for a whole it was the UB-02 form, now I think it’s the UB-04
form so its become a decade process that’s morphed into a 12 year process. And
one of the members of our staff who spends a lot of his time thinking about
ways to use information to drive a value purchasing agenda, and to use
administrative in a responsible way, realized that we had this window of
opportunity to try to influence the NUBC process, literally a once in a decade
opportunity to try to get more information that could support better
performance reporting and pay for performance demonstrations.

We brought that to the attention of the Consumer/ Purchaser Disclosure
Group under the leadership of Arnie Millstein who serves as the medical
director for this initiative. We convened an advisory panel, a couple of people
in the room today were a part of that advisory panel, hopefully they remember
that, so we were able to vet a number of suggestions and what we came out with
were six broad areas of expansions of administrative datasets that match very
nicely with the recommendations in your report. Our recommendations align
perfectly with recommendations one through six in your report. With all due
respect and deference to Karen’s comments about the importance of functional
status measures, that was not on our list, but what is on our list are the

Secondary diagnosis codes present on admission, we would rate that as a top
order priority in terms of our ability to differentiate complications from
comorbidities and begin to move forward with more appropriate severity
adjustment of hospital performance. We experienced this two and a half years
ago when we published our own report on hospital performance and referenced it
in very strong terms using some of Judy Hibberd’s research on the importance of
evaluability and talked about the need to select hospitals that have fewer
rates of complication deaths and medical errors, not often terms that you see
in public performance reports associated with hospitals. And one of the
criticisms that was levied against the report is the very legitimate concern
about complications that are present on admission versus those that arise
within the hospital setting and we think the addition of a secondary diagnosis
code to flag those conditions present on admission would be a significant
improvement in administrative datasets.

The second area that we are interested in is unique physician identifiers
for each hospital procedure, which allows an outcomes assessment by hospital
procedure lists and we believe is more precise then retrospective linkages of
records. We’d like to see as a third area vital signs at admission, things such
as heart rate, blood pressure, temperature, respiratory rate. We believe it’s a
powerful predictor of mortality for common conditions like AMI, congestive
heart failure, asthma.

Fourth area we’re interested in is key lab values at admission
notwithstanding some of the technical concerns that were suggested earlier. A
fifth area we’re interested in is do not resuscitate orders that are present
within the first 24 hours so that we can distinguish most pre-complications
from post-complications to patient do not resuscitate requests. And then lastly
we’re interested in time of admission, discharge, and procedures so that we can
do a better job of understanding and reporting delays in treatment for patients
within that acute care setting.

A couple of comments about the ability to use this type of information for
pay for performance and some of you may be aware that our organization, the
Alliance, hosted a pay for performance conference in Madison in early May which
was funded by AHRQ and which was designed to begin to introduce some of the
concepts in an evidence based practice center report that AHRQ is issuing soon,
the current state of the art of pay for performance initiatives. Some of the
takeaways for me from that conference is that we don’t know a lot about what
works in the area of pay for performance and it would be I think risky for me
to sit here and suggest that all of the items that we have outlined as being
important to us in terms of expansion of administrative datasets will easily
and readily lend themselves to pay for performance demonstrations. The reality
is we don’t know what works, we don’t know the size of the incentives, we don’t
know exactly how to motivate changes not only in provider behavior but
beneficiary behavior. But this is a time of critical importance in moving
forward with experimentation, you’ll hear in a minute about an important
demonstration known as Bridges to Excellence, and we are of the conviction that
employers are motivated to begin to experiment and with this type of
information I think you will see a change in behavior on the part of purchasers
who are looking for strategies and techniques to begin to combine information
with meaningful incentives to not only impact provider performance but also
change beneficiary behavior and decision making.

With that I will stop and yield whatever time I might have left to Jessica
and look forward to the comments and questions.

Agenda Item: Purchasers – Panel 1 – Ms.

MS. DILORENZO: Good morning, my name is Jessica DiLorenzo, I do work for
General Electric, I also work for a not for profit company called Bridges to
Excellence. And I say that kind of in a tongue in cheek way because there
really is not staff at Bridges to Excellence, I think you’re looking at it,
with all the partners like MedStat, and all the partners that we have we’ve
been able to do a lot of work with the people that we have involved.

Really what I wanted to do today was to talk about the recommendations
within the framework of Bridges to Excellence and also what GE’s thoughts are
about transforming the health care marketplace, and review the mission, talk a
little bit about the specific Bridges to Excellence products, and then do a
high level crosswalk between your recommendations and the metrics, the
performance metrics that are related to Bridges to Excellence. And I have
probably the advantage of being last because I can just say I can agree with
everybody previously, I won’t do that, I may be a little bit quicker but that’s
okay, I have to admit that up front.

First of all I just wanted to say that from the GE perspective, so I have
two hats on, I’ll try to delineate myself when I speak here, but from the GE
perspective we imagine a health care market where service providers are free to
compete based on the published value of the services they provide and are
rewarded for high performance. So this is really what’s motivating us and why
we think what the committee here is doing today is so important because there
are some very key words in there for a very good reason and the transparency
piece of it as we have heard before, I have to agree 100 percent, is what’s
going to drive the marketplace. So the published value of the services that
they provide, and we mean here also that the networks are not necessarily
restricted, the consumers have choice and that they are free to choose and that
they are sensitive to the value of the services that they consume and they’re
engaged. So that’s a very lofty mission and something that we’re very committed
to through Bridges to Excellence and Leapfrog Group and some of the initiatives
that we’re working on.

I think it’s not really news to anybody that to create this market we’re
going to have to solve a few problems the first of which we’re focused on today
which is the harmonize the performance measures on effectiveness and efficiency
across the industry and that will go you to that value quotient for the
providers or for the purchasers in getting involved in this. So working with
the health plans and developing the products that will encourage consumers to
shop for value. So as a purchaser and as speaking from GE’s perspective we are
at this point involved in these initiatives to be a catalyst in the market but
to also work with our health plans and other purchasers to move the market so
we can continue to leverage all the good work that we’re doing. And it’s
important to succeed in a significant market experiments like Bridges to
Excellence so that we can demonstrate that the change is possible here and we
have started to do that with Bridges to Excellence in the four markets that
we’ve rolled out in.

Bridges to Excellence is an answer to some, I think that should be in
quotations here, maybe not problems, but Bridges to Excellence is an answer to
some of the problems, it is a multi-stakeholder, for those who don’t know much
about Bridges to Excellence it is a multi-stakeholder approach to creating
incentives for quality. It is purchaser funded so we’re made up of large and
small employers, Proctor and Gamble, Verizon, UPS, Ford, hopefully I’m not
forgetting anybody, and some local and regional employers as well, retail
market chains and such in upstate New York.

When we came to the table to talk about Bridges to Excellence we didn’t do
it alone, we did bring the health plans, we brought experts in industry, we
brought the providers and the consumers and really hashed it out for a good two
years as to what the measures should be, what the incentives should look like,
what was important to everybody, we used a sick saving quality rigor in the
project so that involves getting all of the customers point of views in this
process and all the stakeholder point of views in designing this so that we
broke through as many barriers as we possibly could and could have answers to
potentially any of the sort of responses we got along the way.

You can our mission there, is to improve quality of care through rewards
and incentives that encourage providers and patients so again, it had to be a
two-fold effort, it wasn’t just focused on one segment or the other. And we
knew that we had to do that just based on the economy of the way that it would
set up and also the physicians told us that you can’t just beat us with a
stick, you have to give a carrot out there on the patient side as well to be
motivated to change. And we do have that as part of the program as well that’s
just getting ready to launch so I don’t have any specific data to share on

The focus is on office practices and then two specific condition areas,
diabetes and cardiac care, were rolled out in selected markets. There’s four
areas right now that we’ve brought up in the last year and what the very key
selling point for me on Bridges to Excellence is that we’ve been able to create
with our partners an operational model that is very scaleable to pretty much
any area and I think that what we’ve seen in some of the other valuable
demonstration projects is that they’re very specific to an area or a situation
and it’s hard to sometimes pick those up and move them to other areas and
replicate the same results. Bridges to Excellence was that the way that we’re
set up with the rewards and the measures that are standards of NCQA. We are
able to pretty much bring that into very different markets with very different
physician levels of organization and still make an impact. Program costs are
pay fully by the participating employers.

Specifically Bridges to Excellence has like I said a diabetes care link,
cardiac are link, and physician office link, those three names are the names
that are related to the Bridges to Excellence product. The measurement sets are
administered by the NCQA, the NCQA had national measurement sets prior to
Bridges to Excellence coming along in diabetes and then later in heart/stroke
recognition program. We haven’t really had much push back from the physicians
at all on the specific measures and I’ll show those a little bit later so
they’re fair but they’re tough and I think that with the proper infrastructure
I place many more physicians will be able to report on their activities and
gain recognition.

The objective of course is to improve the outcomes for patients with
diabetes and cardiovascular disease and we’re targeting the primary care
physicians groups and the specialists that treat those types of patients.
Physician office link is something that rewards physicians for all covered, so
it’s population or population management at the practice level, redesigning the
processes of the care and of course the Quality Chasm and all primary care
physicians that can produce the certain amount of records on those patients are
eligible to apply.

The fourth one which is new is hospital care link and we’re working with
the Leapfrog Group on that and we’re looking to basically use the Bridges to
Excellence operational model to leverage the hospital care link and I think
that we’re very new in those discussions.

This is kind of an interesting slide that I like to show, I mean what you
can see here that stands out to me immediately in our discussions today, at the
very center of the slide is standard measures and pretty much all of what we’re
trying to get out here is focusing around the standard measures and of course
you can’t get the standard measures unless you have ways of reporting them that
will ease some of the administrative burdens for physicians and hospitals to do
that. So basically you can see how we’ve organized around this and how we’ve
approached this, you could say that with Leapfrog Group and Bridges to
Excellence the first step really was there to organize a large amount of
stakeholders around an objective and to get a lot of focus and visibility if
you will for the missions. And then create some standard measures that focus on
process and outcome improvement so we can move, shift to hospitals and we can
change the way the physicians are practicing.

And then again, using measures that will, and this was important to
purchasers of course, find measures that will be actuarily sound, so that there
is a return for the purchaser and there is a business case for purchasers to
get involved in moving the market. We do want to improve the quality,
absolutely, but there’s always, we’re looking at the savings at the very
individual metric level as well especially when we take a look at blood
pressure and what they control.

So that’s the framework that we’ve come up that we think is going to break
the gridlock and why we think the discussion today is so critically important.

I wanted to show what the process and outcome measures were for, and
actually DPRP is actually the correct title for that, the Diabetes Provider
Recognition Program through NCQA, so you can see that the measures are HbA1C
measures so they’re not just process but they are actually lab measures as
well. And I can tell you the burden of getting this information for some of the
practices is a barrier that’s keeping them from going forward. And it’s
especially true in some of the smaller practices —

PARTICIPANT: I have a question, is that correct, cholesterol less then 100,
less then 130?

MS. DILORENZO: That’s a range, correct. So these are the NCQA measures, I
also wanted to point out here that we have two measurement sets, there’s a
three year and a one year measure, the one year measure is actually
specifically for Bridges to Excellence and this was in response to what the
physicians had said early on that they wanted to just sort get their toe into
the process and maybe they didn’t want to go for the full measurement set, and
provide a subset of measures for us that would sort of give us an in. No one
has elected for the one year certification process at this stage of the game.

What also comes along with the one year certification is a risk adjustment
option, that again was in response to what the physicians asked for, our
patients are sicker then everyone else’s, we’re not going to be able to make
it, please provide some sort of risk adjustment for us which we did design in,
again no one has opted for that. So while I think that making sure that the
measures around risk adjustment are sound from our perspective, at least at the
physician level, early in the process no one has opted for taking the risk
adjustment. And it’s free, there’s no cost associated with that as well.

Taking a quick look at cardiac care link performance assessment, again
there’s a three year and a one year. What’s different for cardiac care link is
that there’s no reward associated with the one year measure. And the reason
that that is true is because when we took a look at the actual return on
investment for passing each of these performance assessment metrics we were
hard pressed to find a savings quite honestly with some of the measures and
blood pressure and lipid control was really going to give us the bang for the
buck. And so what we did instead was rather then coming up with a blended
reward fee or blended reward for physicians and sort of weight it out between
very, very high performers and then some maybe okay performers we decided to
come up with a tiered reward. So someone who is performing a the highest level
of blood pressure and LBL control they will get, or blood pressure control,
they will get $160 dollars per patient per year as opposed to the $80 dollars
for the three year.

The one year again, the subset of measures that we did come up with and the
reason we did keep it in there was that part of the recognition is not only
rewards but getting publicly recognized as well so we wanted to at least keep
that in there as an option, again, these one year options may sunset depending
on how we proceed, we’re still very new in the roll out, we just completed the
roll out of our fourth market so I think time, we’ll still need to play that

I wanted to from the physician office link try to just quickly crosswalk
for you between some of the physician office link measures that are there,
whether or not they have a basic registry or an EMR so the span there is
anywhere from a basic shoebox that’s organized really well to a fully
functioning EMR within the office so that does qualify for a year or so
physician office link. So the education and screening measures, quality
improvement, they need to show that they have quality improvement of health
outcomes and performance goals programs within their practice. They have to
show that they have chronic care management going on at the office level, that
they’re doing something about preventable admissions and working on
identification and follow-up with those patients, and what are they doing with
high risk patients.

So I wanted to just lay out some of those, they’re our candidate
recommendations that I thought would specifically help in that physician office
link application process and how this would ease the burden. Again, the
physician office link application process is very long and it’s very intense
and there’s some redundancies and we’re working on that. And I think that’s why
it’s good to call Bridges to Excellence a pilot and we’re learning as we go and
we’re willing to learn as we go and we are getting ready to sit down with
everyone who has just gone through the application process and we have a good
sample of them now in Boston, we’ll have somewhere around 37, actually we’ll
have about 67 practice sites that will have applied for physician office links
so that’s a good amount of people to sit down and say okay, what can we do
better about this and what are the process changes that at least we can make on
our end. I think there’s work here as we talked about making sure that some of
these standardizations happen across the board that would even make that less
burdensome to the practices.

Our overall response is basically that competition should happen at the
individual physician level by disease and by procedure. Today competition is
really at the health plan and the network level and we can’t have competition
if we don’t have robust comparisons and accurate data. Competition at this
level is only possible through improvements in the outcome and administrative
data and making it transparent, I think that that’s a theme that you’ve heard
over and over already today. Bridges to Excellence will continue to move
forward with the status quo because we need to sometimes just go about bringing
about change in the environment that you’re in. I think that we’ve been pretty
clear that we’re kind of a catalyst to change and we need to keep marching
forward and hopefully improvements will come as we go.

For purchasers to buy health care based on value outcome data needs to be
readily available and transparent in the marketplace, value is created when we
have a market where providers compete and consumers can shop with their feet, I
mean I think that everyone would pretty much agree with that.

I have to agree with Karen as well that number one and number two are the
highest priority for us in the sense that test results and vital signs and
effective data is where we see the largest room for improvement and that this
would definitely ease the barrier in provider participation. What we’re seeing,
I think that what we’re seeing in some of, in actually two of the markets I
know two specific cases where they’re going through the application process
with Bridges to Excellence or for NCQA and they found out that they can’t pass
and that’s where we’re finding that the large amount of value Bridges to
Excellence quite honestly is that they’ve noticed just going through the
application process and burden and assigning resources and physicians are
finding very creative ways themselves to get through this process. They’re
getting graduate students in there, they’re contracting with grants to get
resources in the offices to be able to this, they currently just do not have,
most of them don’t have the resources within the office structure to do pay for
performance applications in the current format that it’s in and we’re trying to
help them with that as well, QIOs, I mean we’re, we’re just trying to figure
out a way to get over that hurdle for some of the small to mid size practices.

Boston and very highly organized situations, they have the resources and
they’re able to bring in people to get through the application process and they
have the infrastructure and some electronic data now so they’re able to submit
a little bit more easily. But what we’re seeing is that when they do look at
the day they’re going to make changes. And they tell us they’re making changes
because they weren’t able to pass and even if they were able to pass it was
well we found room for improvement so obviously this is what, this is the crux
of where change is going to happen and when they have the data readily
available and it’s accurate and it’s at the patient level.

Secondary admission diagnosis flag and operating physician would improve
the precision of hospital performance measures, of course we’re interested in
that as well through the Leapfrog Group. Dates and times for admission and
procedures would allow for measures of timeliness of hospital care, quite
honestly a lot of these, some of them are relevant to Bridges to Excellence,
most of them are more focused on the hospital side so I’d have to agree with
Karen and her assessment of those. Functional status coding, coding of health
status and severity of illness would better account for individual patient
differences which of course is critical for us to know. And obviously if we are
able to get that data on the physicians at a more accurate level we would be
able to reward them and create an incentive system that would better reflect
the performance.

In summary we want to reward physicians for doing it right the first time
and for us that creates the business need for accurate and complete data at the
individual physician procedure level. Standard data formats will help
physicians and practices leap the administrative hurdle of self reported
performance standards. And I also wanted to leave the door open that Bridges to
Excellence is open to exploring any of the changes in performance measures in
the promotion of data exchange and interoperability.

MR. HUNGATE: Thank you very much, I commend all your presentations for
clarity and content, very good, very helpful. Let’s move to questions on
content and clarifications, spend about 20 minutes doing that, and then I think
we’ll take a little break. And then come back and I’d like to come back to
start to try to articulate what could be called a business case around these
recommendations in terms of what are the tradeoffs and how do we quantify the
merits of the various tradeoffs in efficiency. So we’re open for questions.

MS. POKER: I have two comments or questions and it’s kind of an overall
question and I think I’m going to pick on Trent just because he was the first
one that triggered this thought but I’d love input from anyone about this. One
of my concerns as we’re going ahead with making very high level health policy
decisions based on the quality of the data that’s being captured, do we know
that we’re capturing quality data? My concern is with the MDS data, in my
previous life I’ve been both an MDS coordinator and an MDS trainer of other
coordinators and I’m here to tell you it wasn’t captured accurately
unfortunately from the level I saw. Now I can’t state nationally to that but I
have talked nationally to other MDS trainers, it’s a poor pitiful issue out
there. And in other words if you’re going to be basing serious decisions in
health policy are we looking at the reliability of this data. I’m only talking
of one aspect of it and how are we going to look at it.

And as another real quick comment too to leave out to the panel to think
about, one of my concerns as we’re going forward with this rewarding and
incentive program for physicians, and this is just a food for thought for
anybody here, are we going to be increasing the disparities of care and access
to care because are we going to be getting towards a system where the docs are
going to say well, it’s worth my while to work in areas where the population is
less sick, there’s less chronicity, I’m going to have a better rate. I’m just
throwing this out there, I don’t know the answers to this but this is one of my
concerns, I work on the disparities reporting and we do have a huge issue in
this country about that and do we want, just kind of tread cautiously.

DR. HAYWOOD: I’ll start off and the wealth here. As to do we know that
we’re capturing quality data, I think we are still in the earlier phases but I
think we do know to the level that we’re actually measuring data how reliable
and valid and for what its intended purposes are. In other words the measures
that we absolutely incorporate into our activities, we actually do that in a
process whereby for instance with MDS is actually to go back and compare those
measurements with activities that we’ve actually not only with dementia but
with actually activities that were occurring in those nursing homes to
determine whether or not those nursing home standards were actually having
activities or processes or systems in place that would actually be linked to
those actual individual measures thus improvement in those particular measures.
So that’s how kind of the nursing home quality initiative was actually piloted
was not only looking at the validity and reliability of measuring but also
comparing that to a processes that you would know shows improvement in care
along those particular continuums. So that’s how that was tested to find out
whether or not they were actually accurate.

Similarly I think in the other things that we are looking at particular
measurement, that’s why you have so few measures right now to be honest because
there’s a lot of people that are concerned about your particular interests, is
that we don’t want to move so quickly the science isn’t sound and I think
similarly on the purchaser side whether the purchaser sector or even more so on
the private sector we continue to hear that no one wants a system built upon
something that’s fallible. And so we continue to work not only with kind of the
experts but we also work with the individual providers and clinicians to make
certain it’s not only kind of works from the scientific statistically sound,
that type of perspective, but also from the true at the bedside perspective as
to whether or not it actually has an impact at the clinical level.

So I think we hear your concern, we always continue to have that concern
and actually with each measure that gets added it goes to the full process and
part of that is not only kind of internally at the measure developer stage but
also when you actually roll it out to the National Quality Forum at the
consensus level stage as well, that it’s re-vetted again throughout that whole
process taking a look at the particular concerns that you raised.

As to the second question, I’m going to broaden it a little bit but I share
your concern which is unintended consequences. And particularly from the
private sector because I’ve continued, and even to my private colleagues where
we are in total agreement I think with a market based approach to actually
having some improvement strategy that allows that it may have unintended
consequences. Now the unintended consequences that you specifically mentioned
was disparities and access, because there is the possibility that the rich get
richer and the poor get poorer if it’s built upon a system in which people have
to already have resources to actually do certain activities so for instance
Bridges to Excellence in Boston, as Jessica indicated it was easier because
they already had the infrastructure in place. And so if you were to design a
model in that capacity, if you didn’t actually try to go into the particular
communities and get them on a certain footing they would be disadvantaged in
this market.

Having said that there are activities that are underway where this may
actually improve that in the sense that if you are really actually doing risk
adjustment and looking at people that have more comorbidities there may
actually now be a financial incentive which there currently is not to actually
take care of those particular patients. And so it really becomes I think a
matter of how we actually one, increase the infrastructure in those particular
communities, and if we really do do kind of the risk adjustment as we pay based
upon that risk adjustment that may finally be a financial incentive to actually
start taking care of those particular populations.

MS. TITLOW: Just a couple of comments. One is I think we need to draw a
distinction between concept and execution. So the idea of collecting the data I
think it’s hard to be against that, the execution of it I think gets more to
your point, do we have, how well are folks trained, how accurately do they
report the data, I think that’s a very separate issue and one that could be
considered separately. And of course you want to have execution to be the best
that it can but because it isn’t currently being executed in a perfect way it
shouldn’t interfere with the appropriateness of the concept that this data
needs to be collected or we will not continue to improve at the rate that we
need to improve.

