[This Transcript is Unedited]
DEPARTMENT OF HEALTH AND HUMAN SERVICES
NATIONAL COMMITTEE ON VITAL HEALTH AND STATISTICS
March 2, 2012
Doubletree Hilton Hotel
8727 Colesville Road
Silver Spring, MD 20910
CASET Associates, Ltd.
Fairfax, Virginia 22030
TABLE OF CONTENTS
- Call to Order, Review Agenda – Justine Carr
- Standards Letters – Justine Carr
- Subcommittee Report outs, Strategic Plans and Next Steps – Justine Carr
- CMS Line of Service for Information Resources
- Opening remarks by the NCVHS chair – Justine Carr
- How the construct fits within the Department’s data Strategy – big picture – Todd Park
- Opening by HHS Staff, Presentation on the plan – Niall Brennan
- NCVHS presentation – Reactor Panel – Bill Scanlon and Bruce Cohen
- Data Users and Perspectives Panel
- Kerry Hicks, Healthgrades
- Brian Kelly, Aetna
- Harley Geiger, Center for Democracy & Technology
- Bill Davenhall, Esri
- Joshua Rosenthal, RowdMap
- Cont’d Reactor Panel
- Public presentation/testimony and discussion
- Wrap up
- Adourn Meeting – Justine Carr
P R O C E E D I N G S (10:08 a.m.)
DR. CARR: Welcome everyone to the day two of the NCVHS full committee
meeting. Before we go around the room, there’s a couple of housekeeping issues.
Our hearing this afternoon goes until 4:00, and we need to make some plans for
how we get from here to our means of departure. I think what we would like to
do is, if you could, just have everybody write down what time they are leaving
and where they are going. If you write it on your notepad where your departure
point is, what airport and what time or plane or train, whatever your
departure, then we will go around and collect that. And we will try to get
enough cabs, if you want to travel by cab, or you could mention if you want to
travel by metro.
Bruce, did you look up how long it takes by metro from here to the airport?
DR. COHEN: Jack and I are just going to leave at 3:30 to take the metro to
DR. CARR: Okay, just to remind you that it is important that we respect all
the participants who are coming here for the hearing this afternoon that goes
until 4:00. So where possible, I am hoping that we will have a good showing
until the end.
MS. GREENBERG: Those of you who met with us yesterday and were interested in
the new workgroup, we are hoping that Todd Park will get here about 12:30, if
he can. He is coming from Baltimore, and so that would give us an opportunity
to talk with him, so just when you are planning your lunch plans.
DR. CARR: Okay, I will start. I am Justine Carr, of Steward Health Care,
chair of the committee, no conflicts.
DR. FRANCIS: Leslie Francis, University of Utah and Visiting at Oxford, and
DR. GREEN: Larry Green, University of Colorado Denver, in this committee, no
MS. MILAM: Sally Milam, West Virginia Health Care Authority, member of the
committee, no conflicts.
DR. FITZMAURICE: Michael Fitzmaurice, Agency for Healthcare Research and
Quality, liaison to the committee, staff to the standards and quality
subcommittees. If I had conflicts, they would have to put me in jail. I’m a
MR. SOONTHORNSIMA: Ob Soonthornsima of Blue Cross Blue Shield Louisiana, no
DR. COHEN: Bruce Cohen, Massachusetts Department of Public Health, member of
the full committee, no conflicts.
DR. TANG: Paul Tang, Palo Alto Medical Foundation, member of the committee,
MR. BURKE: Jack Burke, Harvard Pilgrim Health Care Boston, member of the
committee, no conflicts.
DR. HORNBROOK: Mark Hoonbrook, Kaiser Permanente, member of the committee,
DR. WALKER: Jim Walker, Geisinger, member of the committee, no conflicts.
DR. SCANLON: Bill Scanlon, National Health Policy Forum, member of the
committee, no conflicts.
DR. MAYS: Vickie Mays, University of California, no conflicts.
DR. WARREN: Judy Warren, University of Kansas School of Nursing, no
DR. SUAREZ: Walter Suarez with Kaiser Permanente, no conflict.
MS. GREENBERG: Marjorie Greenberg, National Center for Health Statistics,
CDC and executive secretary to the committee.
MS. SQUIRES: Marietta Squires, staff to the committee.
DR. CARR: One other housekeeping issue. We have set up SharePoint, and the
goal of SharePoint, going forward, is now when we have documents for the
subcommittee to review, and people want to edit, it will be on SharePoint and
that will assist us in version control. If you have not signed on yet, just
follow the directions in the email, and make sure that you are. And then, as we
have letters or things for us to share through the subcommittees or the full
committee, we will find it there.
Walter, I’ll turn this over to you. Wait, I have one other announcement
actually. We mentioned yesterday that a number of us, six people’s terms are up
in June, and some of us, Judy and myself, have reached the maximum. And because
Judy serves as co-chair of standards, we want to ensure the continuity of the
work. So Judy has graciously agreed to step into the emeritus position, and Ob
has graciously agreed to step into the co-chair position, so thank you, Ob, and
thank you, Judy. Now, Walter?
DR. SUAREZ: Well, what we want to do is cover the three letters that we
discussed yesterday. The good news is basically we don’t have too many edits
and we appreciate the comments from everyone that we received. We are going to
go through each of the three very quickly, and highlight the changes that were
made from the version that was distributed to the full committee.
And then, we want to also spend some time in a fourth letter that was
drafted, that focuses on ICD-10. As we discussed yesterday, it is going to be
very important for the national committee to make a statement about the recent
announcement in the delay in the adoption and implementation of ICD-10. We
wanted to do that and take the opportunity of having the full committee here,
and prepare a very short letter. This will be a first letter really, of what
would become a second, larger, more detailed letter later on, with respect to
ICD-10. But we wanted to sort of make a statement at this point about ICD-10.
The first letter that I think we are going to review is the claim
attachments letter, and I think there are only just a few edits done to that
letter, mostly edits related to consistency in the way that structure of the
sentences were written. There were no substantial changes made, so I don’t know
if you can open that.
Let’s go to the ACA 10109 letter. The Affordable Care Act, section 10109
letter is the one that dealt with these four areas that we cover in the
hearing. And this letter, there were two changes, one on page 3, I believe, so
if you could go down. This is the section 10109.
Let’s go to the first change. There’s some small editorial changes, so this
one is, I think, on page 3. On page 3 of the letter, on the second paragraph
that starts with use of the terms enrollment, at the end of that paragraph,
there is a sentence that references litigation history. And we are changing
that word to malpractice claim history, mostly because not all go into
litigation. So really, we didn’t figure litigation history was proper. We are
changing that to malpractice claim, that is one change.
And then, the second change in that letter is on page 5, under section 4,
consistency in claim coding edits, the bulleted list of issues. The first issue
that does not belong there, it is a copy from actually section 2 before so. We
are deleting that first bullet that talks about reviewing the e-billing
initiative in Texas and California. I believe those were the only two changes
of this letter. Is that correct?
DR. CARR: Do we have a motion?
DR. SUAREZ: Do we have any other changes or was that it?
DR. WARREN: No, that is it.
DR. SUAREZ: That’s it, okay.
DR. WARREN: Just those two.
DR. CARR: Do we have a motion to approve?
DR. SUAREZ: I move approval.
DR. WARREN: Second.
DR. CARR: Any further discussion? All in favor?
(Chorus of ayes)
DR. CARR: Opposed?
DR. CARR: Abstaining?
DR. CARR: Pass.
DR. SUAREZ: Great. The second letter that we will be looking at is the one
that we call the DSMO letter, which is a letter addressing some of the updates
on maintenance process of standards and operating rules. So here, I believe
there were no changes in this letter at all.
DR. WARREN: No. I am just scrolling through to make sure though. Yes, there
was one. It was brought to our attention that we had left out the word
“operating rules” in this particular list. So there are transactions,
operating rules and code sets.
DR. SUAREZ: So this is the only change, and we are passing it around, the
revised letter with the red lines already. That is the only change in this
letter. I guess I will move approval of the letter.
DR. CARR: Do I have a second?
DR. CARR: Any discussion? All in favor.
(Chorus of ayes)
DR. CARR: Any opposed?
DR. CARR: Any abstaining?
DR. CARR: Unanimously passed.
DR. SUAREZ: Thank you. The third letter is the letter on claim attachments.
And again, this one we had a little more changes that we made, all of them
really where primarily editorial and consistency in terms of the phrasing and
the style of the writing. So you will see several, in this particular one that
we are highlighting here in this screen is basically creating a more
sentence-like set of texts in the bullets, rather than leaving the bullets too
loose. We tightened up some of that in this particular paragraph.
I think the next ones are the same thing, primarily correcting some of the
phrasing and the wording and the paragraphs, but no substantive change in terms
of the contents. These are all just kind of editorial changes.