And so I would caution you against mixing those two issues up, I think
that’s something that has really, this is something that we see from the
purchaser side, this is how we see the system, which is that at NQF, often at
CMS meetings, there is such a push back from the perfect getting in the way of
the better. I mean it is not a good health care system now, there is a huge
capacity for improvement and because there is such a frequent push back yes, it
could be perfect, it’s slightly off, this won’t be exactly right, that’s true
but that shouldn’t stop us from trying to take the improvement steps that we
could take. And I strongly urge you not to let that happen, to continue to let
that happen, that’s been the way our health care system has worked for 20 or 30
years, there is a huge amount of room for improvement before we get to the
point that we say that this is imperfect and so we can’t do it.

And I guess that’s just the biggest point is that it is a fallible system,
I was a provider, I’ve been an injured person, a health care person, and health
care is fallible 100 percent, I mean that is the way that it works, we are
human beings, we don’t understand our own physical bodies that well, providers,
there’s an emotional aspect to treatment and so it’s going to be a fallible
system, it’s not like gears working together and I think that we should
recognize that, that it’s a health care system.

MR. QUERON: Just a quick comment on the first issue that you raised and
that is the whole issue of the integrity of datasets. I think you get the sense
from Karen’s comments that you won’t hear a lot of sympathy from the purchaser
community that we need to wait for perfection in data before we make movements
and we’re a living example of that with the experience that we’ve had in our
community using administrative data to publish hospital comparative performance
reports. I do think that there is an argument to be made that using information
makes it better, that it becomes an incentive to improve the quality and the
integrity of datasets when it’s used responsibly but perhaps pushing the
envelope to invest the time in coding practices, coding procedures, data
auditing techniques and so forth to improve the integrity of the information.

About a year and a half ago the National Quality Forum had a vigorous
debate about the adoption of a hospital performance measurement set. And one of
the things that that debate highlighted is that there is a fundamental
difference between the, we characterize it as values, the values of the
industrial purchasing model and the values of the professional model within the
health care industry.

If you look at how the decision criteria are applied and weighted in
vetting performance measures strong emphasis is placed on the scientific and
clinical credibility of a measure, on the feasibility of collecting the
information, the burden issue, and certainly everybody agrees that there needs
to be an importance to the measure, that it’s linked to a clinical process that
can be improved. But usability is often weighted significantly lower and
purchasers and consumers are looking for usable information with which to guide
decision making, we would place a higher emphasis on the usability of
information for pay for performance, for consumer decision making and for
public accountability then we would on scientific and clinical credibility.

That’s not to say that that’s an important value that should be discounted
but in our world we would weight the decision criteria higher. And we have
often said sometimes to the consternation of our clinical and scientific
colleagues that 75 percent accuracy is sufficient for decision making rather
then 99.9 percent accuracy.

DR. COHN: Actually I have a couple of question, first of all I want to
thank the testifiers for some very interesting testimony. Now first of all
Christopher you had commented that you had had conversations with the NUBC, and
the good news is I’m not on the NUBC so I don’t have to recuse myself, that’s
one of the few committees I’m not on. Potentially can you update us on the
status of those conversations and then I have some very focused questions in
relationship to some of those.

MR. QUERON: No, I really can’t, I would have to say that the UBC process is
a Byzantine one that I have not figured out so maybe defer to your questions.

MS. GREENBERG: I am on the NUBC, as staff I don’t have to recuse myself on
anything I don’t think. But I’m into disclosure of course and I am a member of
the NUBC.

Arnie Millstein did present on a few occasions actually and the UB-04 has
not been finalized but it’s pretty, the current version of it very likely to be
approved in the next, I think probably at the August meeting but definitely by
the November meeting, does include the indicator for the flag for secondary
diagnoses, which already is actually part of the 837-I, the institutional claim
standards, the HIPAA institutional claim standard, although it’s not currently
in the HIPAA guide, the implementation guide, but it is part of the standard so
to move it into the actual implementation guide, the next version would not be
that difficult because it’s already in the standard.

And one of the interests of course of the NUBC is to align itself as much
as possible with the electronic version recognizing that for hospitals in
particular the vast majority do use hospitals. On some of these measures the
committee advised Dr. Millstein that paper solutions were probably not that
desirable and that electronic solutions would be much more likely and easier to
implement. But he responded that there still are hospitals that do not bill
electronically and they really don’t want the patients at those hospitals to
just be forgotten. So that will probably be added, at least the current version
has that on the form, and there also are, I’m calling them code to code values,
there are some fields that will be on the UB-04 that would be used for the
vital signs and the lab values if one were to choose to do that, there’s real
estate there that’s kind of been set aside in which you would indicate, sort of
like value codes or if you know anything about the UB forms where it has
different types of code values so you would have a code that says this is, now
you’re going to see the blood pressure and now here is the blood pressure or
something like that. The infrastructure could support some of that and probably
that’s true about several of these other elements.

But the only, I believe the only one of, and as you said they’re quite
aligned with these first eight, the only one in which it would be specifically,
it seems to be strong consensus on is this flag but I think there would be
potential capacity in the new form for some of the others as well.

MR. QUERON: When would a decision likely to be finalized on that?

MS. GREENBERG: The next, it’s certainly this calendar year, there are two
more meetings, one is in August and then the last meeting of the year is in
November and I can’t remember at which one of those two meetings they’re
actually going to vote on this because this is also out being looked at by the
state uniform billing committees as well and so we have to get feedback from
them but I would say at the latest November.

DR. HAYWOOD: Simon, I can I just ask just Marjorie real quick from a timing
standpoint for the activity at this staff, that the Workgroup on Quality is
doing, how does that relate to if some of the recommendations is directly
related to EUBC and it’s only going to occur for a decade, I didn’t know how
the timing that this activity relates to that.

MS. GREENBERG: The committee is aware of the fact that this, the Uniform
Billing Committee is aware of the fact that many of these issues are also under
consideration by this committee and it’s true, if it’s August it would be
different then November. I’ll find that out, I’ll actually try to find that out
at the break.

DR. COHN: And I guess I should clarify, you commented about the role of the
standards development organizations is a little unusual, one obviously needs to
recognize that most of hospital data submissions is electronic these days, I
mean CMS, virtually all of your stuff comes in now on an 837 or what used to be
the NSF format and all of this, I mean obviously the paper form has less input
but it’s more the intellectual leadership of the NUBC in all of this.

MS. GREENBERG: If the UB-04 has this flagged for secondary diagnoses and
said this is a required element it would have an impact on anyone reporting
hospital data.

DR. COHN: The other piece on this of course is the number of states
obviously already adopted this many years ago, California for example has been
using this modifier 15 years, 20 years —

MS. GREENBERG: New York and California. Actually another one that isn’t in
these recommendations but is in the consumer disclosure recommendations, that’s
the DNR, and the Public Health Data Standards Consortium on behalf of
California, the Public Health Data Standards Consortium has a state and a
federal representative, I’m actually the federal representative on the NUBC. On
behalf of California and other states who have expressed interest in this is
bringing that forward as a proposal in August.

DR. COHN: I was going to ask just a follow-up question on this but would
you rather have —

MR. HUNGATE: No, go ahead.

DR. COHN: Okay, so now that we know what the update is I actually had a
very specific, I think Trent you’re right, actually please tell me how many
emergency physicians we have that we’re dealing with, I had no idea that Trent
was an emergency physician, Shawn Toomas(?), myself, all the physicians,
actually not all, but many of the physicians on the NCVHS are all emergency

But I find myself continuing to be intrigued by item four and I’m curious
about all of your comments about all this about operating physician identifier
codes. Now I guess I need to understand from all of you about this issue, now
Trent didn’t think that this was a big deal and he thought the administrative
burden was probably worse then the data capture value but the rest of you
obviously want this very badly. Now when I tend to think about going and
collecting data I typically think about going to the source as opposed to a
second or third party for data if I’m trying to find reliable data, and one
would observe that probably the physicians know best, whether they’re the
operating physician or not and indeed in the world of physician billing things
are not quite so tough, there’s the main physician but then there’s also things
that identify whether it was a group procedure or a co-lead procedure, I mean
this is the world of modifiers because obviously there are lots of as you
described nuances to all of this.

Now knowing that this stuff is already being routinely submitted because
obviously there are bills from physicians into the health care system I guess I
would need to understand why this is an item that’s being listed on the
hospital form and is this really just a convenience issue for the people
dealing with data or something to get around physician resistance to this. I
mean can you explain to me why this is an item that’s on this level of

DR. HAYWOOD: I think I can try to explain some of that which is there is
confusion because the Part B billing doesn’t easily match for your if you’re
trying to get at that Part A stay. So in other words if I’m the physician
that’s doing a particular procedure on a Part A stay I may bill it, my group
may bill it, so you may have, you may never get a bill that specifically allows
you to say that Trent Haywood actually did the procedure, what you would get is
a group in which Trent Haywood is a member of Billed Medicare Program, and that
relates to a particular HIC(?) number, but after actually getting to the
individual date and service that’s where you continue to have the disconnect.
And so this is a way of actually making that link pretty straightforward and
expedient so that you don’t have to actually crosswalk between trying to get
down to the physician level on the Part B claim which you apparently aren’t
doing a lot of times and then trying to walk that over to the Part A to
actually get at the service that you actually have one of the interests. So I
think that’s where the biggest issue is, Simon, with actually trying to use the
Part B data for this.

DR. COHN: Any comments? Is this the issue or is it something else?

MS. TITLOW: I guess my only comment would be that we were asked to react to
it as opposed to develop it so I don’t know —

DR. COHN: [Inaudible.]

MR. QUERON: We weighed in with the UBC committee and the rationale, I’m
going to read it so that it’s clear for this is that we believe that this would
enable outcome assessments for physicians performing inpatient procedures that
complements existing UB-92 data elements but more importantly eliminates the
need to link institutional and professional databases. So it could be easily,
well, I’ll say easily, it could be collected by the hospital and reported as
part of the UB-04 for procedures that are performed on an inpatient basis. And
it identifies specific physicians so that we can begin to look at more
granularity, more outcome assessments of inpatient procedures.

MR. HUNGATE: Cathy and then let’s take a break.

MS. COLTIN: I wanted offer one additional rationale and that is that not
only is there a very real problem in that the professional claims frequently
only provider a provider ID that represents some sort of a billing group or
billing organization rather then an individual but also professional claims
data are generally proprietary, they’re not out there in public domain whereas
most hospital datasets are available, most states have publicly available
hospital datasets that can be used in the public good to create these kinds of
measures and this kind of consumer information and therefore it would make this
information more widely available.

MS. GREENBERG: Well, I just had a quick question for this one person on the
panel —

MR. HUNGATE: Same subject? Same subject, not new subject, right? Or is this
a new subject?

MS. GREENBERG: Well it’s the subject we’ve been discussing all morning. I
don’t know how long they’re going to be here —

MR. HUNGATE: Can you folks stay for the rest of our discussion or do you
need to leave?

DR. HAYWOOD: Is there a way that we could actually, I know I unfortunately,
and I’m feeling bad because I’m on the staff but I’m going to the airport from
here —

MR. HUNGATE: Not take a break —

DR. HAYWOOD: Right, skip the break.

MR. HUNGATE: Okay, we will continue. I’ve got Stan, I’ve got Harry, I’ve
got Peggy, I’ve got Marjorie, and Julia. Peggy, can I let her go first?

MS. HANDRICH: Thank you. Thank you all for your excellent testimony, it was
very interesting to hear the agreements that you shared and also some of the
variation in the comments that you made, and questions and comments that I have
go to one of the areas where there was a little bit of variety in your reaction
and that had to do with the priorities you placed on the functional assessment.

I thought it was interesting that Karen and Trent spoke somewhat
passionately to the importance of standardized data to inform functional
assessments and the two of you, Chris and Jessica, did not. And I perceived
that the perception of the importance of the functional screen from the
perspective of Trent and Karen had to do with actually very different
populations. In your case Trent you were looking at basically the long term
care population, and Karen I thought it was so interesting that you made the
comment we want to know when they’re ready to go back to work, a different
population. And if you can confirm that that’s the case when I get done that
would be great but this is kind of a question for all four of you.

I wonder why it is that Jessica and Chris didn’t assign as much of a
priority to that, if you could speak a little bit more to that, and what it is
more specifically that you think when you spoke to the need for standardized
information, if there is any role that the workgroup or the committee, you
would be interested in seeing played out in terms of the development of that.
And I want to invite you to launch on all of this with a reaction that I had,
you all spoke to kind of the need for moving forward the quality agenda by
influencing provider behavior and also by informing and energizing the consumer
community so that they can make informed decisions.

We have a very small but very popular project in Wisconsin known as Family
Care that includes a personal functional screen being performed for every
individual in this model in conjunction with a case plan being developed. And
this project has been evaluated at least three times already and in each of the
evaluations it has come to the fore that consumer participation in a functional
screen has informed and energized the consumer to make better decisions about
seeking health care, which provider they wish to have that health care be
provided by, using information that’s available, and also making decisions
about choices, where for example a person might say yes, that would be the
ideal thing for me but I don’t think I’m going to go there. So I was interested
that none of you kind of saw the functional screen as being a lever on the
consumer side and I wonder if you could all react to that.

MS. DILORENZO: Well honestly the functional status and my perception of
that was more of the elderly population, that was my take on the
recommendations and Bridges to Excellence specifically is focused on active
employees so that’s why I fell a little bit lower on my priority list, not that
it’s not important, if you’re talking about health risk assessment and
functional assessment and that way, important, yes, difficult to get done on a
wide scale basis.

MR. QUERON: My only comment, and I don’t disagree with the importance of
functional status and I think representing an employer constituency we would
love to be able to have an idea of a holistic perspective if you will of how
health care services are provided, particularly as it relates to productivity
and presenteeism, which is of keen interest to employers. But part of where we
were coming from is we’ve been through an iterative process in part through the
NUBC and we want to be somewhat parsimonious in the request that we make of the
system. And our perception is that the greatest return on investment in the
short run would be through the addition of these six data elements that we are
now focusing our advocacy on.

MS. GREENBERG: If I could just comment because I specifically, I personally
have a very strong interest in functional status assessment professionally
because of again truth and advertising responsibility for the international
qualification of functioning disability and health, I asked Arnie specifically
at the NUBC meeting why functional status wasn’t on your list and he said that
there had been a lot of discussion about it and there was recognition that it
could be extremely valuable but it wasn’t, they didn’t feel that it was
standardized at this point and so it would be more difficult to implement and
they were looking more at things that could be implemented now. But I think
actually the recommendation here is to do the kind of work that would be
necessary to develop standardized measures and so my sense of the way he
responded was that the consumer disclosure project would support that, but I
don’t want to put words in your mouth but that was the dialogue I had with him
on that.

MR. QUERON: Well more is better from a purchaser perspective but I’m trying
to be realistic with you in terms of does our reach exceed our grasp,
sometimes, and I have great appreciation and respect for being responsible and
using what you ask for. And I’m not persuaded at this point that functional
status is something that’s as actionable as the other datasets that we’re
putting our emphasis on.

MR. REYNOLDS: It feels like Groundhog Day, yesterday I left North Carolina
and walked out of the building at 6:00 after having presentations on Bridges to
Excellence and Leapfrog. If I wake up tomorrow morning and you’re back I will
be concerned.

I tend to focus on how the health care system operates and that’s part of
why we’re on the standards situation and I’ve drawn myself a picture here,
which I won’t bore you with, but the interesting thing is that if you think of
the way information flows right now and you think of the way we work you have a
care setting and then standard transactions go to some kind of a payer, whether
it’s a private payer, Medicare, Medicaid, or sometimes a third party for some
of you large companies. And you pretty much call a lot of the shots, it’s not
the standard underwritten situation.

And then out of each of those payers drops many constituencies, United,
Aetna, Signa, us, have a lot of employers and a lot of individuals, whereas
Medicare has a very structured population, Medicaid has a structured population
and so on. When you think of quality, quality to me doesn’t seem to focus on
maybe that singular chain anymore, the normal drop down, because a hospital is
a hospital is a hospital and it sees all its patients and the quality of that
should be based on its whole population as well as a doctor, not just whether
if in North Carolina we’re X percent of the business then I only get a look at
X percent of their patients.

So when you think of, and obviously one of the things we have to address in
the Standards Committee in helping this particular committee, is when we look
at standards you can look at the claim form, or you can look at the attachment,
or you can look at something else, and if quality truly does supercede the true
financial payment process I would love to know from you if you see quality more
of a population situation, not who they have health care, and we’ve got a
significant other population called the uninsured that is growing and changing
and so on.

So I struggle trying to sit here as a member of the Standards Committee and
get a sense of what your model is but it seems to be saying that you want to
step clear back from that and I know some of you as employers literally go work
directly with care settings, so that kind of takes part of that structure out
of play. So I really would like before you get going, I really would like a
sense of where you see this data flowing to and is there one, like you think an
NCQA, is there one, I’m just using them as an example, but using them as who
decides quality and all the data flows to them or flows to somebody else versus
having a financial transaction that pops down and then somebody’s looking at a
very small subset which is how we’re kind of organized now as an industry. So I
would love just some comment so I can feel a little better about where this
might fall out.

DR. HAYWOOD: I’ll start it off because there’s a lot of layers there to say
the least, it’s kind of an intriguing question. Let me up front probably state
one of my biases which I heard what you said which is kind of population based
that made me think of the kind of a traditional public health approach. Which a
lot of times ignores the way you started the conversation which is how we
actually finance health care, there’s always this disconnect between what a
traditional public health approach is and the reality of the marketplace. And
so I think what you’re hearing today is kind of a realities of the marketplace
type of approach to actually getting at what you ultimately want to get which
is improved quality across that particular population.

Now as to whether or not there needs to be one data repository is a
separate issue to me then whether one individual institution can actually
define quality because I think there are quality measures but then there’s
value propositions associated with those quality measures that I don’t think
we’ll ever have consensus around everyone coming to an agreement as to which of
those particular measures they find more valuable then others irrespective of
who that arbiter might be as to qualities are if you will.

We are currently in the process at NCQA of building out the QIO data
warehouse and we are trying to build that out in a way, at least the way that
we’re envisioning that, in a way that should actually allow for some, the way
we’re looking at it to be honest I think a lot of people are trying to get out
of the data collection industry, there’s derivatives that are financially
desirable in the marketplace but the actual primary data collection, it’s a
question if that’s a declining industry as far as a financial model. And so I
think the government is going to at some point take that on anyway and so we’re
already starting to build out a kind of the QIO data warehouse on the hospital
side and on the physician side to actually be able to be somewhat of a data
repository where you can have that information put into it.

The thought is, in a very simplistic way, is that if then individual
practitioners or providers want that information disclosed to whoever then they
can actually have that disclosed so that individual entity, kind of giving the
consent of that entity can have the information kind of in the way that they
actually value that information.

So that’s a little different then actual passing judgment on whether or not
we’re defining quality but part of that is that you’re going to buy, virtually
by standardizing what information is coming in you are apriory(?) making a
determination as to which measures you actually think should be kind of part of
that decision making process so I think that portion is occurring now. And
that’s occurring kind of collectively because with the NCQA, with AMA on the
ambulatory side as well, with CMS, with Joint Commission, and kind of through
that NQF process as well as all the specialty societies that play a role, I
think we’re having that now starting in June where you’re getting a consensus
around what the actual spectrum will be as far as the measurement set and then
you’ll continue to have people argue whether or not they prefer within that
measurement set they like on measure, they value that more then the other.

So I think that’s at least kind of my view of where the current state is, I
don’t think we’ll actually get to the point of actually having one quality czar
kind of dictating what should be considered quality but I think we will get to
a consensus and NQF is the place to where you actually go to actually make
determination as to what those measurement sets should be.

MR. REYNOLDS: I didn’t mean necessarily that one person would decide, but
if it is in one place then anyone can go out, philosophical, I’m not
recommending that, I’m talking, you go out, access that data, and decide what
they decide, and then you go back to free enterprise and like you said value —

MS. TITLOW: I want to answer from the purchaser, purchaser side which is
we’re working as hard as we can to standardize what we asked for so all the
Leapfrog purchasers ask for the same thing and we are really working hard to
get our health plans that we contract with to ask for the same quality measures
and that is very difficult because that is, you’re both with health plans, you
know that is a differentiator in the market. So we get a lot of push back for
that question as well but we are, that is we would much prefer to have the
standardized questions and we think that this process will help that occur and
this is a step, a very important step in that process but I do want to
reinforce for you that that is absolutely our goal is to have as little noise,
as little variation, as possible out there.

MS. DILORENZO: I would agree with that and purchases is a little bit
different because it’s on the physician side but I think there’s somewhere in
between that we’re trying to strive to, we’re asking the health plans to use
national standards, third party objective standards, that’s what the physicians
will, along with, not to say that is an exclusive set, the only set, but at
least the health plans and the purchasers would all be asking for the same
thing because again, that’s part of the barrier we’re finding is they’re asked
to report on so many different performance measures that at this point it’s
like which one do I pick here. So there needs to be some consensus with littler
levels of variation.

MR. QUERON: Harry, my only quick contribution in response to your question
is the Consumer/Purchaser Disclosure Group is in the process of formulating
what we refer to as a set of ground rules that could be used by purchasers, by
payers, by others, in terms of what measures should we be adopting and
requiring for public reporting, for pay for performance and for other purposes
And if you’re interested in seeing that we’d be happy to share it with you
because I think it gets to the heart of your question, how can we forge more
standardization around a common set of measures for not only the burden
considerations but to enhance consumer and purchaser use of information. So if
you’re interested in seeing that we hope to use the bully pulpit of our broad
base of participating organizations to force more standardization along the
times of what we’re advocating.

MR. HUNGATE: That would be helpful.

DR. HOLES: We brought up the topic of risk adjustments and you responded in
terms of your recommendation saying that adding a particular data element would
support risk adjustment. But as we all know risk adjustment is a very murky
issue and I wonder when you speak of aiding risk adjustment for the purpose of
quality reporting or pay for performance do you have a particular risk
adjustment methodology in mind, are there differences in terms of a risk
adjustment methodology, whether you’re talking about outpatient reporting,
inpatient reporting, and would they ever come together. And perhaps Trend you
might be in a good position too to talk about what CMS has done with respect to
trying to actually take data elements and put them into a methodology with
which to risk adjust.

MS. TITLOW: For the measures that Leapfrog collects we look to risk
adjustment that’s done by the states for the publicly available state
information so they have their own process of risk adjustments and we just
defer to the four states that do that for those particular, when we take that
information, we also, we use STS, the Society for Thoracic Surgery, and what’s
the other one, thank you, ACC, somebody who knows more then me, ACC’s
recommendations, so we are deferring to the societies for their greater
knowledge. So no, we haven’t, we’re not going to develop our independent risk
adjustment methodology but we’re also working with JCAHO for IPS, for the

DR. HOLMES: Would it also be necessary that just as we have common
standardized data elements to have a common standardized risk adjustment

MS. TITLOW: And if anybody could identify for us an authority that could
help to develop that then we would be happy to look to them.

DR. HAYWOOD: I would have to talk to some kind of local CMS experts about
whether or not standardization of risk adjustment, what I mean by, I think
within individual settings I don’t have any particular issue, what I was trying
to get a sense of was whether or not it actually allowed us to do it across all
our different settings, which is a different issue to me. So that I don’t know
if you were asking kind of standardization for instance around particular risk
adjustment, hospital and mortality based, kind of getting some standardization
around that risk adjustment model, versus what we may do kind of in our
classification condition schemes that we’re doing on the managed care plus
side, whether or not it’d be the same kind of risk adjustment model once we
switched to home health and the long term care.

So I guess I would have to talk to some of the experts, Jared maybe later
on can comment to you as to whether or not it actually is that something that
would actually be valuable to do across all those settings or whether or not we
would continue to have to use different methodologies with different variables
across those settings.

DR. COHN: I actually had a question and a comment, Cathy is it about this
issue or something else?

MS. COLTIN: It’s about the issue that —

DR. COHN: Okay, so your last issue. Let me just make a comment about risk
adjustment and then ask a question and then Cathy can go back to the previous

Actually risk adjustment is sort of interesting since obviously CMS has a
major payment methodology which includes both Medicare Advantage as well as all
the Part B payments, it’s a Medicare model, it’s a payment model, not a risk
adjustment model for identifying other aspects of quality. I don’t know if CMS
is doing it, some sort of standard in my book.

Now having said all that let me get back to a more concrete issue. We’ve
been obviously talking about a number of these recommendations and I guess I
just need to understand from all of your perspectives sort of the frequency of
collection of some of these data elements. And obviously for example we’re
talking about lab results and vital signs and objective data as well as times
and dates for admissions and procedures.

And I guess I just need to understand is that something, are you
anticipating in your own vision of, I’m assuming that you even like these
things, would you be getting lab results on every admission, on every member,
every time you have a lab test? Or every time they have an X test done? Same
thing about vital signs, every time a patient is seen in the clinic are you
expecting blood pressures to be submitted with it every single time that they
have a visit?

The same point about the, I guess these times and dates of admission and
procedures, the only thing that I really ever heard referenced about those were
time to get in the ER until they go thrombolytics and what time they got IV
antibiotics before they had a surgery. Is there something more generalizable
here about all these things or are these, and the reason I’m asking is is
because if it’s indeed a generalizable every time a patient is seen you’d want
to put it into a claims transaction. But if it’s sort of an unusual thing that
you only get from time to time you might want to use for example something like
a claims attachment to really send out along the specific data related to that
very unusual circumstance. So can you give me a vision of what you think about
all this?

MS. DILORENZO: Actually it’s a snapshot of a period of time, for at least
the NCQA measures it’s a snapshot via claims database, claims information, so
it’s not ongoing, it’s a one time application, a snapshot of where their
patient population is and then the recognition that’s good for three years. So
it’s not an ongoing submission of data to NCQA, it’s an application, you do
chart extraction, you fill out, you provide evidence that you’ve achieved these
outcomes with your patients, you have to pull a sample of patient populations

DR. COHN: So you wouldn’t want the blood pressure on every visit.


MS. TITLOW: I want the blood pressure on every visit, because from my
perspective I think that Chris made an extremely important point which is that
we are at a very early stage of figuring out incentives and what works and what
doesn’t work, we’re also at an early stage of developing standardized quality
measures. So if we, the broader amount of information that we can collect the
better the quality data is going to be and the more flexible we’re going to be
able to be in the future. So now if we collect, my assumption would be that
it’s important from a quality, from an actual quality perspective to know your
patient’s blood pressure every time they come in and that should be something
that on a lot of check off lists for various conditions is checked every time
and that would be a quality measure. So I would want to know that in my process
measures and that we had process measures for CABGs(?), we have process
measures for diabetes, etc., etc., the question we have right now is do you
check blood pressure every time it comes in and if you saw high blood pressure
did you do something automatically to correct it. So if it was something that
was only collected once a year or once every whenever then I don’t think that
would count as quality from our perspective.

MR. QUERON: Our advocacy is to make these added fields mandatory on the UB,
I guess what would be the UB-04, so it would be submitted every time there was
a hospital discharge for these additional datasets, data fields.

DR. EDINGER: I wanted to just ask as you’re sitting there now about the
issue like lab tests, like lab results, if you did it by four different methods
in the same hospital and had that recently happen to me at Bethesda Navy, you
may see a trend in something that isn’t real because the methodology is
different, the techniques for doing the test is different and the normal ranges
are different. If you capture every lab result every day and then look at them
you may get a misleading picture of what’s going on with the patient without
additional information about how it was done or the methodology that was done.
For example, let’s take a glucose, it was done by different methods, you might
get a different, it may be normal but it may be in a different range which may
be abnormal in a different methodology.

DR. COHN: You’re asking that question?

DR. EDINGER: Yeah, I’m asking if you capture that data, just the numeric
data, is that going to cause a problem in doing an analysis for looking at
patient safety issues, what day you see abnormal results may become a patient
safety issue which may not actually exist.

DR. COHN: Can I try to answer that one? This is something that obviously we
could ask the LOINC person but I actually understood that that was part of the
LOINC capability, it got specific enough that it actually could deal with
variations in how the test was run.

DR. EDINGER: Well, also somebody has to look at that.

DR. COHN: Yes.

DR. EDINGER: And understand that that’s an issue, that’s the other half of

DR. COHN: Yes.

DR. CARR: But I think glucose is pretty consistent. But my question is, but
it does sort of raise the issue because the kinds of things that we’re looking
at, many of them are pretty standard but your point is well taken.

But I’m still trying to envision, if you have a claim for an admission you
might have had 50 glucoses so I guess that would be a different scenario, and
knowing the glucose control in the hospital is a very, very important quality
indicator but it would be unwieldy to have on the claim form.

MR. QUERON: Getting to this level of granularity is important, in
submission that we made to the NUBC we’re looking for lab values at admission,
not each and every lab procedure that is performed during the course of a stay,
we’re looking for vital signs at admission, we’re look for DNR orders at
admission or within 24 hours of admission, so that there are limitations on the
burden, not that this wouldn’t be a burden to capture but that it is precise at
the time of admission.

DR. CARR: Just to respond to that, I think in addition to trying to
standardize the data elements we actually need to standardize the questions
because you’re asking was the patient high risk when they came in and did the
outline line up, and you’re asking was blood pressure properly controlled over
time. And someone else is asking, Jessica was asking if you have a high blood
pressure on a snapshot did it trigger a response. And I think that it’s just as
important to standardize the questions as it is the data elements.

MS. TITLOW: And again, this is what we were asked to respond to so at the
moment we only have, enough of that’s inappropriate because that’s what we want
to be able to respond to but I think that we are presenting different answers
depending on where we’re looking at it in the continuum of our process. I’m
trying to think in the very long term what would really be able to help us get
down to the quality measures and I think Chris’ information very well informed
by experts is what is the fastest thing that we can get done mostly likely to
be able to pass through and then we could at least get a starting point and I
think both places are valid places to be, it just depends on, some of it’s
politically where do we think it’s most useful to go. And I’m one for
collecting lots of, long for a lot of flexibility —

MS. COLTIN: I wanted to respond to the question that Mr. Reynolds raised
about health plans getting only a slice of the data and it’s something I’m very
familiar with and a problem that in our health plan we struggle with, there’s
only a percentage of our provider network on whom we have petitioned numbers of
patients to base any measures and that in order to evaluate the quality of the
entire provider network we need to pool our data with other health plans in
order to maximize the value of that data. And if Dr. Millstein were here I’m
sure he would be talking about initiatives that he’s involved with, the Care
Focus Purchasing Initiative where employers are asking the health plans to pool
the claims data to create quality measures from these pooled datasets. These
are the same administrative datasets that the proposal would effect in terms of
making available new data elements to permit the creation of more robust
measures of outcome and other measures of quality.

So that’s one perspective on that, another project that’s similar that Bob
Hungate is involved with is one that’s going on in Massachusetts that’s
sponsored by the Group Insurance Commission which is also pooling claims data
to develop measures of provider efficiency and as part of that project some of
these same data elements can be used for risk adjustment, not simply for
outcomes measurement. So I did want to bring that perspective in that
increasingly plans are being asked not just to take these data in, to process
them and adjudicate claims, but to analyze them and take actions on them
themselves and to ship them out as part of collaborative quality improvement
efforts. So I see the data stream as really in and out with a lot of digestion
in the middle.

MR. HUNGATE: Okay, where are we question wise, have I covered —

DR. EDINGER: I have a question since Cathy’s here, Trent says that if we do
pay for performance one of the things we may do is get a problem with
disparities because basically some of the rich may get richer, the poor may get
poorer, and I guess part of the issue is how you brought the data that show
there’s a performance difference, are there certain data elements we might need
that might help in, well, you can’t solve the societal problem but at least in
generating some information that we may not just pay for providing unless you
got a CPOE you get ten percent more or because you have higher risk patients,
but are there other data elements that might help us in the pay for performance
model or especially down the road, don’t pay for non performance models, that
may actually help eliminate some of these disparities issues, at least let us
deal with them in the context of this kind of a free market or pay for
performance model?

MS. COLTIN: Well, I think over the four years of testimony that was taken
in formulating these candidate recommendations we heard a lot of ideas. The
think the candidate recommendations reflect the preponderance of interest in
particular data elements to move this along. It’s not that there weren’t some
other ideas suggested along the way. I think that in terms of the specific
issue around creating disparities and trying to protect against that there are
other candidate recommendations contained in the full report, this hearing is
only looking at the first eight but there are recommendations around the
collection of race and ethnicity data which would enable the transparency of
these measures along various racial and ethnic lines as well so that we could
identify whether or not disparities were increasing as a result of these kinds
of strategies and could realign incentives to make certain that that doesn’t
happen. So I think while we’re talking today about only the first stage we need
to keep in mind that they’re part of a set of recommendations.

MR. HUNGATE: I have a personal need to take a break and I’d like to thank
the speakers and applaud your participation and content, it’s very helpful. And
we would especially like to see what UBC submission from the group is.

[Whereupon at 11:40 a.m. the meeting was adjourned, to reconvene at 1:05
p.m. the same afternoon, Thursday, June 24, 2004.]

O N [1:05 p.m.]

MR. HUNGATE: We’re missing a few people as yet but they’ll get here. I know
that such quality performers as we have for this afternoon will need no
reminding to stay within their 15 minute desired limit, I’m going to switch the
order a little bit and let Jerod Loeb go first from the Joint Commission since
he has some deadlines on the other side. In the panel this morning the comments
were very succinct, the testimony was very helpful, and the discussion
afterward was even more helpful so we’ll hold all questions to after each of
you has done your thing and then we’ll go for questions.

Why don’t we go around and do a quick introductions so everybody knows who
everybody is.

DR. HOLMES: I’m Julia Holmes and I work for the National Center for Health
Statistics and one of my primary tasks here has been to work with wonderful
people like Ed Kelley over at AHRQ on the National Health Care Quality Report
and Ernest Moy(?) on the National Health Care Disparities Report working to
provide probably a third of the measures that go into the report.

MS. GREENBERG: I’m Marjorie Greenberg, also here at NCHS, CDC, and
executive secretary to the committee and I want to thank you for coming and
welcome you to NCHS.

MS. POKER: Hi, I’m Anna Poker from AHRQ and I also have the pleasure of
working with Ed Kelley on the QR report and on the DR report with Dr. Moy. And
I also am part of the patient safety team at AHRQ.

MR. HUNGATE: Bob Hungate, Physician Patient Partnerships for Health,
chairman of the Quality Workgroup, member NCVHS.

MS. HANDRICH: Hi, I’m Peggy Handrich, I’m a member of the committee and a
member of the Quality Workgroup and I come from the Wisconsin Department of
Health and Family Services and have a special interest in Medicaid and the
state collection of health statistics.

DR. CARR: I’m Justine Carr, I’m a physician at Beth Israel Deaconess
Medical Center and Health Care Quality and I’m a member of the committee.

MR. REYNOLDS: Harry Reynolds, Blue Cross and Blue Shield of North Carolina
and a member of the committee.

DR. JANES: Gail Janes, CDC Atlanta, actually Washington, D.C., and staff.

DR. EDINGER: Stan Edinger, AHRQ, I also had the pleasure of working with Ed
Kelley although he won’t give me an autographed copy of his book unless I agree
to give him $50 dollars for it but I’m still waiting. I’m also staff to the

DR. LOEB: I’m Jerod Loeb, executive vice president for research at the
Joint Commission, I’ve never had the pleasure of working with Ed but I’ve been
on panels with him and thanks to Ed and Peggy for letting me go first.

MS. O’KANE: Peggy O’Kane, president of NCQA, I look forward to dialoguing
with you.

DR. KELLEY: My name is Ed Kelly and I work with Ed Kelly at the, I work at
Agency for Healthcare Research and Quality, I’m director for the National
Health Care Quality Report and team leader for the National Health Care Quality
and Disparities Report.

Agenda Item: Quality Measurement Organizations – Panel 2
– Dr. Loeb

DR. LOEB: Any of you that know me, and there’s a fair number in this room,
I almost never speak without slides so this is a whole new experience for me,
but as I sat and thought about what to say it occurred to me that this was
dense stuff and it was going to be really hard to do this in the context of a
short timeframe so I actually scripted myself so that I could hopefully keep to
the amount of time and I have given Debbie a copy of my remarks so you do have
it for the record.

Certainly I realize that you’ve asked us all to limit our remarks
specifically to certain sections of the report but I beg your indulgence for
just a couple of seconds because I do have to make a global comment having read
the full report and that is I really applaud on behalf of the Joint Commission
the basic message in this report, the basic message being the importance of
standardization in the measurement field. It cries out to standardization and
I’m talking about this from the perspective of data collection as well as data
reporting and frankly I don’t think this is necessarily intuitive obvious in a
pluralistic society where the mantra seems to be let 1,000 flowers bloom. So it
really is sort of antithetical to our society but at the same time I think it’s
abundantly clear that absent the standardization that you’re trying to move
toward our ability to establish appropriate benchmarks, to track and hopefully
improve performance over time, pay differentially for performance and compare
providers is difficult, perhaps even impossible.

And inherently woven into this complex fabric as I look upon this is the
importance of data collection becoming a byproduct of the care delivery process
rather then as an iterative activity separated from health care delivery. And
that’s really what we have here today, we have health care delivery here, and
for the most part data collection here, at least within the context of most of
our provider organizations. And the ven diagram is really never connected and
frankly I think it is only with a nationally standardized electronic health
record with consensus generated standards that we’ll ever have any hope of

I know I didn’t have to say all this but I feel better having said it.

To begin I speak from the perspective of the nation’s oldest and largest
health care accrediting body, a little bit bigger then Peggy’s but I don’t want
to say anything inappropriate, but more importantly I want to be clear that my
responses to the specific questions that you’ve asked really represent a
synthesis of our staff analysis of the report and its recommendations. And we
did so in the context of trying to understand how it would impact our existing
business processes, how it would impact the business processes associated with
the vendors with whom we have relationships to transmit data to us and there’s
about 130 of those vendors out there, as well as the provider organizations and
there’s about 16,000 of those that we currently accredit. So clearly the
measurement train has left the station based on what I just said and I speak
here about the enormous but what I view as clearly fragmented measurement
infrastructure that’s in place today in health care facilities.

Now what I will try to do is to use specific examples from our existing as
well as our evolving measure sets to speak to the questions that you asked
based on the candidate recommendations, firstly in terms of business case. The
question of legitimate business needs I think is clear, there is a legitimate
business need for just about all the recommendations that you have made in the
report. Using our ICU measure set as an example we have many physiological
variables in the set that are needed in calculating the risk models, at the
moment it uses Apache-3 methodology and those of you that are familiar with
Apache-3 know that there’s a host of physiological variables associated with
that. But in addition there are other variables such as the results of
laboratory tests that we clearly believe that your first candidate
recommendation relative to laboratory values is really very important in terms
of standardizing it. But one of the questions that we have is how much of the
laboratory values, that’s not really very clear in the report as we see it.

Outside of traditional laboratory values, physiological data, and we heard
this in the testimony this morning, blood pressure, heart rate, respiratory
rate, and so on, all are critical in the context for example of the ICU set.
Since most of these measures require data collection beginning on the first day
of admission to the ICU for our evolving set, but linkage of the test results
or even vital signs to data collection we think is absolutely essential.

Comorbid conditions are important risk adjustment variables in the ICU
measure sets so we certainly agree with the recommendation relative to
secondary admission flags, totally supportive of that. And just to go further
as another example in our existing pneumonia set, which currently has about
2100 hospitals collecting data, this is an even more important issue.

We believe the recommendation respecting required reporting of operating
physician would bring significant added value to performance measurement both
in terms of internal QI as well as in terms of external reporting. And as I’m
sure you’re all aware there is a lot of discussion in the context of the boards
of medicine today relative to maintenance of certification. And if we collect
the data once using a standardized set of variables and stream the data where
it’s needed, to an accrediting body perhaps for decision making, to a board
perhaps for decision making, all with one data collection it would
significantly advance the field and probably reduce the push back.

As to the issue of dates and times for admission and procedures, this is
already integral in our process. Timeliness of care is addressed both in our
pneumonia set and it relates there to blood cultures and antibiotic timing,
it’s addressed in our acute MI set, it relates to thrombolytic timing,
aspirant(?) arrival, etc., but these measure calculations use arrival date and
time rather then admission date and time. And we did that purposefully because
much of the care provided prior to admission in areas such as pre-hospital EMS
and in ED care, that’s an issue, so we’re a little concerned with the way you
currently have this worded. And in many ways many of the entities in the field
are really moving away from admission date and moving toward arrival time,
admission time to an arrival time. So we would urge you to reexamine this in
the light of some of the existing processes and the evolving science before you
make your final recommendation.

In terms of pay for performance the Joint Commission uses data elements to
construct measures and not data elements themselves as a unique dataset, that’s
a fine nuance that I think is important to understand. Thus if some of the
specific data elements that are being recommended in your report could be
derived administratively and not require chart abstraction everybody’s going to
benefit. I think you heard that in the context of the discussion earlier this
morning, from Chris, from Trent, from others as well, and this benefit will
accrue to health care organizations, it will accrue to individual providers, it
will accrue to purchasers as well. It’s really a transformational change moving
from chart abstraction such that these data elements become part of the
administrative dataset.

Risk adjustment, a very complex topic, you discussed it for a while this
morning, there’s clearly many patient level data elements necessary to
appropriately risk adjust outcome based performance measures. And I want to
emphasize to you I’m talking about outcome based measures and not process
measures because we traditionally do not risk adjust process measures.

The ICU measure set which is currently in testing requires nine separate
lab values for inclusion in the risk model and these include such things as
blood gasses, hematocrit, albumin, BUN(?), creatinine glucose, sodium,
bilirubin and white counts, so the committee’s first recommendation respecting
lab data values once again becomes very critical for us although the actual
usefulness is ultimately going to depend on which values, which values you
include in the final set that AHRQ and the National Quality Forum are charged
with developing.

Similarly the recommendations made respecting vital signs, secondary
admission, diagnosis flags, and the admission dates and times so for the
arrival times all have significant bearing upon our risk adjustment
methodologies and standardizing data elements here is going to go a long way
toward reducing some of the push back and some of the controversy that exists
frankly with respect to risk adjustment techniques today.

Data quality, personally I believe data quality should under gird
everything we’re doing in performance measurement but I think that there is, at
least from the pulpit that I occupy, quite a bit of variability in terms of
data quality today. You talked about it earlier with respect to the MDS set,
it’s true in the hospital dataset, it’s true in the home care dataset, it’s
true all over.

Your fourth questions pertains to whether the suggested mode of data
collection would in fact provide data of acceptable quality. It’s clear that
the data being collected are of variable quality but we certainly believe that
your recommendations are sound and will result in data of acceptable quality,
perhaps even beyond acceptable to good quality data. In particular with respect
to our surgical infection prevention measures and our children’s asthma
measures which are now in testing the recommendations once again pertinent to
test results probably require a bit more specificity then you have already
articulated. Test result data should include dates and times and be able to be
identified as to whether it’s inpatient or outpatient.

Value, that’s the mantra that everyone is talking about in performance
measurement today. If standardization stimulates a migration of data elements
from those requiring chart abstraction to the administrative dataset we believe
the recommendations in the report will significantly enhance the value to
everybody concerned. But I also want to at the same time exert a bit of a
caution here, and that is it’s important not to shift general thinking toward
only measuring those variables that can be derived from administrative datasets
rather then from medical record abstraction, I’d like to see a balance here,
certainly in favor of administratively derived data elements but let’s not
discount the fact that there are important variables particularly with respect
to risk adjustments that can only be derived at least today from the medical

And as I think about this there’s sort of three schools of thought about
this in measurement, the first says what do I have, what data do I have and
what can I measure. The second school of thought says what can I measure and
can I find reliable and valid data. And the third school of thought which is
the one that I would hope you would be thinking about is do I ask the question
first, what’s the right thing to measure, can I find reliable data, and what
about the cost benefit relationship, do the costs outweigh the benefits or do
the benefits outweigh the costs. It’s sort of that three part piece to this
complex question that we think is really important.

So in summary I think it’s fair to say that we strongly support the basic
tenets that you are presenting as your candidate recommendations in all cases.
I didn’t speak about functional health data and I know that was an issue this
morning but certainly that’s an important area and we feel very strongly
certainly outside the hospital accreditation program, it’s important in the
home care, it’s important in long term care, so we support that as well.

We believe additional standardization around data collection and reporting
is absolutely essential although it’s abundantly clear that there still are
significant challenges as you begin to consider the implementation strategies.

So thank you for listening and hopefully we can engage in some conversation
before I need to run.

Agenda Item: Quality Measurement Organizations – Panel 2
– Peggy O’Kane

MS. O’KANE: Well Jerod is going to feel really good that he did a written
statement instead of slides, I don’t know, we’ll see how this goes. You’ll see
there’s a lot of wordiness but I think I can get the messages across. We also
think this is a real contribution, this report, and I’m going to speak at a
fairly high level I think somewhat like Jerod did, but I also would invite you
to talk to us because we have a great performance measurement staff that’s been
actively collecting these data and we can talk to you a lot about the kinds of
things that are really important to measure that we can’t measure today because
of the lack of availability of data.