I think beyond that, this one at the bottom of page 3, there was a good
question about when we standard, and then we say request and response. In
reality the claim attachments process, as I think I mentioned yesterday, there
are sort of two ways of thinking about the attachment. One was the unsolicited
way in which the attachment is sent along with the claim.
And then, there is the solicited way, which is a request is submitted by the
payor. That request, there is actually a standard to submit the request. We are
just highlighting here, requesting a claim attachment, and this will be the
standard that would be considered. And then, the second part is the response,
which is a response to a request to submit an attachment, or the submission of
an attachment without a request.
DR. WARREN: So I am wondering if what we need to do with the word standard
is put the standard process is as follows, just to get the same format as the
above? Or if anybody has a different —
DR. SUAREZ: This particular part refers to the actual technical electronic
standard, which is the current electronic standard that is being —
DR. WARREN: I am just trying to get the grammar, because the word standard
still doesn’t tell you what you are —
DR. SUAREZ: Good point. We need some wording in there. We can either say
current standard being considered include –. I would probably say current
standards. So current standard being considered include, so I would take the s
at the end of includes. That is just a clarification change, I think. And I
think that is pretty much, in terms of edits.
DR. FITZMAURICE: Walter, just going back up to where we were before,
something about responding to a requested attachment for submitting in an
unsolicited manner. Is that saying, you asked me for something, but I am going
to pretend that you didn’t, and I am going to send you something anyway?
DR. SUAREZ: No, the two ways the attachments are sent are I send it without
having to be asked to send the attachment, so I send it in an unsolicited
DR. WARREN: Does that help?
DR. FITZMAURICE: Yes, yes, yes.
DR. WARREN: There is one other one on here that I had a question for Sally
on, and that is down here where she recommended we delete the word integrity.
And when Walter and I looked at this, we were thinking of data integrity, which
is as important as the others.
MS. MILAM: My reason for that is integrity is a component of security. I
thought it was a duplication.
DR. WARREN: No, data integrity itself is you have got the right data in the
right format, not necessarily that it is secure.
MS. MILAM: I guess the right data, in terms of integrity, though, is a
domain within the security itself.
DR. WARREN: No, it is a domain within data entry.
MS. MILAM: So it’s not corrupted data, there aren’t missing elements?
DR. WARREN: The numbers aren’t reversed, it’s the correct concept in the
MS. MILAM: I think we are saying the same thing.
DR. WARREN: Except I don’t see it as a security issue. You can have data
with no integrity still be secure.
MS. MILAM: Well, I think security is made up of confidentiality, integrity
and availability domains. But however you choose to leave your language, it is
up to you.
DR. CARR: I think it reflects the different disciplines that people come
from and how they use it. Is there any objection to using data integrity? It’s
a term that we have used in the past. Let’s leave it as data integrity.
DR. SUAREZ: So were we deleting integrity? No, we are not?
DR. CARR: I think what we are hearing is different terminology used in
different disciplines. But I think because we have used the term data integrity
in the context that you are describing in many previous documents, I think it
would be consistent to leave it in, noting that actually it encompasses
security. But I don’t think it changes or undermines the meaning. You are
seeing it as a redundancy, but I think just leave it in.
DR. WARREN: I think that was the last change in the letter.
MS. MILAM: I see that the minimum necessary was added, but it wasn’t
underlined. But I am pleased to see that it was added. It was in your paragraph
under general concerns, third bullet. You need to go up a little bit further. I
don’t have that hard copy in front of me. There, in that third bullet, you
added minimum necessary.
DR. SUAREZ: So that is basically this letter. I guess I will move approval.
DR. CARR: I hear a second?
DR. CARR: Any discussion? All in favor, any objections or abstentions,
DR. SUAREZ: We are going to talk about the ICD-10 letter, so I am going to
DR. CARR: I think we actually had some rich discussion yesterday afternoon,
building on, as Marjorie points out, years and years of discussion. I think
that we are respectful of the Secretary’s decisions and supportive, and I think
we want to reflect that, and offer support or observations that might be
helpful at this juncture.
This is a draft of a minimum necessary letter that documents our awareness,
some thoughts and keeping it at a very minimum necessary. I guess it wasn’t
posted, and again, this is just the work of a small group. So the outcome of
this can be that we accept as written that we revise or that we decide that
it’s not the right time to submit a letter. All of those are on the table, but
this is the straw man to try to capture some of the discussion that went on
yesterday. I think we do have a fair amount of time to be able to deliberate
I am going to read it, so that everyone is aware. Dear Secretary Sebelius,
the National Committee on Vital and Health Statistics is a statutory advisory
committee with responsibility for providing recommendations on health
information policy and standards to the Secretary of the Department of HHS.
Under the Health Insurance Portability Act of 1996, NCVHS needs to advise
the secretary on the adoption of standards and code sets for HIPAA
transactions, including on the transition between ICD-9 and ICD-10.
That is our customary first paragraph, with the exception of the final
phrase, including on the transition between ICD-9 and ICD-10. Yes, Mark?
DR. HORNBROOK: Is it important to put in the CM, because the 9 and 10 are
codes of death?
DR. WARREN: Do we want CM and also PCS?
MS. GREENBERG: Well, I think it is important to put in the CM, and then
after ICD-10, I think the way they have been called collectively is the ICD-10
code sets. I have been particularly asked by World Health Organization to be
careful in this country of not referring to ICD-10-CM or ICD-10-PCS as ICD-10,
because ICD-10 is currently implemented for mortality data, and also in many,
many countries for morbidity data. It causes a lot of confusion. So throughout
the letter, if we can just make that change, I would appreciate it.
DR. WARREN: But I have the wording the way you want it, right?
MS. GREENBERG: Yes, in that sentence, ICD-9 CM.
DR. CARR: And is the term, including on the transition, between, are those
the right words? Including the, take out the on, including the transition from
ICD-9-CM to ICD-10 code sets. Okay, any other comments on paragraph 1?
These are unprecedented times in health care. The pace of change is
extraordinary, and the opportunity to improve the health and health care of our
nation accelerates with each passing month. Ironically, our success in the pace
of advancement has also become our challenge. As clinicians embrace electronic
health records, they are learning not only how to document care delivery, but
also how to use population data and how to redesign care to meet the needs of
With meaningful use of electronic health records comes a further requirement
to express clinical concepts in SNOMED CT, a structured terminology that is
also new to clinicians. Information technology infrastructure demands have also
accelerated with this rapid pace of change, including the recent
implementations of ASC X12 5010, NCPDP D.0 and 3.0. These transactions
standards will bring us closer to the goal of administrative simplification.
ICD-10 also supports administrative simplification as a granularity of the
structure of Ford’s Electronic Communication of Clinical Information, obviating
the needs in many cases, for additional inquiry and or claims attachments. Stop
MS. GREENBERG: So the ICD-10 code sets.
DR. SUAREZ: A couple of quick comments. Where it says, they are learning not
only how to document care delivery, I would just say they are learning not only
new ways on how to document care delivery, because it would be too strong to
say they are learning new ways – they are learning how to document. But
new ways on how to document care delivery consistently, but also how to use
population data, so that is one change.
There is a missing of before the ASC X12, implementation of ASC X12 version.
DR. WARREN: Do I have this sentence right, the way you wanted it, Walter?
DR. SUAREZ: I am sorry, they are learning not only new ways to document care
delivery consistently, but also exactly. Yes, I think that.
DR. W. SCANLON: I think that moderates it some, but I am taking a physician
perspective here, and I want to take an umbrage at the original sentence, on
learning how to document care. And my solution was to insert the word
electronically, how to document electronically care delivery.
And I actually question the other, the second two parts, whether this is
really widespread among the physician community, that they are also how to use
population data, and to meet the needs, and redesign.
DR. CARR: So let me give an example. If you are to get your HEDIS scores
perfect, you need to know who in your panel has diabetes, hypertension, heart
disease. If you have a paper system, it is not easy. If you have an electronic
health record, you can pull a report of show me everyone who has diabetes, so
that is that population. Population can be at the provider level, as well as at
the national level.
DR. W. SCANLON: I think of those things as aspirational goals, if you
started off, if you want to get your HEDIS scores perfect. I can’t believe that
we are yet, as we are talking about the adoption of electronic health records,
they are not being exploited for their full potential on a widespread basis.
This kind of goes further than I think we have evidence for.
DR. CARR: Let’s hear from the docs, and can we get Raj on the phone?
DR. SUAREZ: I think one way to change a little bit what you are pointing to,
is that part that says but also how to use population data. In my mind, if
really also how to use more granular data for population health management, and
to redesign care, because it is ultimately really we’re learning how to use
more granular data to do population health management.
DR. CARR: But let me take it back a level to Bill’s question, because I
think we need to break this down one step at a time. What Bill said is that he
is looking for the level of certainty that HEDIS measures are a significant
driver of physicians, to begin to look at population health.