So with that I’d just thought I’d, I think most of you know who we are but
let me just go over it briefly, we’re a private non-profit, much smaller then
the Joint Commission, maybe a little more ornery then the Joint Commission, we
measure and report, we’re very proud of the role we’ve played in the
advancement of performance measurement in the whole accountability world. And
we I think have done this by uniting diverse groups around the common goal of
improving quality through measurement and public reporting. Our orf(?) reflect
the consensus of stakeholders, sometimes very difficult to achieve but always
achieved. We produce information that enables value based purchasing and we put
a lot of emphasis on pay for performance, I’ll talk about that.

We produce information that can be used by the public for choice and by
providers for quality improvement. And our public reporting of HEDIS clinical
information and patient satisfaction using CAHPS covers over 300 health plans
with more then 70 million covered lives. And we seek to align with key
regulatory requirements because there’s no question that this activity is very
expensive and creates a lot of, I mean when people see their results also it
takes a lot of effort to improve them.

So HEDIS I think, we’re not actually the inventors of HEDIS, it came to
root I guess in NCQA in 1992. It’s a very robust set today of over 60 process
and outcomes. CAHPS came to us through AHRQ and we appreciate our relationship
with AHRQ and I think it’s an example of the kind of public/private partnership
that we all ought to be looking for. We’re very proud of our precise
specification of measures and auditing of results and we have independent
auditors, we audit the auditors. On the satisfaction survey we certify survey
vendors and so forth. So in the course of just trying to have a system that
could withstand the kinds of questions that would inevitably be raised we’ve
created really a whole soup to nuts accountability system.

It’s used to evaluate quality of care for commercially insured and Medicaid
and Medicare enrollees, we have now adapted it for measurement of performance
by medical groups such as physician offices in many circumstances. It’s
required by many public and private sector purchasers and it is, derives about
30 percent of our accreditation scores these days and that’s something that
continues to move up as we build more measures.

I think there’s no question that transparency really works, I’ll just give
you a few numbers from our State of Health Care Report last year. Every year I
say to my staff I hope the plan has improved again and again 30 out of 33
clinical measures showed improvement last year. There were very strong gains on
cardiac measures which were relatively more recent, and if you look back over
the history of this reporting the average improvement is over 50 percent and
it’s really remarkable and it speaks very well of the people in the plans who
have taken this on.

We believe there’s a need to expand this agenda especially today as we move
into kind of a, how do I want to say this politely, 1,000 flowers are blooming
in the health plan world and we need to be able to make comparisons across
what’s the value add of this type of a product versus that one. We’re pleased
that we’re going to have information from hospitals coming out, that will be
useful I think for the plans as well as the consumers. And physicians I think
there’s a lot of very good activity going on so reason to feel optimistic that
this agenda is really going to move.

We believe very strongly that we need payment reform as a way of really
moving this agenda, that means not only the kind of things that we’re involved
with today which is rewarding superior performance, but I think there are many,
many ways in which our current payment systems reward bad performance. We need
to go beyond fee for service, we need to go beyond unadjusted DRGs, and when
we’re paying for complications that are caused by a provider I think we need to
think about that.

I think the resource base relative value scale was useful in its time, I
think we are at a point when we need to really think about what is the value
that we gained from particular health services and think about health as the
true north for value. And all of this of course requires meaningful, sound, and
feasible performance measures.

Just to talk a little bit about some of the pay for performance activities
that we’re involved with, in California there is a project involving seven
health plans and 230 medical groups. It’s $100 million dollars on the table
that will be distributed to very superior performers. I think the level of
enthusiasm both on the sides of the medical groups and the part of the plans
has been remarkable and it really, this was something that started out very
shaky and there’s been a lot of pushing and shoving and now everybody’s seems
to feel good about it. Maybe the ones that don’t get the money won’t be so

The other one I think that you probably heard about is Bridges to
Excellence, which we do collaboratively with GE and other employers in four
different cities, Louisville, Cincinnati, Boston, and the latest entry is in
Albany Schenectady. These are built on our recognition programs for physicians,
we have one for diabetes, one for heart care, and one for systems in the office
including electronic medical records, registries, education and so forth,
patient education. We’re also involved in Minneapolis where again there are
groups, medical groups, which by the way facilitate this enormously I think you
know. Also I think a lot of interest there at reporting at the group level and
also in pay for quality. And then our recognition programs are being used by
others but those are the most prominent examples. We also, you mentioned
Medicaid, many Medicaid programs are now beginning to look at pay for quality
and we think that’s a wonderful thing.

Now to go to the report, as I said we were very impressed with the report
and really agree very much with the thrust. The lack of accessible and
inexpensive data, in other words not from chart review or survey, is a major
barrier to measurement and I just want to reinforce that we are to some extent
the drunks under the lamppost because just from a practical point of view there
are many things that we cannot measure because it’s too expensive or too

As we try to move to the hospital and physician level of measurement chart
review becomes prohibitively expensive and particularly if you think about very
sick patients where they’re seeing multiple specialists there are big questions
about the whole framework for accountability in that kind of a scenario. But
just figuring out whether somebody actually got something when you have
multiple physicians was very difficult.

Electronic medical records offer hope but they won’t be more useful then
paper records unless they are structured and coded for abstracting quality
purposes. So you can have just kind of free text in electronic medical records,
it’s really not very helpful. As administrative data will remain an important
source for quality information we think for years to come it’s reasonable to
improve this data source and so we applaud you for your practical approach to

In addition to enhancing the administrative data sources to enable the
measurement of particular care processes and outcomes it should be clear which
of these measures will be used for accountability, provider feedback,
surveillance, or research purposes since not all measures will be useful for
all purposes to the same extent. So overall while there’s some initial and
ongoing expense expanding current claims data far, you’ll get a much better
bang for the buck then trying to go around the current claims data, so again,
we agree.

So let’s go to the individual recommendations. We combined one and two,
create a mechanism for reporting selected inpatient and outpatient lab results,
vital signs, objective data measurements and standardized transactions. Lab
results and vital signs can represent important outcomes or risk adjustment
parameters. Their inclusion in routine performance measurement and pay for
performance activity will increase substantially with routine reporting through
administrative claims and the plans I think have really done a lot, there’s
been an enormous amount of resourcefulness with the current state of affairs
but you could just make this so much better if these recommendations are

Objective measures could be reported in the standardized fashion, we urge
you to concentrate on things that are likely to change as opposed to things
that are not likely to change. Some parameters only pertain to particular
clinical conditions or specific procedures, we urge you to clarify on relevant
coding forms or to suggest that it be clarified when added elements are
required to be coded during care episodes. Having electronic data on test
results and vital signs would significantly increase the ability to do clinical
measures and risk adjustments at a much lower cost per measure.

Number three, which is facilitate the reporting of a diagnosis modifier to
flag diagnoses that were present on admission on secondary diagnosis fields in
all inpatient claims transactions. Administrative transactional data provided
by hospitals about the type of care provided don’t lend themselves easily to
the calculation of multiple valid performance measures pertaining to inpatient
care. However more measures could be calculated with the addition of a few
routine data elements on medical claims. For example, calculating the rates of
adverse outcomes would be greatly enhanced if diagnoses present on admission
were routinely and validly coded, in other words you’d have some hope of
knowing what happens during the admission.

In addition the inclusion of standardized severity designations, for
example cancer stage which is hugely important, might aid in the calculation of
risk adjustment outcomes or measures of appropriateness. So we very much
support that recommendation.

Number four, modify the use instructions for the existing data element for
operating physician such that it is a required data element for the principal
inpatient procedure. I’m going to just summarize, we think this is a very
useful recommendation, we very much support it and it really does make the
whole accountability picture much clearer.

Let me move to number five, modify the requirements for reporting admission
date and time and selected procedure dates and times on institutional claims
transactions. The IOM identified timeliness as a key aim, we agree, little
information is available now on timeliness as Jerod also noted. Moreover in
order to improve the rendering of effective care that require the adherence to
certain time intervals in order to improve outcomes, like door to needle time,
date and time information is essential. Information about which time intervals
are most critical to adhere to in order to improve clinical outcomes or
minimize patient stress should be identified. And I just want to really urge
you to think about the patient worrying and stress as also an important outcome
and an important corollary of care. Dates and times of services would allow
measurement in an accurate and efficient manner of a major current gap in
performance measurement.

Number six, encourage payers to modify billing instructions to providers to
align procedure start and end dates with services included in selected global
procedure codes and standard HIPAA claims transactions. In general we think
this is a desirable and useful addition to administrative data.

For the set of practices where compelling evidence exists that care must be
initiated within certain time intervals to increase the likelihood of desired
outcomes it’s advisable to require the recording of start and end dates if such
practices are captured using global procedure codes. At the same time in order
to not overburden the system such capture should not be required for services
where time is not of the essence.

Number seven and eight, review the available options for coding patients’
functional status in EHRs and other clinical datasets and recommend standard
approaches. Conduct the research recommended by NCVHS in 2001 and CHI in 2003
as endorsed by NCVHS. Create a mechanism for reporting functional status codes
in a standardized transaction. There’s no doubt that functional status
information represents an important outcome for many care processes and patient
groups that should be captured in a standardized fashion. As with other
outcomes many factors contribute to functional status including patient
variables as well as care processes.

Interpretation of functional outcome data as an accountability measure can
be very complicated, including issues like what is the natural history of
disease and what’s the variation from the natural history. It can be
complicated if functional status prior to initiation of treatment in addition
to other important variables is not taken into account. So I think there’s a
framing issue there that’s really crucial even before the sort of nuts and
bolts of functional status measurement is taken on so I would urge you to
consider that.

So our conclusions, administrative data will remain critical for the
measurement of quality in health care for some time to come as alternative
widely implemented data capture processes have been slow to penetrate health
care delivery. As has been demonstrated with the HEDIS performance measurement
set administrative data can constitute an easily available, not so easily if
you ask many of the people, reliable and valid data source for the measurement
of many aspects of health care quality. Administrative data can further be
enhanced through the alteration of current data capture processes in coding and
billing practices, and one point I want to make is that when these data count
for payment, for accreditation, for whatever it is, for public reporting, the
data get better.

The rationale for the adoption of recommendations could further be enhanced
if it is clear which additional performance measures could be computed, and we
would be delighted to work with you on that and some of the experience we’ve
had, if it is determined how these measures relate to national, regional, or
other important health care quality goals and priorities, and if the health
benefit gained by improving the performance in areas for which these newly
deployable performance measures is quantified. So in other words making the
case for kind of gains that are sort of sitting on the table now I think is a
crucial part to advancing this agenda.

So thank you very much for your attention and look forward to the

Agenda Item: Quality Measurement Organizations Panel 2
– Dr. Kelley

DR. KELLEY: First of all I want to, I’ll say congratulations to the
committee on the report and I see Cathy here, congratulations on the report, I
know this was a long time in coming and a lot of work. I’ll echo what I had
said at the full committee meeting which is that AHRQ has enjoyed a real
symbiotic relationship with NCVHS in general and this committee in particular
supported quite a bit the development of the National Health Care Quality and
Disparities Reports and so for that we’re eternally grateful.

And also I would say congratulations on actually holding these hearings
because I think, I mean the report as Jerod and Peggy were saying could stand
on its own, I think it’s important for I’ll say folks like us but I mean the
folks who are in kind of the measurement and data world to assume that better
measurement and better data which is as Jerod was alluding to what this report
talks about is automatically a good thing. Not talking to our partners who come
from different perspectives doesn’t allow the full consideration of what are
the tradeoffs on this and I think some of the questions that you gave to us to
answer in terms of the value issue speak to that and I’m sure some of the other
panels will have slightly different perspectives on this.

But what I’d like to do, I didn’t have copies of the slides, mainly because
then I would have to put what follows is the opinion solely of Ed Kelley and
not of AHRQ in general, we’ve had a lot of input on this but what I’d like to
do is provide you sort of a little bit more formal official that Caroline who
has been out of town had a chance to see and sign off on as well as the written
input. So with that in mind maybe we can move along.

In terms of, I’d like to just provide a few background pieces and then go
through some of our input on the specific recommendations and one follow-up
item that we think the committee should take into account. Now the business
need for this which is how you had worded the first question to us, we took
literally in terms of considering where is AHRQ’s concern with this and
improvement of administrative data, it’s obviously coming from a research and
quality agency, something that AHRQ holds very dear to its heart, two prominent
areas and I don’t mean to imply that these are the only areas but to prominent
areas where we have done work on this are with AHRQ’s Health Care Costs and
Utilization Project and our quality indicators which have been developed to
assess quality of care at the hospital level using administrative data, and
they were intended for internal quality improvement efforts. And a lot of those
same indicators have been used in the National Health Care Quality Report and
National Health Care Disparities Report to publications, that as the committee
knows that we were tasked by Congress to produce.

I’ll just pause a minute as Jerod opened the door, I know you asked us to
focus on recommendations one through eight but we are sure looking forward to
the follow-up hearings on the rest of the recommendations because I think there
are a lot of other good pieces in there, particularly recommendation nine that
talks about disparities and the measurement of racial and ethnic data elements
and that’s, I mean we’ve had endless conversations about that. And as well as
number ten, standardizing surveys and how things like vaccination is measured
across different surveys.

The reports themselves, I won’t dwell on them too much but again, mandated
by Congress, and this is the language that gave them to us. The one to track
quality of care annually in a National Health Care Disparities Report to track
prevailing disparities in health care delivery as it relates to racial factors,
socioeconomic factors in priority populations. The two reports track quality of
care in the National Health Care Quality Report and quality and access issues
for those, along those particular lenses as I mentioned.

The areas that are tracked and for which this has the most relevance is in
the effectiveness area and that lists off the different areas, different
condition areas, and we talked about at the full committee hearing how we got
to those. There’s also the issue of timeliness and I’ll touch on that in a

Health Care Costs and Utilization Project, it’s the largest collection of
longitudinal hospital care data in the United States and there’s a family of
databases that make up each and it’s a partnership between AHRQ and the states
as some on the committee know. There’s the Nationwide Inpatient Sample and
that’s inpatient data from a national sample of over 1,000 hospitals. The State
Inpatient Databases, which is inpatient care in community hospitals in
participating states and this next years report, National Health Care Quality
Report, will report on about 30 states although membership is up quite a bit
from that currently. That accounts for approximately 80 percent of all U.S.
hospital discharges in total. And then there are two more derivative databases,
state ambulatory surgery databases on ambulatory care encounters, and on the
kids inpatient database for hospital inpatient stays for children.

Now if I can get then to the summary, sort of struck me, we all have the
same issue of what’s the best way to offer this information to you and honestly
I only did these slides because I knew Peggy and Jerod were going to do slides
and then Jerod let me down, so Peggy and I are both mad at him. But Peggy’s
slides offered a lot more information and what I’d like to do is kind of run, I
mean this is very summary so it’s very hard to know our logic behind it, and in
the hopes that you’ll think still waters run deep I’ll kind of go quickly
through this and give you the rationale for the thinking on this and then again
provide you with the actual written rationale.

Clearly it will be while the national dialogue on electronic medical
records has stepped up significantly in the past couple years and even
particularly just in the past six months we all anticipate that it will be
quite some time before that information will be readily available to us in
terms of quality analysis and quality improvement. There are a whole host of
reasons in terms of trying to rely on ICD-9-CM codes for clinical information
in terms of working on improving administrative data and trying to have
clinically important and specific quality improvement efforts. And even quality
monitoring efforts, ICD-9-CM has limited clinical depth, the severity of
conditions for things like congestive heart failure and diabetes are difficult
to ascertain.

Timing of conditions cannot be understood using current administrative
data, while renal failure may be coded it’s not clear if the patient may have
been admitted with the condition or if the condition originated during the
hospital stay. And of course hospitals differ in their depth of coding, this is
something that we’ve had difficulty with in the reporting at the state level
for the National Health Care Quality Report in terms of trying to determine an
equitable way of reporting some of our indicators, many of which depend on
coding which is differentially done for risk adjustment as well for exclusions

In terms of the data, the recommendations themselves, I’ve sort of grouped
them into three areas, the ones that we feel like offer, and this echoes a lot
of what Peggy and Jerod had said, improve data and better risk adjustment and
those are tackled in recommendations two, three, seven, and eight. The ones
that, skipping down, that offer some important process outcome links,
recommendations one, although we have some concerns about burden. And then
recommendations two, six, seven, and eight, and you’ll notice that no one list
is mutually exclusive, that’s designed to be potentially confusing, the value
for effort concerns I think we have some on recommendations two, six, seven and
eight. And what I’d like to do is maybe jump into some of the specific thoughts
on a subset of these measures, specifically recommendations three, four, and
five, and recommendation number one.

Now by listing in that previous slide recommendations in multiple places I
think it kind of indicates some of the ambiguity that we may feel over
recommendations that will make important improvements in data but obviously
there’s a real spectrum of the costs associated with it, it’s not just dollar
costs, associated with some of these recommendations. And we did try, although
it’s not sort of uniquely our area, to consider some of these.

On our overall feedback on this recommendation three will allow providers,
purchasers, and policy makers to distinguish complications from comorbidities
and enhance performance measurement and risk adjustment. And if this is
consistent with concerns expressed by stakeholders such as MedPack(?) and
others of why are we paying more for poor care.

Recommendation three could be a critical component in a pay for performance
environment by enhancing indicator definitions and related risk adjustment, and
by providing clarify around complications versus comorbidities in general. It
could be valuable to internal quality improvement efforts by providers,
comparative reporting efforts by organizations such as hospital associations
and state agencies, and the public reporting of provider performance. We felt
like this recommendation had high value and is as far as we know is already in
place in a couple of states already and would also provide specifically with
the National Health Care Quality and Disparities Report better ability to risk
adjust and analyze across some of the indicators that we’re looking at right
now at least only reporting at the national level and only reporting kind of
one point in time which is kind of difficult, it’s not really our mandate and
it’s not that informative.

Recommendations four and five we feel like could be listed as having a
strong business case mainly based on the resources required to implement in our
view, it should be relatively low cost and we should be able to have confidence
that the operating physician is the one who performed the principle procedure
and would help to expand so the physician is noted with each procedure. It will
assist in establishing accountability for procedure related outcomes, and we
felt that in terms of the value question that you asked us to address that this
is relatively high value.

Now on recommendation five and a little bit on number six, this tackles the
timeliness question and both Jerod and Peggy talked about this. Now timeliness
as I showed in an earlier slide timeliness is one of the four dimensions that’s
tracked in the National Health Care Quality and Disparities Report and is in
fact one of the least populated of the entire framework in terms of number of
measures. And in part I think that it’s because there’s been something of a
schism in what people, in the views that people have on timeliness, what exists
and what came up through a very lengthy consensus building process for the
National Health Care Quality Report on measures were timeliness measures were
patient’s perception of timeliness measures, about getting appointments at the
right time and those types of things.

While a lot of really good timeliness measures were clinically specific,
timeliness measures were kind of lumped into the effectiveness area and more
forward thinking organizations like NCQA, like JCAHO, may not think this way
but I think there’s a significant portion of the research and quality community
that don’t make that leap and link the two. And on our next National Health
Care Quality Report we will present in a timeliness section, not just the
patient perspective which unfortunately a lot of our readers, they go to the
effectiveness section, they read that and then they forget the rest of it, and
maybe the safety section, and they forget the rest of it because it’s all about
those patients and really what the heck do they know anyway.

And the discussion now that we’re having would vastly improve not just the
quality and disparities reports ability to track timeliness in a much more
clinically specific way but would also allow us to kind of take the first step
on linking some of those patients perspectives and provider perspectives on
what’s an extremely important dimension of quality.

Jerod is right on the admission and arrival point and we would, while AHRQ
has not been kind of in the forefront on any evidence reports specifically
looking at this and has generally taken its cue from the from the work that CMS
and JCAHO and other organizations have done in terms of designating these on
these particular measures, out measure in the National Health Care Quality
Report right now on pneumonia in terms of time to antibiotics is arrival, right
now although we’re examining it for AMIs on beta blockers and aspirin(?) at
admission and I think that would most likely change. But those particular
measures we would encourage some examination of. And particularly this
timeliness, the times issues in recommendation number five will help on two
measures that we track nationally right now using chart review data, time to
thrombylitics and time to PTCA.

I should pause first on the uncertain on recommendation number one. Now the
National Health Care Quality and Disparities Reports do track both process and
outcome measures and we talked about it a little bit in the full committee, we
were tasked to do that by the Institute of Medicine that gave us guidance on
this. And in some selected cases we have some very good clinically specific
measures that are provided to us on control of hemoglobin A1C, control of blood
pressure and control of lipids for diabetic patients and these could be
obviously looked at for other patients as well, diabetes just happens to be the
one that there’s a lot of consensus on in terms of the measures.

This is provided by NCHS through the NHANES survey and is great data, so I
guess we have something of ambivalence regarding this in terms of our specific
role which is what I would like to limit our feeling on here, our comments on
here, which is that while this would as Jerod and Peggy add tremendously to
information available and I think if you talk to any physician that works on
quality improvement and having this type of information about a set of their
patients in a given unit they would say that this is great. For our purposes in
terms of national reporting we do have a ready source of data and anyway, we
think it may fall a little bit more on the margin in terms of the value and
immediate need.

I can move forward to a couple caveats, concerns going forward. I guess
I’ll summarize this slide and the rest of my slides by saying that there are a
broad set of recommendations in the report and in general the recommendations
don’t, having read the report and again with all praise, in general the
recommendations don’t present things not to do, they sort of in general present
things we should be doing, sort of additional things to do. And while that was
kind of the point of doing it, figuring out how we can do this better and then
usually how you can do something better is rarely to chuck out something you’re
already doing, we do think that there is some room for the committee to at
least advocate additional examination of what, which is what you’re doing
today, what of the many, many clinically related data elements that could be
added to a standard transactions really make sense. And I just wanted to bring
this up mainly because AHRQ is working right now on a request for task orders
that will look at this exact question.

Again AHRQ QIs are initially developed for internal quality improvement
efforts and there’s been a lot of interest in using these QIs for public
reports comparing hospitals, but there’s a lot of questions as we’ve been
talking about in terms of the clinical validity of these datasets. And the
motivation for doing this request for task orders has come from a lot of
different places, this committee, NCQA’s Measurement Advisory Panel, as well as
the Consumer/Purchaser Disclosure Project.