DR. WALKER: I would say that Bill is right at one level. I think one thing
is the verb, learning, is the right verb. That does not say we are doing it, we
are learning it. I would also say that there are tens of millions of Americans,
maybe 40 or 50 million, who are being managed at a population level. We can
tell you, and Kaiser can tell you, when each of 22,000 diabetics had their last
hemoglobin A1c. And we manage that sucker, so it is happening. I think learning
maybe expresses that we are still in the learning innings.
DR. CARR: I think you are right. I think you are the 99th
DR. WALKER: But you have to remember, there are probably 40 or 50 million
Americans that are at the 16th percentile.
DR. CARR: I think there are mature systems that not only have electronic
health records, they have connectivity, they have informatics expertise to know
how to push data at a time. I think that is the aspiration, to be sure. I think
there are physicians who are just now, as we heard from ONC, getting their EHR,
and they will, in time then, be learning how to aggregate for a population, in
order to manage.
So I heard Bill also questioning whether we can accept there is a mandate
that practitioners see the importance of managing populations. I use the term,
the HEDIS measures, so I think you were questioning that.
MR. SOONTHORNSIMA: Maybe Bill’s point, it would separate a panel of patients
from population, so that we don’t confuse the two, if that makes sense. We are
talking with individual physicians here, and they are going to be much more
concerned at the practice level, the panel of patients, versus the whole
DR. CARR: Let’s think of that, and let’s hear Jim.
DR. WALKER: I think we would be better off to define population. When we use
that word in all of these discussions, we are usually completely unclear. Are
we talking the country, are we talking a state, are we talking 21 million
veterans, are we talking 2000 patients per doctor? Those are all populations.
And the structure, the intellectual task of managing those populations is all
similar. Obviously, there are different demands. So I think we would be better
off to say a population can be the panel the physician stayed at IDN.
DR. CARR: So you are saying add a sentence that says a population can be any
of the following, can be at the local, at the office, at the state, at the
DR. WALKER: And this will come up again and then we have got a standard way
of talking about this that makes sense.
MR. SOONTHORNSIMA: Is it just patient populations?
DR. WALKER: I would use an asterisk, I would just say population can be a
DR. CARR: Yes, so just a sentence. I want to hear from Paul.
DR. TANG: Is there a reason we are limiting it only to population
management, versus the management of individuals? I mean, ICD-10, so I’m just
asking a question, why?
DR. CARR: Yes, individual and population management, I think that is right.
DR. SUAREZ: Part of what the next statement, after population management, is
to try to go to that, is really signed care to meet needs of the patients. So
really, sort of managing population, and the other part was trying to get to a
managerial and individual care.
DR. CARR: They are learning not only ways to document care delivery
consistently, but also how to use more granular data to manage individuals and
populations. It should be manage the health of individuals and populations, and
then parenthesis, populations equals this. Larry?
DR. GREEN: I don’t understand why this matters to this letter. It seems to
me that the letter wants to get to the point that we want the delay to be
brief. And it’s the recent action that provokes the letter.
DR. CARR: We are going to get to that conversation. But I was looking for
the physicians in the room to respond to Bill’s concern that it is
aspirational, that the management of a population is on the horizon or in the
purview of physician practice today.
DR. GREEN: I tend to agree with Bill, that it overstates just a little bit
where we are. I am more inclined, in a background paragraph, instead of arguing
about where we are with adoption and stuff, what I believe is the key issue is
the United States is using a classification system that is obsolete.
It is lacking, so we lack the ability to express important clinical concepts
that matter to people, and that we are continuing to deprive ourselves of the
opportunities to use modern medical knowledge with a classification system that
very succinctly and briefly, and quite simply on a claims form, could be being
used, but it’s not being used. And from a clinical point of view, it’s the
absence of important concepts in what we are required to report that is really
what justifies getting.
DR. CARR: It sounds like we need more work on this to prioritize what we
say, and to say it in a succinct way. Maybe we need to talk about the finish
line of where this letter is going, and then work back from there. Is that
good? So knowing that we have to do more work on paragraph 2, I’m going to —
DR. WALKER: Justine, can I just raise one issue? It’s it two and three both.
But I think the letter begs the question. We are saying for administrative
simplification, second paragraph, third paragraph, what AMA and John Halamka
and lots of other people are saying is, one person’s administrative
simplification is another person’s administrative nightmare. And so, the
question is whose administration, and that is the whole argument here, who
bears the administrative burden, and who reaps the administrative benefits. And
so, I think at some point, we need to address that and say on some basis, we
believe that this burden is justified by this simplification.
DR. WARREN: Jim just caused me to think in a little bit different way with
his comment. The administrative simplification is for the patient, because what
we are trying to do is get the information that we need in order to improve the
health of the people of this country. Not to improve government functioning,
not to create a burden for the physician, but to improve our health, our
individual health. And maybe that is what we need to make clear in the letter.
This is all about patients, it is not about regulation or anything else.
DR. CARR: So let’s hold those thoughts, except for Marjorie, because I want
to make sure we are headed in the right direction. But if you want to make a
comment now —
MS. GREENBERG: One comment on the population health, and this was just a
clarification since I wasn’t at the meeting when this was drafted. I thought
that in mentioning populations in the context of meaningful use, you were
referring to not only obviously the more population approach, and I do think
population has more behind it than panels. I certainly understand what Ob is
But also the fact that meaningful use does require some structured
communication with population health data systems, like immunization, et
cetera. Right now, this is just dependent on if a physician sends a postcard or
something. So there is that engagement with the population health agenda, as
DR. CARR: So I think that the clarifying sentence that we received, I think
from Jim, to say population health may represent a panel, a state, a country,
all of the above. I think that is helpful. I want to read the next.
MS. GREENBERG: But I did want to say one thing about the other thing. I
understand what Jim is saying, but this letter specifically ties administrative
simplification to obviating the need in many cases for additional information,
claims attachments, et cetera, and that should be for the individual physician.
DR. CARR: Okay, let’s hold the comments now, just read the next paragraph to
see if the finish line is where we think we need to be. So the NCVHS has a long
history of support for transition to ICD-10 code sets —
MS. GREENBERG: That’s okay the way it is.
DR. CARR: — because of the potential to efficiently summarize the critical
clinical information, thereby facilitating administrative functions, as well as
understanding of population health. We recognize the challenge before us. We
urge you to limit the delay no more than one year. We also urge you to leverage
emerging technology, such as Mlm’s just released iMagic program, which allows
clinicians to easily translate clinical concepts into SNOMED and ICD-10 code
MS. GREENBERGL: Well, that is actually ICD-10 CM. I don’t think the Mlm
product deals with 10-PSC.
DR. CARR: And I will just read the last two lines, so we can then circle
back. As always, NCVHS stands ready to assist in any way we can. We are already
scheduled to hold a hearing on ICD-10 code sets this spring or early summer. We
would be happy to address specific issues related to facilitation of the ICD-10
So again, let’s go at a very high level. We’ve done a couple of things,
number one, identified that we feel that, given our long history, it is
important that this meeting to say something about ICD-10. And secondly, we
have tried to articulate this, acknowledging the issues that have led to this
delay. And thirdly, offered up our assistance in helping address those issues.
So let’s just go around the room, I think, one by one. Judy, did you want to
say anything? Paul?
DR. TANG: I don’t think, and it’s probably deliberate and I just want to
make sure that is true, it doesn’t address the sort of incorporation of SNOMED
into the discussion and being a vehicle for conversion.
DR. CARR: Do you have a friendly amendment, as Lorraine would say?
DR. TANG: I can read the thoughts. So we agree with the importance of
converting to ICD-10-CM, so that we sort of put that on the table, that it is
important to billing in epidemiology. And here’s the part where we sort of
create the opportunity. So at a time when the US is converting to the EHR at an
accelerated pace, it’s important to establish a standardized clinical
terminology. And we acknowledge that the Department’s issuance of the — for
meaningful use stage 2 recommends SNOMED as that clinical terminology.
So what the recommendation might be is that, at the time the country is
converted from ICD-9-CM to ICD-10-CM, that it leverage the work of the NLM in
mapping SNOMED ICD-10-CM. And recommend a strategy, a voluntary one, but it
puts on the radar where it is not at all today, that the users of EHRs use the
recommended clinical terminology to code diagnoses, and use the NLM-supplied
mapping to translate that or convert that into ICD-10-CM.
DR. HORNBROOK: Which is royalty-free, publicly supported.
DR. CARR: If you want to give Judy a copy of what you have, we can have that
for consideration. Let’s just go around the room, because I want to get the
feeling of the entire committee on is it right to have a letter, is this the
right direction, and any other comments?
DR. COHEN: I think it is right to have a letter, and as strong a letter as
possible. Now, the question, going back to the issue that Judy and Jim raised,
is the goal of ICD-10-CM to improve individual health? And if that is the case,
we should say the committee is supporting a small as delay as possible, because
we feel not only the focus should be to improve patient health, as well as
facilitating administrative functions. So if we truly believe that that’s the
goal of rapid implementation of ICD-10-CM, the focus should be on the
individual health and the population health.