The research questions that we’ll be tackling, or the contractor will be
tackling in this request for task orders which was put out April
12th, we’re still in the process of working on it, but what key
clinical data elements can be added to selected measures from the AHRQ QIs and
to what extent does the use of clinical data in performance measurement add
value relative to the cost of data collection. Now this is, this particular
mechanism is used by AHRQ and by other HHS agencies on kind of evaluation type
aspects and so that’s why it’s relatively AHRQ specific in terms of the AHRQ
QIs but it’s, obviously it’s been explaining, the AHRQ QIs are only kind of a
lens on administrative data in general so we feel like it’s very widely

I can provide, again I will send this information so you all will have it
but the QIs that are going to be examined in this study are mortality rates,
patient safety indicators, and potentially failure to rescue which is another
patient safety indicator, but a subset of those two and of outcome areas as
Jerod was talking about where risk adjustment and specific risk adjustment
would be important.

The clinical data elements that will be examined include history, vital
signs, clinical assessment, procedure reports, lab values, cultures, DNR
orders, and times, and obviously vital signs, lab values, and times would be
especially important in terms of the set of eight recommendations we’re
considering today.

I won’t go into any more detail on that mainly because that’s about as much
as I have, we’re just sort of in the planning process of that but the reason I
bring it up is mainly to say that not just that particular request for task
order but there may be other work that other partners are doing that would help
the committee hone its discussion across the numerous recommendations about,
and even a summarization of these hearings the sentiment that some examination
that is data driven and the efficiency here in terms of the outcome being that
what are the most efficient data elements to add will be data driven and
statistically determined, I think there might be some room for that in addition
to the kind of expertise that testifiers are bringing here to you today.

Finally I’ll just say that we look forward to working with the committee on
just in general, I know this is the first kind of big step that was Bob has
moved the committee on in terms of his tenure but we’re looking forward to in
general collaborating more with the committee on shaping the National Health
Care Quality and Disparities Reports as well as with keeping the committee up
to date on work we’re doing with the QIs and the National Quality Forum, that’s
actually why Ken’s not here, he’s actually been working on the quality
indicators while we’ve been talking. And then the Quality Committee we feel
like would be a perfect spot for ongoing discussions on measure alignment and
that’s I think really what we’ve been talking about here is the standardization
and measure alignment which would probably be the number one legacy but the
committee could leave in its review of this particular report would be improved
alignment and improved standardization.


MR. HUNGATE: Very good, thank you all for excellent testimony, we’re open
for discussion and questions. Can any of you speculate what Ken might have said
differently then any of you?

DR. LOEB: I think he’d agree.

MR. HUNGATE: That would be my guess.

DR. COHN: First of all I want to thank our testifiers for I think what’s
been some very interesting helpful testimony. First of all Peggy I really
wanted to thank you for I think reminding the workgroup that electronic health
records is not a panacea. As one whose been involved in system development
since about ’84 I think it’s, I mean we keep sort of talking about well when
electronic health records come but maybe part of these lists of items is really
also help inform electronic health record developers since you’re right, free
text or just even, just not structuring it right can create great barriers.

MS. O’KANE: I would actually encourage, I don’t know if this is what you’re
looking for but I think that’s a huge topic and I’ve been kind of going around
looking at installed electronic health records just for my own personal
education and it’s quite edifying to learn about the experiences that people
have and unedifying at the same time. So I think NCVHS, I think that’s in your
bailiwick, right? I think you should declare it in your bailiwick and I think
you could make a tremendous contribution by pointing out some of the pitfalls,
some of the ways in which we can sort of leap over some of the early
experiences that people have had where they’ve been very disappointed with what
the records have done.

DR. COHN: Well it isn’t even disappointment, it’s just not engineered to
answer the right questions.

MS. O’KANE: I was visiting Chuck Hylow(?), I don’t know if you know him but
he’s got this Greenfield Medical Clinic in Portland, Oregon, which is kind of a
clinic of the future and he introduced me to this young doc that he’s got in
his practice and they had, I won’t mention which brands they had but the guy
said he just wants to kick it all the time because it’s so clunky and it
doesn’t do what it’s supposed to do, and particularly talking about the
registry function. So I think that there is tremendous, tremendous value could
be gained by just kind of reviewing some of the experience and trying to send a
signal to the vendor community or have some mechanism for feedback from
customers to vendors, it may actually take some kind of institutional
intervention I think.

DR. LOEB: As I view this I think one of the real problems is we develop
performance measures from a whole variety of data elements on one side and
you’ve got the vendor community on the other side and they don’t integrate up
front, they try to integrate at the back end and that creates enormous
challenges, both technical as well as political and scientific. And I think
that’s a real issue, I know that there are a lot of discussions going on today
about trying to do this in tandem as opposed to do them entirely separately but
there’s FTC issues and a whole variety of other drivers that make that really
difficult but I think that’s part of the problem, the thought processes are
really quite different.

DR. COHN: Well if I can just give an example or two, my specialty is
emergency medicine and so I developed systems back in ’84 for the emergency
room. And I have very good systems that would automatically time and date stamp
admission data and when you’re admitted to the emergency room, but of course
not arrival data. Time that you documented on an intervention but not
necessarily the time that the intervention was actually delivered because
you’re obviously in the room taking care of the patient and most systems don’t
allow, I mean there’s some obviously that you can do it in real time but
usually in the midst of a mini code you’re not sitting there looking for your
key board. And so it gets in the way of all of the measures that you’ve all
been talking about.

I had one final comment, and actually it had to do with the work from AHRQ,
and actually I want to thank you for bringing forward this concept of evidence
based looking at these pieces. Now I’m actually not part of the Quality
Workgroup, I’m from Standards and Security, and we’re sort of the people who
are trying to figure out how if this were to happen how to make it happen. And
I think there’s been a sense, at least I’m speaking for myself, of a sense of
like high level concept, right concept but what exactly do we mean here and is
there really a cost benefit on this one and can you be specific enough that I’d
even know how to do a cost benefit. And it sounds to me like you’re actually
going to be going off and trying to do that work with much greater specificity.

DR. KELLEY: Right, I didn’t mention but I should have, I think I was trying
to limit my slides, but we’re partnership with Pennsylvania Health Care Cost
Containment Council to use their data, that will be the kind of data setting
for the analysis. So anyway, it will be, I mean in some ways it’s getting to
Jerod’s kind of synthesis of both ends of the spectrum there but the feeling
was it would be nice to have some statistical experience in a small kind of
pilot setting at least that would get compared with some of the thinking about
on this report and it evolved in part because of some of the recommendations
that were getting formed from this. But as everyone here knows it’s not, this
is kind of thinking about improving administrative data that’s been coming from
a number of different places.

MS. POKER: I just had to relate to what Simon said before when a patient
codes and you don’t look at the time, the patient’s care gets so in the way of
data collection it seems like. But the point is it shouldn’t, in a good system
what we should be doing is capturing all the relevant data without it
necessarily becoming a burden, so that’s hopefully our goal here. But Jerod,
you had mentioned before when you were talking about the right question to ask
in quality and that’s I think what we’re struggling with here, we’re trying to
find the right questions.

In addition to the recommendations that were presented does the panel feel
there are recommendations we haven’t presented? Or are there quality questions
as you put it, the right questions, that you would like to suggest to us to
consider also? We’d really appreciate it.

MS. O’KANE: We’d like to come back to you with some suggestions. I think
there’s a whole set of questions around cancer where we’re completely hobbled
because we have such dysfunctional registries and so forth, we just had a
meeting on this so it’s on the top of mind for me. But I think some of the
other, some of the things that we had to turn down as potential quality
measures because we couldn’t get the data, but I think that there’s been a
framing of priorities, the IOM priorities are I think a great place. I mean if
we could make progress on those things that would just be enormous. And I think
one thing that I would urge the committee to sort of think about is there’s not
going to be a solution that’s the permanent and eternal solution so I think
moving forward, we always learn and we’re limited by our headlights and I think
you can take a sort of best report, best set of recommendations, given what we
know now and maybe looking back in five years you’ll say well why didn’t we
think of that and the reason is because we didn’t have all that information. We
will come back to you.

MR. HUNGATE: A follow-up related question/ observation I guess. I think I’m
hearing from the combination of things that there are some sub pieces of some
of these recommendations that are better then the more global recommendation
itself, in specific places the data is more helpful, the disease specific,
condition specific. And I don’t know what the right approach for us as a group
in moving, how specific should we try to get in our more complete
recommendations. And Simon says very, and so I think that’s part, if it’s an
efficient measure then it will have a strong business case that says the value
exceeds the cost and that probably is quite specific. Now how do we get from
where we are to there?

DR. LOEB: I’m not sure the answer to that question but I would certainly
agree with what was just stated relative to the degree of specificity, the
devil really is in the detail here and one of the great frustrations in
performance measurement is things that at face value do look alike, when you
start getting underneath them at the data element definition level they really
aren’t alike, that’s what’s really dogged a lot of the work, and I see Barbara
sitting back there from her previous life, working together with Barbara,
trying to align and in fact create identity between measures that the Joint
Commission specified and measures that CMS specified, that at face value looked
pretty similar. Its taken us years and right now we anticipate issuing a
manual, an aligned manual, one manual with the Joint Commission and CMS logo on
it roughly mid-September and I mean it’s literally taken years to get to that
point. But that’s going to move we think the field forward much more quickly
and comfortably because it will indeed create a singular data collection

DR. KELLEY: I’m wondering about though, and I don’t know entirely sort of
what your hope is for the outcome of the series of hearings that you’re going
to have on the recommendations but it would seem to me that the broad set of
the recommendations, if implicit in your questions there’s sort of range in
value for the different recommendations that therefore there might be ones that
are higher value and again we’re thinking, not just cheaper but the actual
value that they have, but that the committee, the outcome of the effort might
be to recommend those that are of higher value and I wonder if, I mean because
that kind of work, I completely agree with these comments, that it really looks
like they’re lined up, we have the same measures, let’s move forward with them
and then you figure out they’re very different, there’s a lot of different
specifications that are different, it’s a set. I’m just wondering if you need
to have all of the specificity for the outcome of this particular effort or if
it’s the next step and the piece that follows is a subgroup getting together on
a subgroup of the recommendations.

DR. LOEB: This sets the table though to see if it’s really critical.

MR. HUNGATE: That’s part of my concern is that we have limited resources I
have learned very quickly and so we can make principle statements a lot more
easily then we can make reconciling statements. The standards group does get a
lot closer to reconciling but it’s got a pretty full plate too so I’m trying to
think about what’s the best way to get what’s most useful.

MS. GREENBERG: I think if you look at the candidate recommendation that’s
like for the first two an that’s where I hear, and I’ve heard this discussion
at the National Uniform Billing Committee as well about well which vital signs,
which lab tests, when would you, and it’s pretty clear I think that if that
isn’t highly specified then you won’t get comparable data. And the question of
who should be highly specifying that, and I think the way these recommendations
now is to create a mechanism for reporting basically this information and then
I think the report does go into this a little further about having the
appropriate groups really identify then exactly what should be collected. And
do you feel that that would be useful in and of itself, trying to push the
state of the art forward, calling for a mechanism for collecting this without
actually saying these vital signs or these tests at this time, etc., and then
would you see groups such as your own sort of filling the void there.

DR. LOEB: I think there’s a couple of caveats that under gird what you say.
First is a consensus development process that includes the right stakeholders,
I think that’s really important. Second caveat at least that comes to my mind
would be the notion of being sure that there is a cost benefit analysis that
underlies this, that there really is identified the need for blood pressure,
heart rate, etc. We all sort of intuitively think that but we believe in
evidence here so does the evidence support it, I think that’s going to be
really important as well. I’ve got a few more but I’ll trade the microphone.

MS. O’KANE: Well I think the whole question of who’s going to implement
measures is a big open question these days and there are all kinds of projects
going on, kind of ad hoc projects, there are people like us who I think maybe
have a bigger staff and a little more expertise with this stuff.

I think what’s really important though is to understand that this has to be
sort of a living relationship, that there’s not sort of one set of
recommendations that you could put out that would solve the problem once and
for all, I mean for those of us that are maintaining measures we’re constantly
discovering new things that we didn’t know about or somebody’s data system
where it’s not working. So there’s a need for thinking about how this would,
and I don’t have an answer because what I’m saying to you is I think sometimes
there are organizations where you could have an ongoing relationship that are
more institutionalized and then you also have kind of ad hoc projects where I
think it becomes much more complicated and I don’t know what you can do about
that except maybe realize that some of these things are going to have a limited
half life.

MS. GREENBERG: I don’t want to belabor this but I guess my bottom line
question is does a recommendation like one and two, which generally you
supported with the caveats, all of the caveats that we understand, but would
the recommendations at that rather general level be useful? Do you see a
benefit to a recommendation that recognized that collection of some of these,
some vital signs, some lab test results, etc., and having a mechanism to do so
in administrative data is a thing that should be pursued?

MS. O’KANE: I think you should be more specific.

DR. LOEB: I think it points the car in the right direction but it doesn’t
move the car and I think that’s a real significant issue.

MS. GREENBERG: So you’d like to see the committee actually being more

DR. LOEB: From input from appropriate parties, yes.

MR. HUNGATE: The same subject right?

MR. REYNOLDS: Well, it’s a little different slant and Gail has been waiting
and others so I’ll wait.

DR. JANES: To a certain extent similar to what we were talking about,
something that Peggy said jogged my memory about something that struck me is
the issue of physician profile and while we didn’t really speak directly to
that in this report —

[Inaudible.] Are there certain elements that we have included in these
recommendations or that we haven’t and perhaps should have that would play into
that world and would strengthen our ability to do that given the fact that it
has —

MS. O’KANE: Well, if I could just step back for a second and say that I
think that the issue of measuring individual physician performance is not for
sissies, I think that there are, our approach to this I think has been fairly
conservative and that is it’s a target based approach where you basically see
if you comply with certain kinds of standards, using ADA guidelines, using
American Heart Association, ACC guidelines, and then some structural stuff with
our practice, physician practice connections program. So I think we’ve, while I
think people think of this as some sort of radical thing we actually think it’s
a fairly conservative approach and really quite distinct from the kind of
profiling that I think we’re on the verge of seeing all over the place, some of
which will work and some of which will I think not work.

I do think, there are examples of plans that give profile information back
to the doctors from the claims data at considerable cost to the plans and they
have had enormous success in improving outcomes and the physicians love it, so
I think the thinking about creative ways to take claims data and populate
practice management systems could take us a long way forward while we’re all
waiting for the day of the all singing all dancing electronic medical record.

So I don’t know if I answered your question —

DR. JANES: Yes, just the last set would be and if we want to move in that
direction, are there certain —

MS. O’KANE: I think the question is what does that direction mean, I think
the answer I would give is I believe there is very important utility to having
physicians profiled, I don’t know that it means that everybody’s results get
publicly reported. This target based approach we’ve got you don’t, nobody
reports anything if you don’t hit the bar. The boards now are sort of trying to
figure out exactly what their evaluation methodology is going to be like and I
think there’s a very important, we need to think about evaluation or feedback
as something more comprehensive then evaluation, something I think that you
could learn from.

DR. JANES: I couldn’t agree more, that it has inherent value and —
[inaudible] — but like I said, to get back to my initial question and that is
if we all around the table decided that yes this is a good thing and that no we
don’t want to wait, we don’t want to engage in large scale chart reviewing, we
want to wait for a comprehensive record, are there certain data elements in the
administrative data which could facilitate that process which we have not
decided or —

MS. O’KANE: Let me get back to you on that because I think we’d like to
just go, I’d like to ask my staff that question because this gets to a level of
detail that I’m just not —

DR. CARR: Well just speaking as a physician if I may for a moment and
thinking about physician profiling, I think the categories of volume, practice
and outcomes are a few ways of looking at it and I think we’re looking at
volumes and I think there’s consensus among physicians that that’s appropriate.
Adherence to best practice I think is another thing and if we got those two
done we’d have made tremendous strides and I think we’ve already seen that in
claims data feedback and so on. I think where you get challenges is outcomes
and that’s where the sophistication of your risk adjustment system comes into
play and I think that’s a much, much larger discussion with a different
audience or different participants in addition to the folks here. But I think
on volume and adherence to best practice there’s information that we’re already
getting that we could get —

MS. O’KANE: Well, I think we’re getting it but I think we’re getting it at
massive costs, and with our recognition programs it really involves the doctors
going through paper records and when I see what’s been done for example by
Exelas(?) Upstate New York which has been very innovative I think that there’s
a lot of potential here to give them better information. Now there’s also the
separate issue of the volume, if you’re looking at a primary care physician
they’re not going to have enough of any one particular category —

DR. KELLEY: But I think this also brings up the special point, the issue of
risk adjustment, having the physician specific information and having the data
elements still does not, you still don’t have the risk adjustment yet, I mean
we can witness what is going on, CMS is working on some of their risk adjusted
AMI 30 day mortality and stroke mortality reporting and we’ve been real
interested in narc and wanting to include those in the National Health Care
Quality Report and the questions even of should the models be kind of
clinically derived or should they be statistically derived and the people come
from very different camps on that. So I do think that the devil is in the
detail as people have been saying and it’s a long term prospect to go from
actually being able to use the outcomes even once you have agreement on the
data pieces that we think are most kind of statistically efficient as well as
clinically important to have.

DR. LOEB: Actually could I just kind of dovetail on what Ed just said in terms
of using the data, I think one of the biggest issues that dogs the discussion
in this area is how do you sort out the accountabilities. Even if we had unique
physician identifiers and we’re dealing with that right now in the context of
measures that track such things as beta blocker use, ace inhibitor use and so
on, where do you point the finger and I understand the whole notion is quality
improvement and not accountability but let’s be realistic about this at the
same time, when it means board certification and ultimately perhaps pubic
reporting the stakes are very, very high and the tempers are very, very short
and you’re right, it isn’t for sissies.

MR. REYNOLDS: I appreciate your comment and I can now add to my resume that
I have heard Ed twice so I’m pretty excited, no, actually I’m going to ask him
some questions and that’s even better, it’s getting exciting. As we deal with
this, I deal with this along with Simon and the others on the Standards
Committee, the philosophy of quality is easy to agree on, the reality of
capturing the data and doing it, and as we right now are spending many days
going through e-prescribing and we’re dealing at the individual doctor level
and then you look at quality where we all have to remember that the individual
practitioner in some of the smaller places, the large hospitals may have the
bodies to do some things but as we get down, so what do you as kind of the
quality evaluators, what do you see as the value proposition, that if this
group agrees on what it’s going to be and then the standards group agrees on
how it’s going to come through, what is the value proposition that can be sold
to the individual practitioners as they deal with their vendors and then they
deal with everybody else through the chain that they get this done?

DR. LOEB: I’ll be the idealist in answering your question and it is
improved health care quality, pure and simple, the ability to track
longitudinally how you’re doing as well as compare it with that of others in a
standardized manner.

MS. O’KANE: I think the value proposition is being made by all kinds of
forces in our society that are taking on the issue of accountability in health
care and I think physicians are, I think there’s a lot of frustration and
appropriately so at all the different things that are coming at them. I
personally think we’re not going to solve these questions just through data
strategies and measurement and that there’s a kind of renegotiation of the
whole definition of accountability that has to go on and I think we’re
beginning to see that, I think that the boards are stepping up I think, trying
to, I think that there’s a whole agenda here that’s not simple in any way. But
I think that the fact that this is going to be happening one way or the other
and there are just multiple examples of it, it’s not necessarily an only upside
proposition, if you don’t sort of embrace the agenda and try to figure out a
way that we all can constructively work towards advancing it then we’ll
continue to live with chaos and waste and unfair judgment.

DR. LOEB: And follow the money, that’s the other piece of it.

DR. KELLY: I think that’s a good question because it’s sort of, it would be
easy to get caught in the kind of benefits of improved data for improved data
sake given that generally we’re converts to that line of thought. I think
Jerod’s point is well taken, improved health care quality in terms of being
able to at a much cheaper cost, not even possible for your average physician to
kind of do regular chart abstractions and compare themselves to CMS data or

But in terms of a pay for performance environment that would obviously,
that environment has to kind of be in place for the actual kind of dollar side
of that value equation of what Jerod’s talking about, I mean I would think that
every average physician would want to improve quality for improved quality sake
and improved outcomes but it would also be the pay for performance aspects.

But I would just say one caveat having done in my previous life quite a bit
of research on the implementation of different quality improvement and quality
assurance methods and I had done a research project looking at feedback of
performance to a set of, relatively large set of providers and in terms of
sustained improvement in quality just sort of knowing where you are and the gap
between you and maybe the best performance in your class or even the rest of
the state or however it is only gets you so far, I mean there are push factors
that get us to do things better in our jobs, in any job, there are pull factors
in terms of getting actually paid for our jobs.

But then there’s a lot that’s internalized in terms of just wanting,
everyone wants to do a good job, I mean my garbage collector, you hear this all
the time from anyone you know, you always see it in TV interviews, I’m the best
darn garbage collector, blank, and so I think that there’s aspect of it but
that sort of wanting to be as good as the best only gets you so far, there are
structural barriers that do exist.

So I do think we have to think hard about the value question and where it
ends up and what are some of the remaining barriers that this won’t solve and
be honest about how it gets solved.

MR. HUNGATE: This has been a very helpful panel, I’m sure that there’s a
lot of follow-up that needs to get done and we’ll get back on those things. But
I know right now Jerod’s got to go catch his cab so I think we better take a 15
minute break and return. Thank you very much.

[Brief break.]

MR. HUNGATE: Let’s see if we can move along. Okay, I don’t think we’ve
really had any banjo players so you’re not following a banjo player in this
case. But I will say we’ve had two excellence panels and look forward very much
to your input to our what I characterized this morning has hearing on the
subject of lurching toward measurement. And I don’t think it’s a bad

For both of the last two panels we’ve taken a few minutes and gone around
the table to introduce everyone so everybody’s name, and that’s all the speaker
introductions we do so say what you wish in that context and let’s start with

DR. HOLMES: I’m Julia Holmes and I here at the National Center for Health
Statistics, and I’m a staff member on this Quality Workgroup and at NCVHS I
also work on using NCH(?) data to measure quality and disparities, we’ve been
very heavily involved with AHRQ’s National Health Care Quality and Disparities

MS. POKER: Hi everybody, my name is Anna Poker and I’m from AHRQ and I’m
the staff lead for the Quality Committee. And AHRQ I work also on the quality
report, on the NHQR and the NHDR, and I’m also part of the patient safety team

MR. HUNGATE: Bob Hungate, principal of the Physician Patient Partnership
for Health which is me and nothing more, chairman of the Group Insurance
Commission, chair of the Quality Workgroup, member of the committee.

MS. HANDRICH: I’m Peggy Handrich, I’m a member of the committee and a
member of the Quality Workgroup, I’m from the Wisconsin Department of Health
and Family Services with a particular interest in Medicaid and health data
collection on the state level.

DR. CARR: I’m Justine Carr, I’m a physician in health care quality at Beth
Israel Deaconess Medical Center and Health Care Quality and I’m a member of the

MR. REYNOLDS: Harry Reynolds, Blue Cross and Blue Shield of North Carolina
and a member of NCVHS.

DR. COHN: Simon Cohn, national director of health information policy for
Kaiser Permanente and a practicing physician.

DR. JANES: Gail Janes, metatheneologist(?) at and staff to the Quality

DR. EDINGER: Stan Edinger, AHRQ, and also staff to the subcommittee.

DR. KAZANDIJIAN: I’m Vahe Kazandijian, I’m from the Maryland Hospital

DR. HOCHBERG: Stan Hochberg, I wear a couple hats here, I’m the medical
director at Provider Service Network in Boston, I’m also the chair of the
National Quality Partnership Physician Council.

DR. PAUL: Barbara Paul, I’m senior vice president and chief medical officer
at Beverly Enterprises and formerly was at CMS for five years most recently
heading up a lot of the quality initiatives that were launched under Tommy
Thompson and Tom Skulley.

MS. FOSTER: Hi, good afternoon, I’m Nancy Foster, I’m a senior associate
director of policy at the American Hospital Association, and in that capacity
and perhaps relevant to your discussions this afternoon I’ve been working very
closely with a number of other organizations on an activity to nationally
voluntarily report hospital quality data to the public. Barbara was
instrumental in that when she was at CMS and so have been many, many others.

MR. HUNGATE: Okay, we’re trying to set up the way we do things so that
there’s 15 minutes of presentation, no questions after that so that we save all
the questions until the end, and it has produced productive discussion so I
think we’ll stick with that. Nancy, you’re on your way.

Agenda Item: Provider Organizations – Panel 3 – Ms.

MS. FOSTER: Thank you, Bob, and thanks to all the members of the committee
for allowing us to be part of the discussion, I really appreciate that, I hope
to be briefer then 15 minutes because I’d really like to get to your questions.

A couple of points I’d like to make going in, one is that as you think
about improving health care quality and improving health care quality measures
it’s important to remember that at least the membership with the American
Hospital Association, over 5,000 members strong, represent a very diverse group
of organizations with very different capacities to accommodate changes. Some
are very much at the forefront and we’ve heard from a couple of the practicing
physicians on the panel here who are among colleagues in organizations that are
trying to adopt information technologies and move ahead rapidly. Others work in
very small institutions, seven, eight bed, some even smaller then that. I think
our smallest member is a two bed hospital. Everything in between, absolutely
everything in between in terms of numbers of beds, numbers of patients,
sophistication. The only common thread outside of the fact that they care for
patients who need general acute care is that they are committed to serving
their communities in the best way they possibly can. And they view quality
measurement as an activity that will help them serve those communities better.

In that light when I come to talk to many groups I hear this sort of angst
about the fact that there’s not a lot of quality measurement out and available
to people. That’s not the same perspective that our hospitals have, our
hospitals think there’s an inordinate amount of quality measurement going on,
which is not to say that we think that there’s a lot of quality information out
there, there’s a big distinction. And part of what I think I’ve seen in the
report that you shared with us and asked us to comment on is that struggle to
figure out how best to serve this master of finding a way to communicate good
information, not data, good information, to the people who want to use it to
effectively improve care, which could be providers, purchasers, the public,
patients, any number of organizations. But none of us can do that quality
improvement job without a source of data and the context in which to understand
that data. And we struggle because the data themselves don’t come with a
context and we really need to create that.

At the American Hospital Association, and I know this is true for several
of our colleague hospital organizations that are in the broad based national
voluntary initiative that I referred to before, we want to work with other
organizations to accomplish two goals, one is to provide information to the
public and to our providers that will help them to improve quality. And the
second is to dramatically reduce the cacophony that exists around measurement.
It’s why we started the quality initiative, it’s why we continue to work with a
number of organizations, it is an incredibly important activity as far as our
membership, our board, and our president our concerned. That is not the
President George Bush but my President Dick Davidson.

And we’ve started with a small set of measures, just ten measures that come
out of the Joint Commission and CMS, a common set of measures, as a way of
building a data path towards getting information out to the public. It was
really a learning process for us and I want to communicate some of the
information we’ve learned as we’ve moved ahead with that effort to get data
out, information out, to the public.

Jerod and others on your previous panel referred to the fact that it’s
incredibly important to get standardization of data, standardization of
measures. It reduces the burden of collection, it moves things forward. I
cannot emphasize that enough, that kind of standardization which is reflected
in your report is incredibly valuable and we would fully support the efforts
you’re undertaking to get there. Having said that, it is more then just the
standardization of the measures and the specification, it is around
standardized methodologies for data collection, it is really down at the very
nitty gritty level of the person in each hospital or in each organization
having a common understanding of what they’re supposed to do, how to do it, and
doing it that way in order to build up to where you have data that you can
compare and general real information. It is not an easy task.

In asking us to come here you’ve asked us to comment on the value of doing,
of imbedding certain information in route billing information, administrative
data, others, and I have to say we understand that question only in context of
what we’re trying to do and in moving things forward on quality and I don’t
know how to answer the question, is it valuable to imbed certain information in
the uniform bill. Well only if you have a common set of understanding of what
the priorities are, what should be measured, the measures that will effectively
answer those questions, and then drill down to the nitty gritty level of each
of the specs that needs to be gathered in order to effectively measure whatever
you’re trying to measure.

Your report points to the IOM Report on National Priorities, a great place
to start, 20 priorities that an

esteemed group laid out at the Institute of Medicine, a wonderful starting
point. But there’s a large jump between those national priority areas and
knowing what to measure. It would be extremely helpful if you could help us
understand how to get from those priority areas knowing what are the most
critical elements of care to measure about each of those priorities. Because
once you know what the most critical elements of care are then you can go back
and look and see if you have measures that will effectively do it and begin to
talk about what data elements you need in order to collect the information
needed to create the answers to those measures, a very detailed process.

In one area of your report as I read it you talked about the value of
outcome measures, particularly for consumers, and asserted that outcomes
measures are in fact the most valuable for consumers. I’d have to say as
somebody whose been working in quality measurement for more years then I care
to comment on that I think, and many others would agree, that the door is still
open on that one, the question is really still unanswered. When we look at
outcomes measures such as the Medicare mortality data that was collected and
reported in the early ‘80’s, and other mortality data such as that
reported in New York and Pennsylvania, it’s been incredibly useful to some of
the provider community folks but the fact is that folks only believe it if it
has a good and rigorous risk adjustment methodology which you’ve talked about.
Some very good focus group work with consumers has suggested that when they
look at outcomes that have been risk adjusted they read that as whoever is
collecting the data having played with the data, having fixed the day so it
comes out the way they want it to come out, not having risk adjusted it in the
way that you and I would understand risk adjustment.

There are some very good process measures that maybe as effective in
communicating to the public as outcomes measures and I wouldn’t rule those out
if you would consider some of them.

I want to just finish up here by commenting on some of your specific
recommendations and then move on down the panel. You have recommended
collecting the specific provider/physician identifier and I know that the
previous panel had some discussion around that. We at the American Hospital
Association understand quality to be a systems property, not an individual
provider property. Several years ago I used to work at Georgetown University
Hospital doing quality improvement work for the Department of Medicine and
while I was there there was a picture that appeared in Life Magazine, as the
centerfold for Life Magazine, they came and took it at the hospital, they
wanted to identify anybody, any clinician in the hospital that had touched a
single patient. They identified the patient, got his permission, had him sit
front and center, and there were 99 people who stood behind him. Now this was a
patient who’d only been in the hospital for three days, all of those people
effected his outcome and as we look at it people who cared for him before he
came in the hospital, people who cared for him after he went home, all effected
his outcome. Systems properties are very, very important and by proposing to
imbed this individual physician identifier you may be running the risk of
scaring people by thinking you’re going to come back to them with the big
accountability hanna(?), and you may be misleading folks into thinking that it
is in fact an individual physician who is responsible for everything that
happens to that patient while in the hospital.

You have also suggested that there should be some flags imbedded, like the
flag to identify the conditions that are present on admission. That would be a
great idea actually, that would be extremely helpful. The problem is that
there’s not a lot of consensus around how to identify what was present on
admission for some conditions that you’d want to flag. We’ve been dealing
around pneumonia, so was this a community acquired pneumonia or a hospital
acquired pneumonia, how do you define, what’s the timeline look like, when do
you know, all of those issues come up and there isn’t to the best of my
knowledge any consensus on how to effectively identify that. So having a flag
when you don’t know what to flag isn’t helpful.

The same may be true and you’ve had a lot of discussion already today
around the service dates and times, it’s very difficult, it’s actually a fairly
rigorous process to try and collect that information about what was provided
and what was provided, it’s very time consuming. Much of the data that we
collect right now, especially if it has to be abstracted from medical records,
is extraordinarily time consuming.

We’ve taken a look, we’ve asked hospitals, I don’t have a really solid
number to offer to you but most of the 100 to 125 bed hospitals that we work
with say they have at least one full FTE devoted to simply collecting data
right now, more if they’re actively involved in various projects. That’s a lot
of investment. And to date hospitals have said we see the value in our own
internal quality improvement work, we see the value of collaborating with other
organizations or collecting data similarly. We don’t see the value in the
public sharing of the information, not that we’re against doing it but there
hasn’t been a response from the public in the way that people have talked about
it, so that part of the value proposition isn’t there.

As you move forward to think about imposing more data collection you’ve got
to think very carefully about what it is, what the value is, to whom that value
is, and how you make sure that the folks who incur the burden at least get some
sort of recognition for that burden and response.

Let me stop there and pass it on to Barbara.

Agenda Item: Provider Organizations – Panel 3 – Dr.

DR. PAUL: Thank you and I would like to first thank you for the invitation
to be here, I really am honored to be here and would also like to add my
congratulations on your report, it was wonderful to read. I also am glad I was
able to be here a little early to hear Jerod and Peggy and Ed because in my
former life at CMS I basically would have been belly to the table with them in
that presentation, or this one this morning where Trend Haywood, who was my
deputy, is now taken over the position I used to have at CMS.

And I’d just like to echo a couple things that Jerod and Peggy talked about
in terms of the incredible value of standardization to propelling health care
quality, I hope you’ll see how I see that value from my new lens in the long
term care area in a minute, and also the value of really integrating data
collection into the clinical flow of processes.

And I’d say that I think that what my comments are going to be, I’m going
to drill very specifically into the long term care area because it’s clear to
me that you’ve had a lot of input from hospital centered folks and since I seem
to be the one person at least on this panel representing long term care I’m
going to get very specific, a little different from the discussions you’ve had
thus far, but personally I would say that my comments are yes and, that
listening to Jerod and Peggy particularly, it’s a yes and which you’re going to
hear, yes and in terms of you’ve been talking about the vehicle being an
administrative claims vehicle, I’m going to tell you sort of a yes and to that.

You’ve talked about the role of standardization and measurement in terms of
using it for quality improvement and for accountability, I’m going to give you
sort of the yes and for that. And then you also heard a very, very brief set of
comments from Peggy about the recommendations seven and eight, functional
status, hopefully I can give you a little more depth on yes and there.

I would like to first just let you know that Beverly Enterprises cares for
roughly 45,000 patients around the country in about 350 nursing homes. We
provide physical therapy, occupational therapy, and speech therapy for roughly
another 40,000 or so patients in some additional nursing homes that we contract
with. We also have a number of home health and hospice agencies and so we
really do as a company reflect that sector that’s often referred to as long
term care, so the elder care spectrum.

And we measure our quality, we have a very detailed scorecard that is
heavily weighted toward clinical quality. We align incentives at our company
toward higher quality very, very explicitly and that was one of the surprises
and one of the reasons why I took this position. Also our rehab and hospice
providers really are I believe at the forefront of some of the measurement work
that’s going on in that area which is still fairly developmental.

So my presentation this afternoon really is going to lay out my case for
why it’s important, my case for why I think it’s just absolutely terrific that
we’re going to talk about long term care in this fairly hospital centered
project. Secondly I’m going to really focus on recommendation one, which is
probably our number one need, talk about recommendation seven and eight and
then a few conclusions.

Our hospital patients are long term care patients and long term care
patients are hospital patients, and just one, a couple of quick statistics to
try to prove my point. In 1999 for the 1.5 million long term care residents in
that year of our nursing home admissions nationwide 46 percent came from a
hospital, and of the nursing home discharges 30 percent went to a hospital. So
to think that you’re talking about hospital quality and hospital costs you’ve
got to be looking at long term care quality if you’re ever going to really get
at hospital quality and costs.

My second point to try to set that stage is that the transitions between
these settings are frequent and this was just one interesting study that I
found that looked at these transitions between the various settings with a
cohort of elderly patients, in and out of hospitals, rehab facilities, nursing
homes, home care, site facilities, hospice, etc. Two year period, cohort of
seniors over 65 years old, 18 percent of them had at least one transition in
that two year period and for women over 85, a very fast growing cohort, it was
almost half of them who had at least one transition. And of those who had a
transition 22.4 percent had at least one transition related problem, at least
arguably transition related problem, an ER visit in the 30 days after that
transition, an avoidable hospitalization or a return to institution from
community. So my message to you, and I hope you’ll agree, is that if your goal
is to be a part of improving hospital quality and costs that I think one of
your key objectives needs to be to find ways to help us in long term care to
improve our quality. And I think you have an opportunity with standardization
to do that.

I would have to say that as I’m trying to understand the landscape in my
new world I would say that intense oversight has a bit of a silver lining for
us in which it has resulted in a standardization of care process that is fairly
impressive to me. There is a lot of standardized valid data, the MDS dataset,
the Oasis dataset, very clinically rich data that is standardized, transmitted
to a central repository, validated, scrubbed and cleaned and studied, much more
clinically rich then hospital data currently, certainly from the standpoint
going back to my old hat at Medicare. And the nursing home environment for
better or for worse is very heavily an environment of accountability. So I
think those are, while that’s tough it’s also a silver lining for where we need
to go as an industry.

I would also say that prior government approaches really tended to focus on
the bottom of the bell shaped curve, trying to weed out the bad apples, and
we’ve got to do that, it’s important. But what I think is, I’m really glad
about is that increasingly government and other sort of approaches to quality
in the long term care area, just as in other settings of health care, are
adding additional approaches. Technical assistance from the quality improvement
organizations, consumer information, incentives, etc., and so I’ve really been
delighted to be moving to this part of the health care sector where we can
really start to chase some carrots as opposed to always just kind of running
away from the sticks.

Also just a few other comments, I would just say that I’ve already
mentioned MDS and Oscar data, we are doing public reporting of that right now
as you know on Medicare.gov, a fairly robust set of quality measures being
reported there. And increasingly as I mentioned in terms of my company we are
aligning internal incentives with quality results. And the last bullet point,
nascent electronic health records exist in long term care in pieces, the MDS
dataset is a module if you will, is a piece of it. Our physician order entry
module is very, very clunky but it’s another piece. We have lots of pieces of
it in our nursing home settings in particular and so I’m really looking forward
to your recommendations helping us to move these pieces forward.

So let me go ahead to recommendation one and just really focus in then on
the long term care setting. And as I said before I think what I’ve been hearing
so far in your deliberations is that you’ve been talking about the need for
standardized laboratory result reporting for purposes of measurement and
quality improvement and accountability, and I would say yes, and, what we need
it for however is just to plain provide care for patients. We need it in real
time, you’ve been talking about providing it in administrative claims data,
that’s too late for us, we’re getting those 46 percent of our patients from
hospitals without laboratory data and our clinicians are spending clinical
person time chasing lab data, it’s a real morale buster, it’s a time waster, I
think the business need is so intuitive I don’t think you have to do a study to
know that if we had real time laboratory data for our admissions what a
different it would make with that one stroke in terms of the quality of the
care we could provide tomorrow.

So just to sort of put a fine point on this, the thing that I want you to
understand is that we need it in real time, it needs to be asynchronous with
the billing vehicle. It’s fine if it’s in the billing vehicle for the
hospitals, for these other uses, for risk adjusting hospital outcomes, but for
us we need it in a standardized module that we can get our hands on it as
appropriate. So I’d like you to think about going one better then the idea of
simply recommending that this go on to the UB-04 or UB-05, whatever it turns
out to be.

The next slide in terms of your other question about using it for pay for
performance, I think that you can imagine that again, if we had this
information in real time we would be able to provide better care as I said with
that one step. And right now we operate very, we’re very aware of the fact that
we operate right now under sort of a negative pay for performance environment,
if we fall short we are subject to fines, we are subject to withholding of
Medicare payments and so forth. And by having this information it would allow
us to avoid that negative pay for performance that we live in right now, it
obviously would also help us to chase that positive pay for performance that
we’d love to see in our sector of health care.

Risk adjustment, I think it goes without saying that it would help with
risk adjustment as well as our acuity assessment as we are accepting patients
from hospitals, we would have a much better sense of acuity of the patient
we’re taking care of.

Mode of collection, this is again just to go back to my point about needing
one better, one standard dataset that’s created and maintained by the lab
provider, standard nomenclature, sent in a standardized transaction,
asynchronous with the billing form. I’d go so far as to suggest that Medicare
should mandate that if a lab provider wants to be paid that they should have to
provide it in this format.

The value ranking, of the recommendations on through eight number one I
would have to say is the greatest need for the long term care providers and
from our standpoint, because of the point I’m making about our need, they can
be independent. I think that if you’re talking about the needs that were talked
about with others today you can talk about the need for them to be linked. But
from our standpoint if you could do one thing tomorrow it would be this.

And then some comments to close on recommendation one, I do think as I’ve
been learning about the CCR record, the continuity of care record effort, that
this contains the shell for what I’m talking about and I think that that would
be a way for us all to move together fairly quickly in this regard. So that
kind of closes my comments on recommendation one.

In going to recommendations two, seven, and eight, I think that you can
just kind of hear my voice in most of those giving you a ditto for most of
them. The emphasis is a little sense because I think the need is just a little
less in terms of our just day in and day out need for quality of care. But let
me give you some focused comments on seven and eight.

Mode of collection, again, going one better, what we, we already have
functional information on the MDS and so I think that one real benefit that we
could offer back to the hospitals would be that the MDS functional data
elements might be a useful starting point for a standardized set of functional
elements that would be able to be transferred back to the hospital when we are
transferring patients back and forth. And I’m sorry, I’m trying to remember
exactly what your proposal was but basically once standard dataset again
created and maintained by the provider of care, standard nomenclature, standard
transaction, again, you were talking again about linking it somehow with
billing, the billing vehicle. Again, I think you need to go one better then
that, it needs to be asynchronous, it’s fine if it’s exported to that vehicle
but I think for really effective use of this information it needs to be
asynchronous and in real time.

Some specific comments about MDS and SNOMED, there apparently has been a
crosswalk recently completed which between looking at SNOMED and it lacks
functional status elements in the area of ADLs, and so CMS is currently talking
as I understand it with ASPE about the fact that SNOMED while is a very useful
set of standardized set of data elements lacks functional assessment and we
think that we need to look at that and find a way to get those functional
elements into SNOMED. For right now we would recommend integrating SNOMED the
other way around, integrate SNOMED into the next version of MDS and then use
the current MDS data elements for the functional element gaps.

A few comments on the CCR effort and functional status. CCR as it’s
currently configured also lacks functional status but a crosswalk again was
recently completed to look at what it lacks and as I understand it the agency
for, the Association for, American Health Care Association and others are
working to develop what those new elements would be that go into the CCR. And
as you can imagine we will make sure that the elements map directly to MDS. So
I think that again you’ve got SNOMED, a very useful vehicle needs some
functional elements added to it, and then you have the CCR effort which is more
the vehicle, again, needs some functional status elements added to it.

So with that I’ll close, long term care patients are hospital patients and
the standardization that you’re talking about today I find very exciting
because it would actually help us not just improve our quality, which is
terrific, but would actually help us tomorrow to be providing the kind of
quality care we want to be providing. And also to let you know that we really
are committed to the end result of your work, transparency, accountability, and
the real time communication of information that this would afford.

Thank you.

Agenda Item: Provider Organizations – Panel 3 – Dr.

DR. HOCHBERG: I want to first thank you for inviting me, I want to thank
the committee for thinking of me. If I say anything that you disagree with you
can talk to Justine later.

Let me just say a little bit about where I’m coming from so you can put my
comments in context. I’m medical director for an organization that pulls
together six different hospitals and their physicians for managed care
contracting and we build data systems, we maintain all payer data warehouses
and other systems to support the management. You’ll see a theme through this,
one of our mantras is that data needs to be aggregated. I’m also the chair and
a board member, I’m chair of the Physicians Council and board member of Mass
Health Quality Partners, that’s a statewide collaborative in Massachusetts of
payers, physician groups and public agencies whose goal is to aggregate quality
data across multiple payers statewide and proceed on a path of public release
meeting very high standards of validity.

In that organization the physicians are absolutely committed to the data
aggregation piece and also to long term public release but I think as Nancy
alluded to also in the physician community there’s a lot of tension and debate
about what are the appropriate standards of validity, what are appropriate
timeframes, is there a period of internal quality improvement that is allowed
before data is released.

I’ve also spent time at a large corporation overseeing the building of data
warehouse and quality measurement systems also predominantly claims based so if
there is a claim added, a field added to a claim for the last five or six years
I’d have to figure out what to do with it when I got back to the office. And we
still do that today. I know a little bit about a lot of things so I’ll try to
touch on a few things as we go through.

Before we go into detail I think one of the issues that we have to be aware
of is that really in the past 20, 30 years we’ve taught physicians to ignore
data, the average practicing physician gets data from multiple sources. Even if
the elements are standardized the reporting formats and the report definitions
are not, and so the validity is usually poor, they don’t tie, and the ends
don’t meet reasonable statistical thresholds. And I think actually people don’t
come out of medical school hating data but I think after a while they learn
that most of the data they’re seeing from administrative databases is not valid
and they tend to just ignore it. And I think if these pay for performance
models are going to improve we also have to get physicians more brought back
into valid data.