DR. CARR: Jim, I’m going to just go around the room, because I think we have
to have all voices heard, and out of that will emerge the kind of collective
MR. SOONTHORNSIMA: I believe there are two running themes. One is a sentence
that tries to describe SNOMED, but I believe that sentence really, what we are
trying to do is help harmonize the clinical context systems, whatever it is, in
this case SNOMED, with the classification system, ICD-10. That is one goal.
The second goal is really to facilitate and expedite the implementation of
ICD-10 with limited delay. And I think that is what we are trying to say in
this letter. What we need to clarify is we also urge you to leverage the
emerging technologies. I think what we need to say, what is it that we want
that system to do. Isn’t it to harmonize? That is the point I was trying to
make. Is it to harmonize the classification system with the clinical context,
and limit the delay. The only comment I have, which may be open for question
later on, is why a year and how did we pick that one.
DR. FITZMAURICE: A couple of things, one that I might consider even
eliminating mention of SNOMED. Our focus is on the decision in front of the
secretary right now. And then, secondly, the sense that we urge you to limit
the delay. I would say, if you choose to delay, which means we have some
questions about whether it should be delayed or not, but if you choose to
delay, no more than a year.
But also allow ICD-10 to go into effect, because people can use ICD-10, but
delay the mandatory use of it for a year, if that’s their choice. That would
recognize the tremendous investment that has already been made in ICD-10, and
that people could use it and get something back for that investment.
MS. MILAM: I agree that the letter is important. I think it could benefit
from some additional framing, in terms of impact and risk benefit analysis.
DR. GREEN: I would like to see us have a letter. I agree with particularly
what Mike just said about how to make the point. I like that if condition, or
if you decide to delay. I really don’t think we need most of what’s in
paragraph 3. I think it should be greatly simplified.
And one other thing is, I think the letter just needs a sentence or two that
takes care of June Walker was saying. We need to indicate an awareness and
appreciation and sensitivity to the fact that this solves problems for some
people, and makes problems for others. But the position that I would like to
see the committee take is that aware of that, do it.
DR. FRANCIS: I am not sure that it fully captures the sense that what is
behind the delay is that people feel like they are having to meet lots of
demands at once. And on the positive side, I would suggest anything we can
figure out concretely, and maybe the iMagic is and we can frame it that way, to
suggest ways of helping coordinate all the things that people are having to do.
So if you are going to delay, don’t just delay. Figure out how to make use of
the delay in a constructive way that actually helps people put it together.
MS. GREENBERG: I think all the comments have been good. I actually, although
it may overstate, it may be more aspirational than actual, I thought that some
of paragraph 2 actually was rather eloquent actually almost. That’d be too
strong a word, but in capturing the very issue that Leslie just mentioned,
about all the different demands that are on people.
I think my bottom line is that, and this is your decision obviously, it does
seem appropriate given your responsibilities at this point to make some kind of
a statement, but I really do think that, if you want to do that, you need to do
it now. And it could take a few months, writing the perfect letter, but I would
say that it probably have no value.
DR. SUAREZ: I agree we need a letter. I think at this point, the main thing
the letter can address is the time, how long this will be. I agree that framing
it as the delay should not be more than a year would be important. But I think
there are a couple of other things that are important to mention, that haven’t
been clearly stated perhaps, or at least I haven’t heard it.
Number one, I think that is maybe what the second paragraph is saying, is we
continue to believe that we must, as a country, convert to ICD-10. That is a
statement that we have present, just to ensure that the idea that maybe we
shouldn’t even go to ICD-10, so number one.
Number two, in one sentence or 15 words, we say we urge you to limit the
delay to no more than a year. We don’t give reasons for that. We want to say a
couple of short sentences about because, and the big because is any delay will
delay the ability to benefit from the transition, and will increase the cost of
doing health care business in this country. And we can elaborate more on that,
but I think we need a statement about the reason why the delay should not be
more than a year.
The second thing is, Mike, your point, at some point there was some
discussion about well, maybe there should be a way to allow people to ICD-10
and also ICD-9. The worst thing that can happen in this country is if we have a
prolonged dual system. That would be more expensive than anything we might have
ever thought. Normally, with other standards like 5010, yes, we could handle
4010 and 5010. In fact, that was the idea. We had a year to allow to do that.
But with ICD-10, my suggestion is that we need to emphasize that the delay
should be no more than a year, and that there should be a hard conversation. A
hard conversion meaning minimizing, and again, given the reason to limit the
need to maintain dual systems that support all a new code set.
DR. CARR: I am just doing a time check, because we are technically supposed
to begin a talk on the subcommittee report. We will go a few minutes over, but
if you have a bullet.
DR. SUAREZ: One last bullet is, in addition to we support the delay for no
more than one year, that there should be a deliberate mechanism that ensures
that, during the remaining time towards the transition, there is concrete steps
taken by people to appropriately and efficiently and effectively achieve the
transition. So there is not just we delay, it’s we delay and people have to do
DR. MAYS: I definitely think we should have a letter. I would like to see us
try and keep paragraph 2 in. I think people have said kind of what the problems
are. The other fix, because I didn’t understand the more than one year. But I
also think we recognize the challenge before us, that if we could kind of say a
few more words there. I think that paragraph just needs a little more
DR. W. SCANLON: I have actually, yesterday and today, been more in a
learning mode than anything else, listening to sort of everything that has been
said, and trying to decide sort of which of the statements are hypotheses and
which of them are facts. And I very much like Jim’s framework, which is the
issue of sort of the cost and the benefits, and recognizing that there is a
distribution that is across different types of people.
I find the administration simplification argument more intuitively plausible
than some of the others, that I find a stronger rationale for moving forward.
Having said that, I have no sense of what the real costs of delay are, and I
have no sense of sort of why a year is the right sort of timeframe.
DR. WALKER: Sir William Osler was teaching a bright young trainee, who was
faced with a big patient problem, and started to rush into action. And Osler
said famously what we try to teach all trainees. Don’t just do something, stand
there. We don’t know the benefit and the cost, but what we do know is that the
cost has been exaggerated by a factor of at least two orders of magnitude. The
costs of making the conversion are at least 100 times greater, and this is for
a very capable organization. This is John Halamka. For very capable
organizations, it’s at least 100 times the published estimate from CMS.
I think the first thing we ought to do is call on the director of CMS to
publish a cost-benefit analysis within six months. And it would obviously be
only as good as the evidence is, and the evidence is undoubtedly not very good.
But I think without that, it’s like Bill said, we don’t know what anybody’s
cost is. And the benefits, we haven’t thought carefully about it.
Number two, clinical communication, the standards conceptual just
transmitted to ONC, which largely accepted it. A whole set of standards for
communication, immunizations would be CVX, and we actually have standard
vocabulary for clinical communication, none of which depend on ICD-10, and for
which ICD-10 would be inadequate.
We have got to acknowledge administrative simplification and burden, and
that is one reason the secretary needs to have the director of CMS make some
estimate of what those are. I am not aware of any evidence, or very much
feeling, that ICD-10 will improve patient care. I think there is wide belief
and some reason to believe that the use of SNOMED will, but that is a stretch.
But we aren’t saying it in the letter, so it probably doesn’t matter.
I agree with not mentioning SNOMED. For one thing, we got it wrong. The
image is about translating from SNOMED to ICD-10. There is an NLM map from
ICD-9 to ICD-10, which would be relevant to sort of call out as one of the
things that could be more heavily publicized, so that people could make the
transition easier. As far as I can tell, and I think in NLM’s sense, not very
many people know about that map.
MS. GREENBERG: Is that the different than the general equivalence maps? That
is what the official map is.
DR. WALKER: It is by image.
DR. HORNBROOK: To be a little contrary, I think we are missing the boat
here. The fact that both I9 and I10 are kind of artifactual classification
systems, I10 is a lot more rubrics in it, so it gets much more detailed for
clinical realism. I9, if you get and start drilling down in it, you realize
that there are a lot of categories that have almost no clinical meaning, other
than being different than some other category, which also has very little
So the iMagic transition means that if you are inside an EMR, you’re already
working in SNOMED. Physicians should be working in SNOMED to describe their
clinical, and we should be using that data, rather than I10. So right now, you
have the ability to move physicians into SNOMED very quickly, and then the
administrative business of putting SNOMED into I10 is handled electronically.
Nobody has to learn I10 who is a practicing physician.
And of course, physicians right now have quote learned I9, have their
favorite code buckets inside their specialty area, and that is the mental set
that they are resisting change, I think. I think we have an opportunity here to
make this letter much more pointed and important.
DR. GREEN: The letter should go quickly, obviously. But for me, the reason
is to avoid the disruption of heavy resources that are currently in place in
the industry, not just on the provider side, but on the payor side, that are
vested in achieving whatever the final outcome is, whether it is a SNOMED
intermediary or not. If we are getting to ICD-10 eventually, these resources
have a way of being sapped away to other uses. There is a sum cost right now of
people not working on this as diligently as they should. So that is one of the
becauses for Walter.
DR. WARREN: I would just like to say, I had a chat with Lorraine Doo
yesterday. And just as a caution for us on our speed, CMS is already working on
the NPRM. And if we don’t have our input in there like within the next day or
two, if our comments are not considered, then these options may not be of the
NPRM, and they will be off the table. So we do have a timeframe here that we
need to expedite this pretty rapidly.
DR. CARR: So I think one thing that I heard that I believe that we have
consensus on is that if there is a delay, the time should be used
constructively to address the issues. Do we agree on that? That is one.
Second, do we agree that we want to mention, as was stated in the beginning
of that paragraph, the recognition of the consequences of our own success, in
many ways, our capacity to accelerate change is now the challenge that we are
dealing with, because so many things are converging at the same time.
DR. WALKER: Someone else sort of said it. I think we would be better to say
there is substantial cost being incurred by many entities, all of which have
significant other demands on them. I think that is enough to say.
DR. CARR: Well, the only thing I would say is that, there is a lot of
passion and emotion here. I think it is important to say that because we have
been able to be nimble and make changes and move quickly with high-tech ACA, I
want to put that forward, that that is the good thing. But the unintended
consequence is that everything has come to at the same time.
Then, I think hearing different points of view is the what gets better with
ICD-10 and for whom.
DR. SUAREZ: Did we agree on the one year delay?
DR. CARR: No, we have said two things that we agree on. And I think we agree
that in this room, there are varying perspectives on what gets better and for
whom. And perhaps what is true is that the what gets better, do we have
consensus? If we say what gets better, with this classification, we would have
more granular description, right?
DR. WALKER: Chris Chute has done a study that demonstrates that that is way
overstated. There is some marginal benefit, in terms of clinical resolution,
but it is far less than people believe. And they reproduced a study that was
done three or four or five years ago, and came to the same conclusion.
DR. CARR: So we can say it is more granular, but how much, we don’t want to
put a modifier on that. Obviously, if we have it, it aligns with what the rest
of the world is doing. And is it true that, if we have this granularity, the
need for claims attachments could go away. Walter, you reflected that if we
DR. SUAREZ: It won’t go away, but it will be reduced. I expect that will be
hopefully the outcome, because I mean we are going to have more granular data.
And going back to what Judy said, which is we need to do this in the next day,
and what Larry said, which is what is relevant in this letter, what is the
important relevant point? And the important relevant point, we can argue and we
can bring back.
DR. CARR: What I am trying to do is a strategy, let’s just say what we agree
DR. WALKER: But I want to address the claims attachment. Saying the
procedure that you did or the thing that you did more precisely does not mean
that you have answered the question, what was the justification for doing it. I
think the idea of the claims attachments will decrease needs to be seen.
DR. CARR: It’s not substantiated. So let’s step back from that. Is there any
other statement that we heard that every person on the committee agrees with?
DR. FRANCIS: I wonder whether people agreed with the comment about trying to
use any delay constructively.
DR. CARR: We said that, that was our first point.
DR. FRANCIS: Well, I wasn’t sure that you had said that.
DR. CARR: That was the first that we all agree on. If there is a delay, use
the time constructively. What was the second point?
DR. WALKER: I think we might be able to agree on the dimensions of the cost
benefit analysis. There are some costs that have been mentioned, there would be
administrative costs for provider organizations, different costs for different
sized organizations. There is the payors who will have costs.
My guess is that CMS’s costs might be greater than private insurers, but I
don’t know that. But I am guessing that their information ecology is enough
more complex, that that might be the case. Anyway, we might at least identify
the elements of costs and benefit that we believe are relevant to consider.
DR. CARR: I think in response to that, what we have heard is that there are
costs for those who have not started, there are costs to do it. For those who
have got it done, there are costs to maintain. So there is cost going to be
incurred, regardless of what the next step is. And the question is, by whom?
DR. WALKER: How great? Maybe they are orders of magnitude different, maybe
they aren’t. I don’t think we know.
DR. SUAREZ: The question I still have is, what is the relevance of that to
the fact that the delay would happen? Are you making that argument because we
think that the delay should be longer, or are we saying that we shouldn’t even
go there because —
DR. WALKER: It is specious to recommend a length of delay if you have no
agreement at all on what the costs and benefits are. That is all.
MS. GREENBERG: Well, first of all, the department’s press release stated
that ICD-10 codes are important to many positive improvements in our health
care system, et cetera, et cetera. Now, I think the committee has to think
whether you just disagree with that. Certainly, there are members of the
committee who clearly do disagree with that. I think you are reopening this
fundamental question, which even the press release seems to have settled from
the compartment’s point of view and the rulemaking. I think that is something
that you need to think about.
Also, I believe that whatever you think about ICD-10-CM, and you can find
studies on all sides of this, there does seem to be a recognition in this
committee that we need to move to more modern code sets and terminologies, a
greater linkage between clinical concepts, SNOMED CT and the ICD. The ICD is
not going to go away, unless the US completely turns its back on the rest of
the world. The ICD, in different versions, is being used effectively around the
world. And those countries are getting better health care outcomes than the US
in many cases. I think we have to pause, if we don’t think about.
And ICD-10-CM is definitely a pathway to greater convergence between robust
clinical terminologies and modern classifications. ICD-9-CM is not. I think
before you trash the entire ICD system, I think you need to put this in the
context of what you are saying.
DR. CARR: I think we are converging on things we can agree on. Paul and then
DR. TANG: The time emergency, I think, is moot for us in the sense of the
NPRM is going to ask for comments on the delay. We will have the time to give a
more measured response to that. So we don’t have to hurry up and say hey, don’t
forget to ask about the delay. That is the whole purpose of the NPRM, so I
think that is actually moot.
We either say something about SNOMED, or we even say not to forget about
ICD-10, but forget about the issue in terms of converting from everybody’s use
of ICD-9 problem list(?)to everybody’s use of ICD-10 and problem list, and just
forget that stuff.
Departments are already on record for saying let’s go to ICD-9-CM to SNOMED
for diagnoses. Let’s just accept that. They already are working diligently on
doing the mapping. In a sense, it becomes almost the proposition is we do
better, it’s almost a non-issue for the frontline clinicians if we do it this
way. So we either make that bold recommendation so that you can include it in
the NPRM, or we don’t. We don’t have to worry about saying don’t forget to ask
about why the delay.
DR. CARR: So your point is the timeliness and it impacts on what goes into
the NPRM, and clearly the delay issue will be in there already. Bruce?
DR. COHEN: I agree with Marjorie. I don’t think we have to address the
importance of converting to ICD-CM. I can’t speak for the clinicians, but from
a population health perspective, it is being used to measure more free-standing
morbidity all over the world right now, and it will continue to be. I think
that is not an issue. We just accept that going to ICD-CM is important.
Also, I think talking about SNOMED is a real distraction for this letter.
That is another issue, again we are talking about the clinical aspects of it.
And the bottom line is, do we want this delay to be as short as possible? I’m
hearing actually a couple of different things about that now. One, that some of
us fundamentally just accept that it should be as short as possible, because
the longer we delay the greater the cost will be for implementation. And others
feel that perhaps that is not the case. I would like to get a consensus from
DR. CARR: I think perhaps the length of the delay is less important than the
things that need to be addressed within that delay. We want an efficient,
timely adjudication of the financial burden that is being held by two groups.
That needs to be understood and managed. I do think I agree with Paul that
SNOMED, these data come from clinicians. Clinicians are the ones that are being
asked to retool how they document. And the burden on the clinicians to learn
this in the community practices is in play here.
I think to Paul’s point, pointing out that we could take that off the table
if we could make tools of presentation layer, as Jim described, that is a piece
of it. Paul and then Walter.
DR. TANG: So just to restate something in a different way, which may
actually provide the way to say this, which is, if we instead focus on SNOMED
instead of a conversion of diagnoses from 9 to 10, we will be aligning and
harmonizing with other federal rules, which is meaningful use. That would be a
way of stating this.
DR. SUAREZ: So just in the interest of time here, too, I would suggest the
following. Ultimately, what we are trying to do is recommend to the secretary
certain things to consider, to include in NPRM. At the end, that is what we are
saying. Anything else, we can talk about all sorts of other things, and go back
to cost benefit analysis and documentation, that was actually included in the
2005 publication of the regulations that called for ICD-10. Actually, it was
2007, I believe it was.
There are all sorts of background that I think could create more noise. What
I think we can recommend is the secretary is to consider including the
following five things in the NPRM. Number one, the delay, and that it should
not be more than a year, or the shortest as possible, preferably no more than a
year. Number two, that the NPRM should emphasize the importance of the role of
the SNOMED in the EHR and the translation of the SNOMED according to the ICD-10
for billing purposes and administrative purposes, and the need to establish
requirements on the EHR to be capable to do that. So include something like
that in the NPRM itself.