The other problem with the data, several other problems, it’s lagged, so
it’s often profiling past performance. The second is it has minimal clinical
depth and that will speak right to the next issue, which is most physicians
don’t find HEDIS measures, that the administrative HEDIS measures particularly
compelling, I’m glad Peggy isn’t here, because to them the issue is, and the
classic example, it’s not whether you did the glycol hemoglobin test for
diabetes, it’s how you controlled it. Yet all the bonus programs and all the
HEDIS, everything’s claims based, it has to across these large networks, it’s
very unimpressive to physicians, doesn’t really generate buy-in. I think
there’s a little better situation in some hospitals who’ve used their
information systems to give good information, what’s going on in the
institutional setting, but it’s rare to have that in the broader outpatient
setting. The exception I think is some organizations like Kaiser and Harvard
Vanguard who have platforms to do that.

When I look at selected laboratory tests this really gets a double thumbs
up, clearly it would improve HEDIS measures and allow often administrative data
measurement with more clinical depth. I think also the incentive payments who
pay for performance contract models would also be better. And for some of the
HEDIS measures where people are grouping around 80 percent compliance anyway
the sort of difference in which people moving money around is not very great,
it may not well probably within competence intervals so I’m a little skeptical
of how dramatic that’s going to drive improvement.

Theoretically it should improve the accuracy of risk adjustment models, the
more inputs you put in of this depth you should increase them, I don’t know
that I can tell you the actual improvement in the R squared but theoretically I
think it absolutely would. High risk members in need of outreach is another
area, if you look at particularly some of the disease management programs
again, just to continue to use the paradigm of diabetes, those programs don’t
have the information they need to actually decide who to outreach to, this
would be an enormous movement in that direction and again would increase the
buy-in to physicians for claims based measurement.

On the down side since there’s never any free lunch in this environment it
would dramatically increase the volume of data captured in claims based
analytic systems and current systems would need significant time to prepare for
this. When I was at McKesson and we were looking at LOINC actually, which
hasn’t done much in a relatively long period of time, we were very excited but
we were fearful one day we would get a file with all that in it and what we
would do with it, we’d have to rewrite all our rules, we’d have to expand our
storage capability. Significant operational challenges if you reported
everything is a bit of a data explosion and even the risk adjusted models would
have to be reevaluated and retooled so there’s a lot of work that goes with
this to make this useful.

I think one thing to think about then is the gradual phase-in of measures
really according to what you think is most critical instead of putting all of
them in at one time, to just phase this over a period of years, start with a
few critical ones, give people time to upgrade their data systems, their
profiles, their acuity predictions and then move from there.

So to summarize this I think very, very high value implementation costs are
significant but I don’t think unreasonable at all actually, I think it’s more
of a time issue for depth rather then a magnitude of investment issue. I think
the overall balance of value versus cost would come out very favorable to do
this so I would very much encourage this to happen. And again I think it’s
really about how you phase and introduce this that’s critical to its success.

If we look at the second recommendation which is selected vital signs and
objective data, clearly at least from my perspective collection for reporting
will be burdensome, particularly in the absence of electronic systems. The
linkage to claim submissions I think would be problematic in a number of
institutions where they’re just not used to pulling that clinical data beyond
diagnosis into their billing system, so again, I think there’s a lot of
background work to make this happen in an automated way. And I know, at least
people would raise questions about reliability, I don’t know how standard the
blood pressure is from hospital to hospital, there’s inter-operator variability
and there’s also product and other issues.

One of the questions really is to defer this until clinical care IT systems
improve and this is captured more easily and then fed down to claims more
easily. The immediate value would probably be an inpatient evaluation of
selected procedures or care paths where vital signs are significant
discriminators around what you would do so I think there’s a smaller subset of
things this is going to relate to. It certainly would provide valuable
benchmark information but at this point I would think that the value versus
cost is not as certainly favorable as the first recommendation.

This charge diagnosis modifier or flag for present at admission, again,
this would certainly assist in risk stratification of cases and that would
improve the current quality assessment efforts and support more accurate
measurements from claims databases. Also to the extent there’s a lot of things
being sent in under pay for performance contracts, the more easily you can
capture acuity differences or comorbidities you need to do that, but these
systems again aren’t incenting properly.

I believe this would require changes to the hospital coding patterns,
overall cost to implement to institutions, I would think it’s probably low, and
the value of this, I think it’s moderate value, low cost, it’s probably
something that should go ahead, it’s low hanging fruit in another jargon.

The operating physician identifier code I actually found very interesting
to think about, it would certainly dramatically improve the ability to monitor
and profile individual surgeons by outside agencies or payers. I think most
hospitals capture this now so it probably does not add a lot of value to
internal hospital quality improvement, this is really going to be more for use
from external agencies. I think there’s certainly going to be problems with
adequate sample sizes and risk adjustments, particularly if you don’t aggregate
claims across different payers, so that’s one of the things I think you have to
think about with this. And I think claims based risk adjustments are really not
at this point prepared to take claims and risk adjust procedures and outcomes,
that doesn’t say they couldn’t do that but again, there’s a lot of statistical
and analytic studies to build those models, I don’t believe they exist at this

I think you’d also, if you’re going to, you need improved outcomes
reporting or if you know who did the surgery, but then you don’t really have a
very good ability to actually measure the outcomes and I’m not sure what the
usefulness of knowing who the surgeon was is. I think there’s another critical
piece to this. This would allow tiering of individual surgeons under pay for
performance model, which would scare some of my constituents to death, there’d
be a very high bar for statistical validity, in our organization it would be
about 80,000 feet up. The other question is what is the proper level to look at
things like operative mortality and performance of individual procedures, is it
at the surgeon level or the institution level, I think currently we haven’t
remotely scratched the surface on just looking at institutional performance on
these measures and to jump out and now try to look at individual surgeons to me
may not be the most rational pathway and the best use of time and effort. And
again, I think a lot of institutions have very robust internal quality
improvement programs and pay for performance as they’re going to be incenting
institutions around quality measures does get people’s attention and does drive
some of your focus in some of the things they do. So I think you can get at
this issue potentially without going to this level for now, I think at some
point down the road you may want to think about that.

Individual surgeon performance, a lot of that now is really within the
hospital and that’s the audience, if you move this to claims base then you’re
broadening the audience, you put it up for potential public release. A lot of
dynamics change in terms of right now this is a very, the walls of this are
around the institution, some of this is peer review protected so you have to I
think about how that relates if you made this change.

If you look at dates and times for admissions and procedures, I have some
accuracy concerns, I think someone mentioned, I can stamp a time that I’m going
into the room but I can leave a minute later and not come back for ten, so I
don’t know how accurate this would be of actually pinpointing at what point in
the process this was done. I could stamp going into the room and then not go in
the room for 20 minutes too. I think utility is probably limited to selected
interventions where you have outcomes clearly tied to the repetity of
interventions and there is a finite set of things for which that time sequence
appears to be critical, so that means this has sort of a finite world in which
this would have a lot of value.

I think the other thing is if you were going to evaluate this you’d have to
collect confounding variables. Best example I can give you of a surgical case I
know of last week where someone needed to go to the operating room immediately
but actually needed a transfusion first, so their time to go to the OR was
delayed more then it should have been for perfectly adequate reasons. Now if we
were just measuring time and date and not that confounding variable that would
have looked like a quality problem when it probably wasn’t. So I think if you
go to this you have to think about what other data do you need, so in essence
risk adjusting or adjusting these cases so you can interpret the data.

Episodes start and end dates, I think useful for evaluation and very
selected episodes, in the report it talked about prenatal visits. I don’t know
that there’s a very big universe for this so I think utility is actually quite
narrow here. And again, I’m a little concerned about the probable accuracy of

Functional status codes, analytic models clearly exist and a number of them
like HIMSS and others are well validated. I think large scale collection of
this at any reasonable cost is probably contingent on electronic health record
adoption, and I’d like to say that’s coming in the next two to five years but I
think it’s really a lot longer. It’s really only some of the integrated systems
now in hospital physician systems who are making these investments, there’s
really no economic model right now for most individual practices to make this

The other thing is, and this was touched on earlier is it’s not just
electronic health record adoption, if you’re going to collect this you need to
have standardized assessment and data input. They’ve put forms in these systems
and that standardization is no where in existence amongst the current vendors.
So I think that would require some kind of government action or standards or
some agency to take on standardized fields and enter in these systems and that
seems to me quite a long time to get there. So I think this is probably out of
scope in the near term.

And I think that’s it, thank you.

Agenda Item: Provider Organizations – Panel 3 –
Dr. Kazandijian

DR. KAZANDIJIAN: Thank you. First of all I’m delighted to be here and see
some faces I’ve known for the past 20 plus years in this field, I’m delighted
to see that the excitement is still there and that the topic is also sometimes
still there.

My presentation is probably going to be slightly different from the others
that we have heard so far, I have only three slides and I would like to put my
presentation within a context as to where I’m coming from. I do wear a number
of hats myself, most appropriate for today’s presentation are a couple of
those. I’m senior vice president of the Maryland Hospital Association and in
that sense bring the association’s issues, the relationship issues on a smaller
scale then AHA obviously, more regional. I’m also a president of a center
affiliated with the hospital association that is responsible for international
research only. And I used to say that I have done international work for the
past 35 years and that’s where I picked up my accent. So that’s one of the
problems with causation and correlation but still an important one.

So a lot of the work that we do in six countries deal with performance
measures and accountability so I will bring some of those issues as well. And
finally, I’m certainly interested in bringing some of the academic rigor to
this and we have had that as a goal in our work for the past two decades or
more. So that’s basically the context in which my presentation will take place.

I will not necessarily go one by one in those recommendations but I will be
bringing some issues that I think make sense coming from the field if you want
to address the issues that you raised and make some comment as we go, some
directly, others indirectly, I’m hoping they will be addressed.

So the first of the three slides is the question of the business need, is
there a business need for this report, which I would like to join the previous
presenting group as well as us to say that it’s very timely, very thoughtful,
yet deals with issues from an angle in some situations that I would like to
challenge if you want, if nothing else challenge the angle itself and say can
we take it perhaps to another level, will that make things easier.

From a business point of view the whole concept of measuring performance
obviously is of importance, the distinction I’m trying to make is measuring
quality versus measuring performance, which has created a lot of difficulties
in the past decade when projects or initiatives dealt with quality as the term
and promised to measure quality, it did not work, as well as when it’s
measuring performance, which is a value free concept. So looking at the time
for example as one of your measures, in itself if you say this is the right
time then it becomes quality, but if you say this is the time and this is how
people are doing it it becomes comparative analysis. And I think that’s an
interesting distinction.

The Maryland Quality Indicator Project is where I’m sort of building some
of my comments, I’ve been fortunate 20 years ago to be part of the first group
to design that project, I’m responsible for that project even today after 20
years. It’s the largest project in the U.S. as well as overseas. And one of the
interesting things there regarding performance measurement is actually the term
performance rather then quality. And I’ve spent the last 20 years telling
people that the title, the name of the project is misleading, it doesn’t
measure quality, the Quality Indicator Project. But that was the thought 20
years ago and then we learned better I think. So that’s an interesting issue.

The concept of benchmarking becomes important I think when you start
talking about pay for performance and I’ll come back to that, but under that
rubric I think the distinction between incentives and rewards becomes very
important. How do you deal with that concept, do you provide incentives for
those who are not doing well or do you reward only those who do well? Do you
increase the gap between those who can or cannot by doing that? It is still an
open question.

So the transformation of performance to quality becomes an important
question and I think that’s where the challenge is with some of those measures
because at least in our experience, and in my personal view, the moment you put
a value on a value free measurement it becomes quality. If you say waiting time
of 18 hours is okay then it becomes okay, if you say it’s not it’s not, but you
measure waiting time the same way in both situations. So the issue of quality
is that transformation by putting a value on it and that in our jargon, that’s
what we call evaluation, which is putting a value, so evaluation is a
transformation of performance to quality and that in itself is a different
question, who does that evaluation, who does that transformation, is it the
user, is it other stakeholders, is it some other agency from outside, whoever
is going to put that value on that measured performance is going to make a
difference in how its adopted and I think all of our recommendations have that
inherently, have that in them inherently as an issue.

Again, I discussed that it works for better performance, it is definitely a
business need and all of your measures I think fit into this. So overwhelmingly
I think the business need is there, the question is how do you define it. So
it’s one thing, I think Nancy was framing it as well, it’s one thing to say
yes, it’s another thing to say how do we measure it, the contextual issues and
so on and so forth, so the reward issue becomes a very important one. And it’s
not a trivial issue because the reward is directly linked to evidence based
medicine concepts. If evidence base is the standard, is the goal, then it’s a
reward system. You meet or you don’t meet, if you meet and you get kudos, if
you don’t meet you need incentives, not rewards, to get there and that should
be part of the business need as well and it’s important to see if those
measures can accommodate the concept of incentives, which is my next point.

The incentives towards better performance, at least I’m following my own
logic which is good, incentives towards better performance become very
important because again, as Nancy was saying, just looking at the hospitals but
it deals with long term care and home care as well, the spectrum of size and
therefore resources becomes a very, very important issue. If you look at the
typical American hospital, it’s a mid-size community hospital no matter how
much of the papers within the medical journals come from large institutions
they do not necessarily present unless they are dealing with a generic issue
and I think that’s where the challenge is, how generic is this to be portable
and adaptable.

The next question regarding the business need, which I think is something
that is important that I did not find here and perhaps it is imbedded in the
logic but not necessarily in there is the question of ongoing monitoring. What
is the logic behind the ongoing monitoring? Because we all know that ad hoc or
short term whenever you throw something that projects people know there is an
end date to it, and when it becomes part of the fabric of life then it’s a
different thing, it’s not a project, it’s a shift in mentality and I think
that’s one area to think about. Is this going to be a project? And if it’s a
project people will give you what you want, people will give what they can, but
then they know it’s going to end and they can go back into doing things if
that’s the reality of it. So the ongoingness or the continuous approach to it
is very important.

And it is also important from a quantitative point of view because the
adjustments which are going to be necessary to make, to take those measures,
evaluate them, see how they are behaving over time, recalibrate them, it’s
going to need a long term ongoing monitoring. And that monitoring in itself is
going to be effected by questions of pay for performance as we said,
accountability for sure, transparency, public disclosure, and so on and so
forth. So that’s one issue.

My second slide deals with the risk adjustment need although risk
adjustment is a distinct recommendation here, I think it has implications for
your other measures as well, and I think it’s important to look at it perhaps a
little bit closely. The distinction may be important to make here that in fact
the adjustment concept has two components relative to those measures. One of
them is that it’s immunological in nature which deals with stratification
basically, that’s what it is, in epidemiology it’s called stratification, other
people call it adjustment. But you compare things that are similar basically,
you put them in pigeonholes that are comparable, so that’s one aspect.

The second one is the clinical adjustments, which we heard that it goes
beyond risk, it goes into actual acuity adjustments which I think is very
important because risk in itself deals with something that could happen whereas
acuity is what is there already and how you measure it, those are two very
distinct concept that cut across again.

From the epidemiological certification point of view if you have a long
term monitoring system it may not be, and here I come after a full day of
everybody agreeing on risk adjustment, you needed one person to say maybe not,
so I’ll try to do that. Risk adjustment I think is very, very necessary,
clinical adjustment is very, very necessary, one it is a short term analysis.
When you have a short time span, six months, a year, to look at something and
try to minimize any variability in your measurement that may be due to things
then actually what it is.

When you’re looking over time, when you’re looking over time at stratified
data, epidemiologically stratified by demographics, by minimal stratification
on health status, some of those patterns will emerge, there is enough
literature on that that they will emerge and it is an interesting challenge how
to marry those two concepts, I don’t think it is necessary to throw the
epidemiological stratifications with demographic stratifications that are not
at the disease level, as clinical adjustments for severity are necessarily
going to be, yet they do provide good insight if the design of the approach has
a longitudinal component to it.

So I would say that adjustment in general is actually the case to be made
here and I think all of your measures would benefit from that. But I will maybe
refine the term risk adjustment or clinical adjustment as the only type of
adjustment and include into that epidemiological stratification as well.

Needless to say there’s a whole industry, there has been a whole industry
in the past decade on looking at stratification and people have tried different
things. Some of them have been so difficult to interpret that they have
actually taken, provided tough times for some very good projects and the
Medicare Mortality Report mid-‘90’s definitely is an example of that where
the stratification itself made it almost impossible to be able to interpret it
among those who understood the stuff let alone what it could have done for
outside groups. So in that sense the business need for adjustment and the
provisions by the business field to support adjustment of those measures may be
a given if in fact it is provided in a way that accommodates for those things.

Then there’s a scientific need to uncover and discover and improve on this
and therefore I think it is important to have that long term commitment of
monitoring because as we uncover things and we tweak with them and we change
that, we heard both Peggy and Jerod mentioning that in their work and certainly
we have done that and continue to do that in our work, the scientific needs,
the measures should be able to accommodate for those scientific changes and
novelties in how to do things. And I do think that, I will come back to that,
except the one that seems to be relatively controversial to everybody in your
recommendations which is the physician identifier or physician information,
everything else I think can and could keep pace with the change whereas that I
don’t think can in this area or do a very good job.

Health service needs, I’m going to try to first just slightly different
again because I think we are at crossroads on this and this may be a wonderful
opportunity with your recommendations, with your support, with your guidance,
to actually take it to a level that people have shied away to do it. Health
status has always been discussed in this kind of context of performance
measurement as the health status of the patient when in an institution or when
they return to the institution. But to link the discussion between process and
outcome that you have raised, this afternoon it was raised as well, I would
suggest that perhaps health status is the ultimate outcome when it comes to
public understanding of quality. And if you think about it health status is
what you’re supposed to change if you did well. You’re going to change it if
you didn’t do well but again, it is still health status.

So in that sense health status perhaps should be looked at and promoted to
go beyond the hospital, beyond the nursing home, beyond the institutional
boundaries. Why am I saying that? If you look at the most frequently done high
cost procedures in the U.S., and other places but primarily in the U.S. but
western Europe is catching up very nicely with it too, they are mostly
electric. If you look at hysterectomies, if you look at prostatectomy, if you
look at C-sections, if you look at evascination(?), and if you look at
orthopedic surgery, unless they are for cancer or for pathology that have
nothing to do with functional status or quality of life of the patient there’s
always that discussion between the provider and the recipient as to why they
are doing that. And functional status in that situation is the ultimate
outcome. Is it different from those both percentages or is it ability to walk,
that whole discussion becomes very, very interesting and I do know it goes
beyond perhaps the scope of looking at the databases that we’re talking about
here, but start putting the bug in people’s ear that health status in itself
should not be only in that sense but also part of accountability and the true
outcomes, maybe an interesting, that’s an area we’re embarking on now in our
work trying to actually link information on health status in community when
patients are discharged to the care that was provided or to the education it
was given, it’s a delightful challenge.

It does make sense to community, and by community, I use community in a
very lax way and a very generic way and we have the communities of payers, we
have the communities of recipients, we have the communities of potential future
recipients. And we have the communities of folks that actually put policies
together, so that’s, it does make sense to practically everybody. Purchasers
primarily if you look at productivity as being one of the measures when people,
when you go to a business group and say here is a model in order to improve
performance or improve quality, one of the first question is of incentives and
measures, is measures of productivity and so on and so forth, so it does have
an inherent attraction to the people.

As I mentioned I want just to be on the record on this, I know it is
probably before its time and probably before even out of context slightly to
today’s discussion, but I couldn’t resist to discuss health status by saying
hey, thinking a little bit out of the traditional circle may also benefit the
cause because for me that is really the ultimate outcome.

And that’s where a lot of the interesting comments come about outcomes
research, when people talk about outcomes research and they say it’s limited to
3.4 four average length of stay days, those are outcome measures or output
measures or whatever you want to call them but they do not have in there the
issue of health status as much as perhaps we should, especially with the aging
population and some of the chronic conditions.

So to end my presentation I would like to bring a couple of thoughts to the
table. From our experience years ago when we were challenged with this kind of
issues, mid 1980’s when there were no projects, there was no initiative in this
field, when we started the quality indicator project one of the challenges was
how do you describe the actual nature of the indicators, of the measures. And
the best way we have found over the years to explain it to various audiences is
the follow which I think also fits into today’s discussion.

We have called the indicators pointer dogs and we have said that they are
basically, they do exactly the job of a pointer dog when you go pheasant
hunting or whatever hunting you do. That in fact a good indicator is a valid
dog that points to a pheasant when you’re hunting a pheasant not to rabbit. But
the point is I can give you the best dogs in the world, you will never get the
pheasant if you’re not trained how to interpret that pointing or to know how to

So the real success of this is in training the user, in training the
hunter, of this information and not giving people the best dogs. And that has
always been the challenge, organizations that say we got the best dogs and put
it out there and people say how the heck do I interpret when the tail goes to
the left, well, that’s where the challenge comes in. So my sort of, and by the
way that’s, I wrote an article on that in 1990 and the word pointer came into
the literature coming from that article, it was absolutely fascinating, it’s a
very simple imagery.

But again, it brings the issue back that the approach should not only be is
this going to be part of the database, is this valid, is this reliable, is
this, that aspect is important but that’s the dog. How does it translate to the
education? How does it have a parallel program for the users? And the users are
the ones not only health care users but the communities, because if
accountability is part of the model then the community should be able to see
and make a difference and I think thinking about that, at least bringing that
to the forefront as the next topic will be a wonderful thing to do.

Those are my comments and again, I do thank you for the opportunity, it’s
delightful to see that this issue is still very hot.

MR. HUNGATE: Thank you, having the pointer work means the information is
actionable in some other language and I couldn’t agree more. I know that the
lady who sits two doors to the right of me here would like to comment on the
health status discussion.

MS. GREENBERG: I couldn’t agree more and again, I think the nature of this,
the recommendation on functional status which actually the World Health
Organization would broaden to call health status, which is kind of a
combination of functioning and maybe health conditions as well, is to do more
research and identify standardized ways of capturing this and sort of to
facilitate eventually being able to say something, say comparable things about
health status or functioning. And then when that should be collected and where
it should be collected, I think you could see collecting it at admission if you
want to use it as a risk adjuster but certainly generally when people get
discharged so quickly out of hospitals now that they’re in worse health status
when they leave probably then when they got there and it’s certainly not in the
improved health status that you would hope they’ll be down the road.

So I think these are all a lot of questions but I appreciate your
recognition that it may be the outcome that makes the most sense to consumers
and ultimately purchasers, we heard that about back to work, etc.