Number three, I would argue that the NPRM should include specific steps to
achieve the meeting of the new deadline, and that should be established in the
NPRM. Number four, I think that NPRM should also state basically how the
federal government itself will intend to meet the deadline and expect that, for
example, Medicare would not accept transactions that come in the standard that
are using the old code set. And that gives a sense to the industry that this is
going to be a deliberate way of having everybody move to the new change.
And the last thing I think you need to include is that the transition should
be a hard conversation transition, that there should not be a dual system. We
all know that we are going to have to go to a dual system, but that the length
of the dual system will be directly related to the cost that this transition
I think what we can suggest is recommend the type of things that should be
included in the NRPM, and then just move along. We are going to have to have
the chance to comment on the NPRM, we are going to have hearings that will
bring up new issues about this whole thing. But I think we should concentrate
on the four or five things that we should have.
DR. CARR: I think you are right, I agree. I think we should have a letter. I
think if we articulate these are the concerns that need to be addressed during
that delay as quickly as possible, we don’t even need to express consensus on
which is more burdensome, the expense to the industry or the expense to the
doctors. We simply say the expense to industry and clinicians needs to be well
understood. And we need to create a roadmap that will achieve success. We need
to identify the roadmap that will achieve success. We need to harmonize around
the clinical terminology direction of SNOMED with a seamless backend to take
the work out, but to use that mapping. Do I hear any objections to putting
those three things? Walter had some other things, but those things that I just
DR. WALKER: So we are recommending a cost benefit analysis, based on the new
DR. CARR: I think we are not going to go as far as that.
DR. WALKER: Then I want to point out to us that the whole reason we are
talking about this is because AMA deposits a different cost benefit analysis
than others do.
DR. CARR: Understood.
DR. WALKER: So this is all about estimated cost and benefit, and for us not
to mention it is silly.
DR. CARR: I don’t think we want to be as granular as do this, do that, do
the other thing. But address, adjudicate, understand the concerns on multiple
fronts of the costs, the cost of continuing in two systems, the cost of
converting to the new system. I don’t think we need to get any more specific
than that, but just to say that.
DR. TANG: On your last one, about terminology, was there an ask in the way
you phrased it?
DR. CARR: What would you suggest?
DR. TANG: If the theme is let’s take advantage of this, if there is a delay,
let’s take advantage of it to do something constructive. I would actually put
in language that says here is our recommendation for your consideration of what
would be an alternative, but we view a constructive approach.
DR. CARR: Again, I am trying to keep it as objective as possible. But SNOMED
is the way meaningful use is going, we’re asking that we harmonize with what we
have already said.
DR. TANG: That’s the ask then.
DR. CARR: So are we in agreement with that?
DR. TANG: Yes.
DR. CARR: That we have already committed to SNOMED, we harmonize with that
and facilitate with whatever back end mapping.
DR. SUAREZ: This is a fundamental component of this, yes.
DR. CARR: Walter had two other things, we are going to talk about that, and
then we are going to stop and go onto the committee reports, and come back.
Over lunch, we will try to put this together. But you also were specific in
asking that it be addressed to meaningful use. Again, I think to do that, I
thought you were saying that —
DR. SUAREZ: I was referring to what Paul had —
DR. CARR: You also said that we avoid having dual systems.
DR. SUAREZ: That there be a hard conversion, and we need to use those words.
DR. WALKER: That is back at cost benefit analysis. You have got to
understand that for the provider, you could easily have a system, where the
provider is allowed to provide either nine or ten, and it is the insurance
company’s problem. So there again, it’s whose cost, how much, and it is being
DR. CARR: I think we agreed on three important things, and if we say those
three important things in a timely fashion, with a letter going out today or
tomorrow, we will have added value, brought something new to the table,
informed the NPRM, knowing that when the NPRM comes out, there will be ample
opportunity for us to have a more in-depth discussion. Any objections to that
DR. SUAREZ: So just to clarify, the next step we are going to take this
letter, draft it and have executive committee –
DR. CARR: No, I am going to see if we can do it over lunch. I would like to
have it looked at before we get out of here today, a draft. So we are now going
to move into the report outs of each of the committees, and we are going to
have to be doing some parallel processing here. So Judy, if you can kind of
call out some of those things. And I am going to ask that we keep the kind of
opening sentences in section 2, paragraph 2, about this is an unprecedented
Paul, would you like to report out on the Quality Subcommittee?
DR. TANG: I could go to conclusion recommendations or I could stay at some
of the observations. It’s up to you.
DR. CARR: Why don’t you take eight minutes?
DR. TANG: Christine Bechtel from National Partnership, started out with the
story of her trying to find her own doc, which was very illustrative. I will
summarize her journey in this way. There is a lot of information out there,
varying values. But even the most motivated, knowledgeable, persistent consumer
can’t find information useful for choosing physicians in hospitals. That is
sort of the bottom line of her experience.
The cause of that, complexity of information makes it impossible to
understand health-related metrics, whether it is to choose a plan or to choose
a doc. And the way it is told, the legalese, the academic way it’s told makes
it, once you even find it, impenetrable.
One of the enlightening things that came from consumer union was in their
surveys, basically coverage concerns and non-profit costs quality. The
implication is important to us because they are not even going to get to the
measures that we are talking about. So that, we have to consider, that an
important ask of panelists were that measures needed to be designed with the
patients, for the patients and tested by the patients, just a design principle.
Measures that were important to consumers are things that not a general like
how are they rating, but things that they could apply to them individually,
their specific conditions. These all place requirements essentially on the
measures that matter to consumers. A big impediment to these measures that
consumers could use if they were meaningful to them, are the lack of standards
and the cost of prosperity tools. We had someone from the Physical Therapy
Association, the vast, all of them, except one, had a license cost to it. So
the topic of, well, gosh, if this was paid for by public funds, why all of a
sudden are license fees getting in the way of their appropriate use.
The movement in population health status needs to be dealt with at a
community level, strong voice from a group from Rochester, very strong
community. We heard it from CVHI, as well. And some notion that the experience
of care is better than a patient set rating. We have to, as health care
professionals, do a better job at learning how to engage, activate patients,
and it may not be with simple scores. And of course, no one pays for outcomes
or improvement, so therefore nobody wants to measure it.
So some conclusions in the lumping style, is that consumers and patients
don’t have access to the relevant information they need to help them select
their health care team or insurer, or to make informed choices about their
care. So what is out there is siloed, complex and hard to understand. That is
To support health reform, both patients and payors need standardized,
understandable, useful measures that matter to me, and usable things that I
could actually understand and interpret. The lack of standards in propriety
interest in measures that do exist inhibit or prohibit the development of
useful comparative information at any economy of scale. Somebody might decide
to spend money on this measure, so they can’t combine them.
Recommendations, first is to take advantage, we do have ACA. There are going
to be insurance exchanges, many or most insurance exchanges are going to be
federally supported. So therefore, not only does the government have a stake
in, it has a say in the health insurance exchange rules. Our proposal then is
that the insurance exchanges have standardized comparison benefit tiers, with
clear, easy to understand comparison charts. These tiers have a specific
standard benefit description, that doesn’t mean any plan can’t add up to any of
these tiers, but at least gives a consumer patient the ability to compare plans
at some kind of standard level.
At a minimum, there would be a nutrition summary like, that appears on the
can, a way of saying what would be the reimbursed, give some transparency into
what the insurance company pays the provider, and the out of pocket costs for
someone on this plan for these common transactions. That could be very
illustrative, but that is saying it more specifically than, let’s say, let’s be
transparent with cost. Again, that goes back to understanding by the consumer.
A second recommendation is that again, for all of these federally-supported
health insurance exchanges, that they also publish, we will just call it next
generation measures that matter to consumers, and will describe how those are
created, in a standard form, it’s easy to read and understand comparison charts
that would include things like demographics and location we know are important
to them, quality measures of the kind we are going to describe, experience
measures, and that these publically reported measures be standardized, free of
licensed cost, and include condition-specific measures.
Recommendation 3 is that we take advantage of what already exists that, as I
said, were siloed. So on the same webpage for these insurance exchanges, that
there be consolidated federal and federally-mandated reports, such as whether
they are already meaningful use qualified, the PQRS scores, HCAP, CGCAP. In
other words, the things that people were having to report anyway, but consumers
can’t actually even get their hands on. That beyond this single side with
software, and here is the ask, to help them visualize consumers, visualize data
in meaningful ways. So ideally, you could plug in some of the parameters
important to you, and you would end up with, well, what does that plan look
like for you? It’s a little bit like the Medicare Part D thing.
So these things, we thought one, take advantage of something already in
progress and mandate, i.e. insurance exchanges, and take advantage of data
already there, and report it. But make them much more understandable, much more
tailored to the individual.
So the second set of major recommendations really our conclusion was we do
not currently have measures that matter, just like we don’t have measures that
matter to clinicians. We have to develop consumer measures. There is additional
research and development needed to understand the factors that consumers use to
make health plan and provider choices, and how to support those decisions with
appropriate information. We need training materials or software to be
developed, that would support the consumers’ assessment process about
health-related matters. And these measures should be developed with consumer
input and tested with consumers. So that is sort of our what can be done by
2014 and what can be done with future research agenda.
DR. CARR: So again, I think it was a great model to have time to deliberate
the day after the hearing. That was great, and it shows what you synthesize. I
think what you are doing also ties in nicely with what Todd Park is trying to
do. I think it is almost a perfect use case for what they might be thinking of.
And so, you’ll be putting together a report or a letter?
DR. TANG: We will put together a letter with these elements.
DR. CARR: And do you have any immediate plans for another hearing at this
DR. TANG: We do not. Let me make sure other people on the subcommittee have
anything to add to what I summarized.
DR. COHEN: Great summary. The question is, it was a rich discussion, I don’t
know that we can convey it all in a letter. I would like us to think a little
more about whether there can be, I don’t want to spend a lot of time doing a
full-blown report. But maybe there is something between a report and a letter,
to capture the richness and subset plans for the future.
DR. TANG: So the process we thought we’d do to accommodate that which is
let’s get the letter out, because that is one of the things we have. And then,
if we find such rich and voluminous things we have left out, we either create
an appendix or an additional report. How does that sound?
DR. CARR: Very good. All right, Populations?
DR. GREEN: We did three things. We planned our workshop five days from now,
refined the questions, the background, identified where we still had gaps, and
got it to a point where it can still be posted and ready to go. The second
thing we did was have a discussion about dissemination of our working reports.
I think the crux of that discussion is there is a performance gap for NCVHS,
that after NCVHS does good stuff, almost no one seems to know about it. And we
don’t have a systematic way of defining what Justine often refers to, he’s a
customer for this report.
And at least this subcommittee of the full committee believes that we should
get much better at knowing who needs to get reports that we do and the work
that we do, and that we should adopt a multi-faceted dissemination system. We
miss Linda Klaus. She provided us with a template, and the committee agreed to
receive the template by email. And over the next few weeks, we plan on
circulating this template to identify potential audiences for NCVHS work,
population work in particular, and how that might go. And we will continue that
The third thing we did is we had a robust discussion about what we might
want to do next. There are eight possibilities, and we are going to try to
summarize those and get them circulating with the committee again, so that we
can look at prioritizing some next steps. They are pretty much unified around
the communities and learning health system.
The committee feels it is totally capable of addressing more than one thing,
and so we are excited. Vickie has done yeoman’s work here in getting this
workshop together for next week, around the SES stuff. So the committee, I
believe, feels comfortable, feeling like we can run on parallel tracks with
more than one thing at the same time.
DR. CARR: I don’t know if you and Walter have had a chance to speak, because
there is some work on population health standards, or public health standards
DR. SUAREZ: We haven’t talked about it, but I mention it in my report.
DR. CARR: We want to make sure that that gets in the mix of the discussion.
Anything else to add? Great job. I want to go to Privacy.
DR. FRANCIS: Just very briefly, we are planning a hearing of the
17th of April. With having learned from Quality, the hearing will be
on the 17t,h and then there will be subcommittee discussion on the
18th. The topic is next steps for community data use. Our audience
is local communities who want to use data.
Our problem, I will just go back to saying, trust, and our focuses are going
to be what do we know about possible modes of governance, that would be beyond
data use agreements. What do we know and could recommend about dealing with the
problem of small area, and potential problematic inferences. What are good ways
of communicating to the public what uses are being made of data. And finally,
what is known about, there has been some pretty significant research and there
also are groups with some important views about.
So what do we know about people think about what would surprise them
unfairly, with respect to data use. That is on the table for the
17th. People, feel free to enhance the accuracy of my summary, but
that is a rough quickie.
MS. GREENBERG: I just want to say I am so pleased that the quality
subcommittee really set the standard for leaving time and actually setting
aside like a half day for processing what you hear in the hearings. I think
that is so important. It is something that we have tried to encourage in the
past, on the staff level, but we recognize that you all have day jobs, et
But I think in the long-term, it is a very efficient way to use your time,
because you are already here, and we all know how difficult it is to schedule
teleconferences and all of that. And also, how you can kind of forget what
happens when the days go by. So I really strongly support that and we are glad
to work with you, to accommodate that for future hearings by all of the
DR. SUAREZ: Okay. Actually, we have prepared slides and a lot of animation
and all that, but in the interest of time, I am just going to talk. So very
quickly, the subcommittee on standards, we are looking at the follow activities
over the next year basically. We just completed, and thank you everyone for
your engagement and participation in the letters about claim attachments,
section 10109, and the standards and operating maintenance process.
So the next step is going to be very quickly complete the process for
identifying and recommending and coordinating authoring entity of the operating
rules for the remaining transactions. And then, bringing that recommendation
very quickly back to the full committee, and we expect that to happen in the
next two months actually. We are going to be moving very aggressively on that,
because of the timing of the development of those operating rules for the
remaining five transactions.
We are going to be developing, as part of the letter that we just approved,
a strategy for the section 10109 next steps. As you recall, the letter called
for committee to develop a strategy by June of this year, so we are going to
develop that strategy and what needs to be put in place, in order to address
all those five areas on section 10109.
And then, as we mentioned, we are going to hold a hearing in June, around
June, we don’t have the dates yet. And this will be a relatively large hearing
in the sense that, the intent originally was to cover before the announcement
of ICD-9 delay. Normally, we have ever year a review of where things are with
the implementation of the standards every year.
So this year, we are going to have how are things going with 5010, how are
things with going with the planning of ICD-10, how are things going with the
planning of the operating rules implementation. But this particular hearing is
going to have to have a dedicated component on where are things with the
standards, in terms of 5010, what are some of the issues around that, and the
preparation for the operating rules.
And then, I think we are going to have a separate complete hearing on
ICD-10, that addresses all these questions and brings together the NPRM
hopefully by then, if there is an NPRM out. We can include that, and then
prepare a more detailed separate letter with recommendations regarding ICD-10
transition. So that will be sort of a major activity in our June timeframe.
And then, we are going to be working on the development of the
11th HIPAA report to Congress, which we hope to be able to finish by
the fall of this year. And then, in the fall, we are going to be looking at
working jointly with the population health subcommittee on a new topic, new in
the sense of a topic that is separate from HIPAA. And that is the topic of
public health data standards, the status of the development and implementation
of public health data standards. And when I mean data standards, I refer to
generally the standards are used to communicate data and messages between
public health and the rest of the industry, and between the providers and
payors in public health.
So we are looking at convening a hearing again jointly with population
health in the fall, to invite public health groups and organizations, and
invite certainly a number of other groups, to talk about the status of public
health data standards. So that is our agenda, as we have it planned now. We
even have an early version of the agenda for 2013, but we will just have it
there. Any comments, Judy?
DR. WARREN: No, other than we are also looking at some alternative
strategies for developing the HIPAA letter that, as we work with Justine, will
probably be contacting the rest of the committee about how we will go about
that in a much more efficient way than what we have done in the past.
DR. CARR: Right, the report to Congress, I think that is right And again, I
think we ought to begin to think about what is this year’s letter short,
focused targeted, informing Congress, this is where we are trying to get to,
this is where we are today, this is where we see the challenges.
DR. WARREN: One of the organizations that we thought of the letter, instead
of the letter being drafted by the standards subcommittee, that we actually get
a task force of the full committee, representing each of the subcommittees, to
work on that letter, because the letter really has become one of the full
committee. Initially, it was just looking at the status of implementing the
standards that HIPAA had, and now it is has kind of moved a little bit beyond
that because it is so involved in quality and so involved in populations. And
it has always been with privacy and standards.
DR. CARR: So I think as we said before, the co-chairs of the subcommittees
ought to think about what their message is. Sally, did you have a question?
MS. MILAM: I have a question for Walter. At I think it was our May community
health hearing, we heard from NAHDO that one of the challenges around public
health data were the differences in definitions utilized to public health, and
that would thwart our efforts at linkage. And I’m wondering if the work with
the standards group in the fall will take up that sort of issue, as well?
DR. SUAREZ: Absolutely, I think that will be one of the components of our
focus. Thank you, that’s great.
DR. GREEN: I would like to echo Sally’s quote about that is an exciting
proposition. I want to stake out a little territory here. I don’t think we
should go into that hearing or workshop with the assumption that there is such
a thing as public health separate from individual health and vice versa.
This is a 100-year old concept that many people think needs to be healed.
And in between now and the fall, there will be IOM reports coming out about
trying to integrate public health with health care. And that can make the
September meeting very timely. But I just want to advise caution about framing
it as if we are going to come out with recommendations for one side or the
other side. What we are really looking at is integration.
DR. WARREN: That has always been a problem for me, to try and figure out how
things are different for the individual and how they are different for the
public, because I see it as continuum.