Though I would agree with whoever made the comment about process because I
think, that can also be an important way of communicating to consumers because
I think it has a real educational benefit too and the more you know about the
what the direct processes are, the best practices, you have to be an advocate
to yourself and your family in the health care environment and that’s helpful
to people and it helps them to make distinctions.

So anyway, I thank all of you for very provocative testimony and I’m really
chewing over the issues that Barbara has brought before us because you’re
absolutely right that you need this information in real time and that putting
it on the administrative transactions will not get it to you in real time —

MR. HUNGATE: Won’t do it.

MS. GREENBERG: And there we get back to either the electronic health record
in all its glory or maybe pieces of it which is I think what you were
suggesting, but as we all know that’s not here yet.

MR. HUNGATE: Thank you and we’re open for discussion.

DR. COHN: I actually have a question for Marjorie, I don’t mean to put her
on the hot spot here but let me just ask her, there are actually two
recommendations relating to functional status and I don’t know whether, I think
we are talking about functional status though, health outcomes and health
status is obviously another piece of all of that. But one of the
recommendations is sort of to study what the right answer is and in 2001 the
NCVHS said the government needs to study this, and then 2003 CHI said we need
to study this. And then we have another one that says go out and do something.
I guess my question is has there been any study either in 2001 or based,
actually it was 2004 recommendations I think, or is there a plan to study

MS. GREENBERG: Well, some study was done through the CHI process and it
was, because it was done basically, not with a detailed research project or
resources or whatever it was inconclusive which is why the CHI process, the
Consolidated Health Informatics process which identified and now many have been
adopted by the departments, DOD and VA, standards for various fields, they did
not adopt a standard for functioning and disability because they felt that they
weren’t there yet.

I think what you’re, so some study has been done but no, I think that a lot
of, and there are some individual projects out there but a lot of resources
have not been significant, certainly not adequate resources have not been put
into the research that was recommended by the national committee —

MS. HANDRICH: If I may just add to that, I would venture a guess that in
the field of long term care what study that has been done on specific measures
for functional assessment has been done by state Medicaid programs.

DR. COHN: The reason I’m asking this is because actually, maybe it would be
a little more clear, 2001 they said geez, ICF, which is the WHO —

MS. GREENBERG: They said it was the most promising —

DR. COHN: Promising but it really needs further study, and then in
2003-2004 when it was looked at again by CHI what they said is geez, there’s a
fuzziness in this whole area, we don’t know what questions to ask much less
what answers to expect. And a code set is great, if you know what it is you’re
doing a code set can fill it in, but if you don’t exactly know what questions
to ask, or if there’s no agreement on all of that which is the LOINC code
issue, it’s hard to figure out what the code set ought to be or what it ought
to be doing. And so I guess I was just sort of wondering, obviously it seemed
very premature to create a mechanism for recording if we don’t know what the
questions are or how to answer them and so I was just wondering around the
basis of the CHI piece whether or not that somehow needs to be expedited. Am I
misstating it?

MS. GREENBERG: The response to that would be that there is some functional
status information currently being collected through administrative processes
but it’s basically free text and so to have the capacity to translate that into
some standardized coding which might then be, one could do something with a
little bit better then free text, is I think partly behind that recommendation.
But it may, even then the standardized coding has not been agreed to at this
point as to whether it should be ICF or SNOMED-CT, and I will say to Barbara
that the clinical terms that have been added to SNOMED do include a lot more
information about functioning and ADLs then SNOMED basically didn’t include
much of anything. But there was actually a real effort by the people in the UK
to make sure that a lot of the terms from ICF, International Classification and
Functioning Disability and Health were added to the clinical terms.

Then when the clinical terms got married with SNOMED SNOMED-CT tends to
have, it’s not adequate at this point but it has more in it, and again, it’s
the question of a terminology, the whole continuum between terminologies and
classifications. So SNOMED-CT is actually, if you haven’t looked at that you
can see that there’s been more added then there was in SNOMED which was very

But Cathy, do you want to comment on number eight, recommendation number
eight, which is made in conjunction with the need for more research and whether
you think, do you want to come to the table?

DR. COHN: I mean seven seems to be pretty reasonable, I just was trying to
figure out how you could do eight given that we need to do a call for some

MS. COLTON: Well I think they were intended to be linear rather then
parallel but I think it was also recognized that the process for adding new
data elements is a very lengthy one and as was pointed out here that the
process for implementing is also a lengthy one. So once a data element gets
added to a standard, a transaction standard, there’s a period of time for
organizations to gear up to implement that standard. And so that if we’re
looking forward and saying when is the earliest time that such a code could
actually be included in the standard and then when would it actually get
reported, one could conduct the research during that period of time. It’s quite
a lengthy period of time.

MS. GREENBERG: It’s an enabler, for example the institutional standard
allows collection of this flag for diagnoses, well, one state or two states
needed it so they were able to get it into the standard but I mean everyone
agrees now it’s kind of a low hanging fruit, it’s been recommended since, the
national committee recommended it back in 1992, only 12 years ago. But one
reason it is low hanging fruit is because the capacity is there, but now if
you’re going to start to try to get it into the standard it would take you
another two or three years. So there’s a reason to, I think back to the 2001
report said that standards organizations should be alerted that the national
committee considered this an important element. And so I think I would
definitely agree with Cathy that there’s, there’s time to do the research and
maybe put in the capacity in the transaction standard would actually get people
to start doing the research, who knows, but there’s certainly time.

MR. HUNGATE: Related question, there is this uncertainty about whether the
standards are there or not but I have the feeling that in the rehab community,
that across that community that they do have their own set of standards that
they have developed —

MS. GREENBERG: Sure, they have to.

MR. HUNGATE: They’ve got to, and I wonder if that isn’t true in the long
term care community, that there isn’t some set of measures that have evolved
that are fairly common. And what worries me in this discussion is that we try
to get something that covers everything when it means that people that could
things now have to wait until its standardized for everything and that worries
me and maybe it’s just lack of understanding on my part.

DR. PAUL: I think comments on this were hesitant because I think you guys
have more depth of knowledge then I do, but in the nursing home setting right
now we are assessing functional status, right now we are struggling with the
fact that a hospital based physical therapist uses terms different then a
nursing home based physical therapist for certain aspects of functional status.
Right now we are working with CMS and ASPE and others to try to take MDS to its
next level, to take MDS to incorporate SNOMED, to take SNOMED-CT to make it
better for functional status, all of that is happening right now. What I don’t
know is how that effects the world that you’re actually living in but right now
it’s happening for our world.

DR. COHN: Barbara you bring up a very good point, I think that the reality
is is that this stuff is being collected and I think you made a very salient
point which is that different people ask different questions, or they ask the
same questions and they mean different things. And I think that when, I’m just
reflecting on what the CHI did around all of this and it might be very useful
for the Quality Workgroup to ask for a briefing from the people that were in
charge of this area to sort of get a more in depth feeling. But there was sort
of confusion a, because there were a whole bunch of questions that seemed to be
sort of the same but sort of different asked on sort of different forms in the
federal government about functional status, and then there’s sort of the
question about whether you leave them all the same as different or whether
there’s some way to standardized them and do a smaller group. And then of
course once you have that then the question is to try and standardize the
answer, which is really where SNOMED or ICF or whatever sort of may fit in. But
it was apparently a complex enough issue that that was sort of held for further
study as I understand and it’s why that LOINC issue as well as the ICF/SNOMED
issue came up. And it sounds like you’re working on pieces of it just within
your work at this point.

MS. HANDRICH: I would agree with the comment that you made but the long
term care population being defined as the population that is over age 65 or
persons with chronic disabilities desires to live in the community and the
science for measuring functional status for the non-institutionalized
population, which is where everyone wants to be, isn’t there. So that’s the
area of opportunity that could be identified as a priority and pursued perhaps
with some endorsement of this committee and others and they are the very people
who go in and out of the hospitals and may have short term stays in nursing
homes but the whole trend of long term care life is for people who may formerly
have been in nursing homes to spend much longer periods of their lives, if not
all their lives at home with intermittent stays in nursing facilities. Would
you agree with that Barbara?

DR. KAZANDIJIAN: If I may just kind of comment on that, I fully agreement
with your overall assessment and I think the reason it’s attractive is that it
brings back the issue of the continuum of care and it brings back the whole
concept of linking the different sites to different activities, to different
outcomes. And I find that the opportunity to use your recommendations if it’s
possible, at least one recommendation that goes beyond one setting, may provide
very good incentives for future developments.

DR. PAUL: I’m sorry, I can get home tonight if I get to BWI. Thank you.

MS. GREENBERG: I might just if I can say Barbara that I know the committee
has a lot of interest in pursuing long term care, quality issues related to
long term care and certainly we can call upon you again.

DR. CARR: I’d like to just integrate all the very wonderful input that
we’ve had today but I want to make sure that I understand it. I think this
morning we heard a lot from JCAHO, CMS, Leapfrog, quality groups about the
tremendous opportunity that lies ahead if we can just get quality going or
something. And this afternoon I’m hearing I think from you of the great risk
that lies ahead if we go in the wrong direction or without careful thought. And
I just wonder, and I think you’ve all sort of spoken about it, but if you could
leave us with one thought of what is the biggest mistake that could be made.
I’ve never been to one of these hearings before so if we could get, I mean
everybody likes apple pie and motherhood but what’s the biggest danger that
lies ahead in this initiative?

MS. FOSTER: I’ll start with one but I’ll just go back to the previous
conversation and say that I was excited as Vahe was about the opportunity to
look at the continuum of care and for hospitals and other organizations to be
able to get feedback on the outcomes of their patients or the functional status
of their patients well after they’ve been in the institution, something they
don’t get right now.

One huge danger, I’m not sure if it’s the biggest because I’d have to think
about this a little bit, but one huge danger is that if you identify the wrong
thing to measure you are going to push people towards doing the wrong thing
clinically. We’ve seen examples of that and whatever you would call it, the
continuum of that is if you don’t have a plan for continuously updating your
information, your requests for data to keep current with current science as
Vahe was saying that you will keep people doing something that is perhaps not
the most effective clinical approach to care. And if you want further
elucidation on that I’m sure Jerod and others can talk to you ad nauseam about
ARBs and other things that are sort of at the point where we’re measuring
something that’s not exactly what clinicians believe should be the right thing
to do.

DR. HOCHBERG: I would add another comment which is that I think all the
focus on measurement also has to create the business case or economic
underpinning for institutions and practices to make the investment to success.
I can share a vignette with you, I was at a meeting about a hospital that was
going to tier people based, hospitals based on Leapfrog criteria, particularly
computerized order entry. And a person next to me was an administrator of a
small hospital said God, for a year we’ve been talking about the need to find a
way to gather money to invest in things like this and if I get this thing right
now I’m actually going to lose money because I don’t have it, that’s the exact
opposite of what I need to go forward. So I think this has to incorporate some
provision of tool system investment otherwise people scramble in relatively
inefficient ways to do this and don’t get where we need to be.

DR. KAZANDIJIAN: I don’t know the answer to your question but just an
observation, I think health care has shown that it likes to be a very wide
swinging pendulum that goes from one to the other and rejects things in
between. I think one of the most dangerous things that could happen is to focus
only on the systems concepts. We heard a number of people today, it’s very
politically correct statement nowadays regarding the safety situation and so on
and so forth that it should always be looked at, only looked at as a system
issue and I think it’s an important component of how it should be approached
but I do think that health care is still delivered to individuals. And you
heard Nancy’s example about the 99 doctors or whatever it was and also systems,
if you shut your eyes for a minute and try to see a system what do you see?
Most people see people, they don’t see processes, they don’t see IT things,
they don’t see structures. So I think we do still have to learn what the system
is and to put everything into the system may actually cover or allow
individuals to believe that there is nothing that could come back and haunt
lesser performers, lesser performance. So one of the dangers I see in all this
is the swing from the 1980 small area variation analysis that we all remember I
think where the only goal was to look at physician practice styles, remember,
that was the term, change the physician practice style you will decrease
variation, ACPRs(?) concepts were based on that back in the 1980’s. Then we
swung from look at the physician to look at the system without mixing those two
and I think it is perhaps a cautionary note that looking at outliers and
inliers at the same time is better then just outliers or only inliers.

MR. HUNGATE: I have a question I think is related and it’s addressed both
to Vahe and Stan really. You both participant in project that people have
joined to improve quality. What data is used in those? Is in either case is it
claims data or is it always some other source of data? So what do your projects
depend on and how do our recommendations effect those projects specifically?

DR. KAZANDIJIAN: That’s a very fair question and I think it depends on the
nature or the reason for participation. For example in our case many of our
projects are projects whereby the measures as well as the data elements are
designed by us and the participants based on consensus using all the necessary
steps of literature reviews and consensus and expertise and all that. But it is
de novo type of data many times and not using existing databases just because
some of the questions could not be answered by those data. Now that’s a
voluntary component whereby organizations have reached their internal culture
if you want of saying we would like to know more and then voluntarily would
participate, so that’s a totally different type of people or organizations.

Then there is the more mandatory component of things and for example in our
work we are the largest vendor for Joint Commission in the U.S., more then 700
hospitals have chosen us based on our relationship with them to be the vendor
for the core measures for them. So we collect the core measures as defined by
the Joint Commission so in that sense also our participants use information
that are there with process that we built, IT components and so on and so
forth. But that distinction of voluntary and non-voluntary is important.

MR. HUNGATE: Is the non-voluntary portion, is that a dataset that the
hospital has to generate as a separate dataset from all the other datasets that
they do? Or is it an offshoot?

DR. KAZANDIJIAN: That’s a fair question. In many situations it goes back to
something Jerod said which is when you look on the surface the titles of the
measures are the same, a C-section is a C-section and so on and so forth. But
when you look at how the data elements are defined there’s always a difference
someplace excluding the denominators that throws the whole thing into a loop.
So there is some overlap but in many situations there are some parallel
databases. And the reason they continue to do that after 20 years is really the
whole question of the business value or the value itself because they see a
value in that as well as in this and they continue doing that and I think
that’s an important consideration. The validity from that point of view, was it
able to make a difference.

MS. FOSTER: Could I just clarify Vahe’s point, the Joint Commission, that’s
done basically by individuals in the hospitals going back and abstracting the
information from charts.

MR. HUNGATE: So it’s all chart review, labor intensive —

MS. FOSTER: Labor intensive chart review, somewhere estimated between 25
minutes and 45 minutes per record —

MR. HUNGATE: So it’s an expensive dataset.

DR. HOCHBERG: Well I would just add that all the ambulatory stuff is claims
based and all the pay for performance programs need to be off administrative
claims, they can’t incorporate chart review, they don’t have the infrastructure
to do that. So everything keeps devolving back to the HEDIS measures which are
relatively good administrative databases and the more challenging measures
which involve chart review don’t end up in incentive programs or as part of the
drive toward improvement.

MR. HUNGATE: So it seems to me that our challenge in this agenda is partly
to try to see what’s efficiently done by the relatively cheaper administrative
information then the relatively expensive chart abstraction. And so in a way
it’s saying what could you gain in making your program better for your members
then now exists from this data. And that’s hard for me to get a handle on —

MS. GREENBERG: [Inaudible.]

MR. HUNGATE: Well, to the participants because they have to pay for the
chart abstraction right?

DR. COHN: But I was also going to add the point that you have this claims
transaction or whatever and obviously every time you add a data element to it
that costs money too so you can’t win —

MR. HUNGATE: I understand, but the magnitude of the different costs is part
of the big business case isn’t it?

DR. COHN: Yeah, exactly.

MR. HUNGATE: And I’m saying I don’t know how to grapple with that, I don’t
know how to get at the real elements of that. See what I mean? I think it’s
important but I don’t know how to get an answer.

MS. FOSTER: And if I may add it would depend on what you’re asking to be
added to the administrative claim, because if in order to get that data someone
in the hospital has to go back to the medical record and abstract it you’ve got
that same cost and then there may be additional costs associated with the delay
in the bill.

MR. HUNGATE: Then my sense is, my hypothesis is that we have to kind of
come to a tentative this is what we think we heard and ask for feedback along
the lines of what I’ve just articulated. Is that a rational approach to the

DR. COHN: I think it is, I though it was actually also informed by
hopefully work that AHRQ is going to be funding or in the process of funding so
I think that actually, I guess from my view it seemed to me that if we identify
things that are very obviously low very low hanging fruit, that cost little but
have some value it becomes very easy to say go forward now, some things we may
say geez, we’re going to know over the next year what the cost benefit is on
all those —

MR. HUNGATE: So my mind says that we want six or eight specific low hanging
fruits, if we could identify, there might not be that many, I’m an optimist,
may never get there. But let’s say that I’m trying to model what we’re trying
to get to, so I’m saying let’s think about, that’s what we’re trying, where
they fit in the categories of candidate recommendations so there’s a specific
that relates to recommendation one or a specific to recommendation two.

DR. COHN: Next step?

MR. HUNGATE: Lab value measurement, which enables a performance measure,
where it is cost effective to do it and where providers would believe it was of
value to them.

DR. CARR: Okay, again, I’m new at this so this is like a really dumb
question. But where does this information reside? So for example Stan’s data
that are claims that go to Harvard Pilgrim or TOPS(?) or wherever they go, or
Blue Cross, they go to those places. And so the creation of the field and the
abstraction of the data elements is what we’re talking about but it goes
somewhere and either is used or is not used. So again, closing the loop on the
cost/benefit, there has to be a recipient place and as we heard also today that
the measures that we see coming out of different groups will have perhaps
different subscribers. So I guess I’m a little fuzzy on after we’ve done this
work and standardized it where it goes or would it also reside in the hospital
data set, I guess it would.

DR. KAZANDIJIAN: I think it would depend on the right issue, first of all
is where do you get it from, and the second one is where does it go once you
generate something out of it. I think where you get it from is, it is in part
to medical record obviously but not exclusively, because if you’re dealing with
process measures, if you’re dealing with issues of systems as people would like
to define in many situations, that is not in the medical record. That is in
understanding the process, in understanding the flow, and understand the
interaction, understanding basically the fishbone diagram, understanding step
wise where things go. It’s an algorithm, it’s purvical(?), so that doesn’t
exist. So there are no data sources for those kind of things, however they have
to be established if in fact that is an important thing, it has to be
established as a new system. Years ago that was tried to be established in the
activities around clinical pathways if you all remember that, it was the term,
everybody was doing 700 pathways in their hospitals, and those pathways were
painstakingly drawn and everything, and again, they missed things that didn’t
work well and today you don’t see them as often so it is important. But when it
goes, where does it go back, it depends on what the message was, if the message
was change the system, change the communication, it has to go to a different
group then you’re not applying, you’re not giving aspirin, when you have to
give you’re not applying evidence based medicine then it goes to someplace
else. I don’t think there’s an answer to that before we know what the question

MR. REYNOLDS: I think a key thing is, and we’re going through this a little
bit in e-prescribing and we’ve had many hearings, we’ve heard one day of
hearings from one set of people and we’re trying to, I sense we’re trying to
jump to an answer or jump to how many fields we can use and where would they go
and what would they be, and I think regardless of what we do is if there is
some low hanging fruit we still have to I guess to maintain a discipline for
ourselves for one or two more steps after that to make sure that whatever is
approved or whatever is decided, if it fits in the current structure and can be
captured in the current structure then it’s a good tactical move. If it is
something that creates a new way of doing things and there’s no structure that
we recommend to put it into place so that when you, if you pick the first five
things you want to do, and I think I heard that from a number of you, you pick
the first five you want to do and then we end up turning out in the end we want
25, and after we do the five we’ve got to redo the whole thing again for the
25, it just looks like we kind of jumped off the cliff so we could look like we
said we were in the air but we’re making a parachute on the way down if we’re
not careful. So I think it might be a little —

MR. HUNGATE: I hear you and I don’t disagree. My experience in industry I
guess is the best way to say it is that we had some people that could follow
the ready, aim fire, we could have some people that just fired, fired, fired
and finally got there, and we’ve got some that aimed, aimed, aimed and never
got there. And I guess I have a bias to try to do a little firing at a target
conclusion if you will and put my mind on the table and let everybody hammer at
it because I feel like we need to have something to work from to move from the
range of the nebulous to the specific where we’ve been called upon to do the
specific and then do the shooting at it that tests whether we’re going to have
to do redo.

MR. REYNOLDS: I agree, that’s what I’m saying.

MS. FOSTER: I’ve never been in a medical meeting that talked so much about
shooting, it’s a good analogy. The second thing is or as you ponder this
question, as you hear more testimony, I would encourage you to think about
costs, about value, but in each case asking to whom or to which parties because
it’s likely to be of cost to more then one and of value to more then one, and
to Justine’s point is that if you have constructed the perfect data collection
you know it’s of value to someone but you don’t have a pathway of getting the
information back to that group of individuals who would find it of value then
you’ve missed the boat too. And all of that is critically important as you go
forward and in most of the data collection activities with which I’m familiar
there’s been some dissonance between who bears the cost and who gets the value.

MR. HUNGATE: I understand. Did you have a comment Stan? It looked like you
wanted to say something there.

DR. HOCHBERG: I think also you just have to think about you’re going to
drive hospital system investment and other investment when you make these
standards someone is going to pull them for reporting and they will be incented
on you put them on the table. So I think you have to ask questions like do you
want more reporting or do you want computerized physician order entry and which
do you want hospital field more necessary to do in the short term. I think you
also have to think about how this plays out in the list of things hospitals
should and need to do and are you diverting them from any of that by creating
additional overhead or reporting requirements.

MR. HUNGATE: I agree. Well, my gold standard personally is that it has to
benefit patients, providers, and payers, and it’s harder to do that then most
anything but I think that defines the low hanging fruit. Where you find things
that do provide benefit on those three fronts you’re likely to have made a
better contribution but that’s just my own definition.

MS. COLTIN: I would also suggest that you look at things that serve
multiple goals, so if hospitals have on the list of things we expect that they
move toward electronic medical records, and we’re asking for increased
reporting of clinical data elements, that those relate to one another and that
hospitals that implement electronic medical records will have an easier time
reporting these clinical data elements as part of a claim but likewise putting
them in the claim adds to the incentive for investing in the clinical
information systems. So there’s a synergy between the two goals and the more we
can identify those types of synergies I think the better.

MR. HUNGATE: Okay, well I suspect we’re close to thinking capacity of the
assembled group and that we should probably adjourn until tomorrow morning for
the rest of us, but thank you panel for your very helpful perspective and

[Whereupon at 4:40 p.m. the meeting was adjourned.]