DR. COHEN: I just wanted to add to that comment. When we think of public
health, we should think of public health more broadly, not in its
disease-specific focus, but that it embodies a quality of life and well-being
in the broadest of UHO sense. This really expands the notion of developing
public health standards to less traditional concepts that people envision as
part of public health.
MS. GREENBERG: I might point out that your statement associated with the
60th anniversary symposium was very eloquent on these very issues,
and it never hurts to go back to some of these things.
DR. MAYS: Talk about timeliness, the issue of quality of life and well-being
was issued in terms of Healthy People 2020. And they haven’t developed anything
about that, and that is going to be the next work. The report from IOM that I
sat on the leading health indicators that actually outlines recommendations
about what should be taken into account, so you might want to do that in terms
of also the other report coming out.
DR. CARR: So Walter has got a very full agenda, standards has a full agenda.
But as populations thinks about your priorities, I think this is an important
one to consider, and it might be able to start sooner and align with the
Healthy People 2020, as well, or at least to inform, find out what everybody is
DR. SUAREZ: I just want to say this is so exciting. I think this is really
bringing us back to one of our origins, which is public health. I am just
extremely excited about the opportunity to work on this particular topic this
year, and certainly partnering with public health, the population of health
DR. CARR: So a couple of things, Debbie has a couple of announcements. I
want to be realistic getting this letter done. Do you have a draft that is
ready? I think we agreed on the three concepts that need to be in the letter,
and I think what we have to do is have the executive subcommittee sign off on
the final thing. So it is those three concepts only, in the way that we heard
from everyone in the room. Is there any objection to having the executive
subcommittee finalize this letter, either later today hopefully or Saturday?
We went through and we said do we all agree that, if there is a delay, we
need to use the time efficiently to understand how to eliminate the obstacles
that currently exist. Second, we want to make mention of the fact that there
are stakeholders on all sides that are bearing a financial burden. So slowing
down fixes one and solves the other, speeding up fixes one and solves the
other. We wouldn’t say it that way, but simply to say there are stakeholders on
all sides with financial challenges that need to be better understood.
And then, thirdly, that we need to harmonize on the fact that SNOMED is the
clinical terminology mandated in meaningful use, and that we need to find ways
to work with that and use the mapping, et cetera, to make it easy to map to
DR. SUAREZ: Are we going 6to be silent about the time?
DR. CARR: I think what we want to say is, if there is one, we want it to be
as short as possible. But during that time, it is important to understand the
DR. SUAREZ: Can we say as sure as possible, and I don’t know, hopefully it
is the right word, but no longer than a year.
DR. CARR: The NPRM is going to provide ample opportunity to address, I would
imagine, that timing.
MS. GREENBERG: But it is also going to make a recommendation on that timing.
DR. SUAREZ: We are trying to make a recommendation about what to put into
the NPRM. That is the main purpose of this letter. Otherwise, we can wait. We
can make arguments about why it would be valuable.
DR. MAYS: The one thing this letter addresses that will not be in the NPRM,
and if we don’t send, won’t even show up in the NPRM, which means no one can
address it, is this whole thing of harmonizing the use of SNOMED to get us to
ICD-10. And also, to call attention to the expenses of those people who already
implement it, and will have to maintain dual systems.
DR. CARR: So let’s have a show of hands, how many people would favor the
specific language of limiting the delay to one year, all in favor? Any opposed?
Do you have an alternative recommendation, either Jim or Paul?
DR. WALKER: Yes, that it be guided by the analysis of the cost and benefits.
To me, it is specious to say it should be a year, when we have no agreement in
this committee whatsoever about what the relevant cost and benefits are. What
are the costs exactly? If we are going to bear a bunch of them, but I don’t
know what they are in any even semi-quantified way. So my concern is that we
will just be making a recommendation, I don’t know based on what, but not on
anything like evidence.
DR. TANG: I guess I am a little less worried about the quantitation of the
cost so much as what if she adopts this other strategy. Let’s take the time to
get that strategy right, that could have long-term benefits.
DR. SUAREZ: What is that other strategy, Paul?
DR. TANG: Basically focus on clinical terminology.
MS. GREENBERG: Do you want to wait until everyone in the country has
implemented SNOMED CT, the hospitals and the physicians?
DR. TANG: Exactly.
DR. CARR: So this is an important question. It seems like with two
objections, the consensus of the committee is —
DR. SCANLON: My issue is the specifics of it, and the notion that as soon as
possible is strong language. And it can be buttressed by the fact that you
point out that delay is costly. And so, I think that is the position that is
totally defensible. If you try to defend a year, and people have evidence, it
DR. CARR: Okay, as soon as possible, in bold, underscored. I think we then
have unanimity, the members of the committee, any objections? Okay, as soon as
possible, caps, bold, underscored. I believe with that, this meeting is now
adjourned, and we can have lunch. And then, Todd Park, I believe, hopefully
will be here by 12:30 and we will have a chance to chat about this data
initiative, for those who are interested, followed by the rest of the afternoon
session that begins at 1:00.
MS. GREENBERG: Is the plan then, if the executive subcommittee will final
assist by when, by Monday?
DR. CARR: Yes, by Monday, unless Monday is too late, but yes, by Monday, as
soon as possible.
DR. JACKSON: I just wanted to underscore what Larry reported out for the
dissemination plan for populations. Really, it kind of applies to all of the
subcommittees. What you will be receiving for the executive subcommittee is the
template for dissemination that Linda Klaus has pulled together so nicely. And
in your work projects, just start keeping the template in mind, start filling
out listservs and programs and organizations and meetings that you see would
really use our products.
I had to step out of the room because the tenth report is being prepared in
the same manner as this, as the community report, in time for the tenth summit.
So you will see that when we go to the meeting at the end of March. All of
those things, just kind of keep in play to keep our materials out there to
public, the folks who really want and need them.
(Whereupon, a luncheon recess was taken at 12:03 p.m.)
A F T E R N O O N S E S S I O N (1:00 P.M.)
DR. CARR: Before I call the meeting to order with regard to the letter to
the Secretary about ICD-10, so let me just see if we have a quorum. I believe
we have one, two, three, four, five, six, seven, eight, nine, ten, eleven
members. And are we able to find Bruce or Marjorie?
DR. SUAREZ: Just a matter of process here, since you before lunch, adjourned
DR. CARR: I adjourned the meeting – Marjorie corrected me. We will be
just a moment while our visitors get here.
The letter up there is the usual introduction. The next paragraph, this is
the thing we agreed on, and just taking us to the next paragraph, the same as
we said, unprecedented times in health care, pace of change, ironically our
success and the pace of advancement has become our challenge.
Nearly 20 years ago, NCVHS introduced the importance of timely conversion to
ICD-10, and we are going to attach our timeline code sets. We wish to recommend
to the secretary, if you choose to delay the scheduled implementation of
ICD-10, you address in the notice of proposed rulemaking the following three
issues. One, use the time to identify and address the obstacles to
Two, evaluate the financial impact on the communities that are ready to
implement, and will carry the financial burden of maintaining two systems, as
well as the financial burden borne by those who have not been able to yet begin
their transition. And three, take this opportunity of conversion, converting
the ICD-10 classification system from ICD-9 to 10. Do it in such a way that
aligns this rule with the meaningful use rule set, that specifies SNOMED as the
standard clinical terminology for quoting diagnoses.
That would mean that clinicians would document their diagnoses in EHR using
SNOMED, which would be mapped to ICD-10 using the NLM-developed map on the
back. And in this way, each coding standard is then used for the purpose for
which it was designed, thereby helping to mitigate the ICD-10-CM user interface
challenges. The NLM-developed iMagic tool is a good example of user friendly
interface tool for the conversion from clinical language to the structured
And this is four, the committee strongly urges that the issue be resolved as
soon as possible. Now, this is the point that I want to raise. At the break,
there was significant discussion about the impact of not including the one
year. And we did not officially take a vote on this, and I realize a couple of
our members have left, but we do have a quorum. So what I want to do is put to
a vote for the committee, the consideration of adding the within one year, and
follow the rules of the committee that a simple majority is sufficient to pass.
And so, I will ask then for a show of hands regarding the inclusion of the no
later than one year as stated, I guess.
DR. SUAREZ: Couldn’t we make a couple of comments about that letter, or did
you just want to?
DR. CARR: Very quickly.
DR. SUAREZ: The first one is this number four, the committee strongly urges
that it should be resolved. I don’t know that really the idea there is that the
committee strongly urges that, if there is a need for delay, that such delay be
DR. CARR: Concluded as soon as possible?
DR. SUAREZ: Exactly, and no later than one year.
DR. CARR: That the delay be concluded as soon as possible, and no later than
one year. So I want the committee members –
DR. WARREN: Where does the delay go in this?
DR. CARR: That the delay be concluded as soon as possible. I want to get to
your critical point, Marjorie, so we will go back and let the Executive
Subcommittee fix that. I think we understand the intent of that. We want to be
respectful of our guests, and I want to just get the issue resolved as to the
will of the committee, with regard to adding the phrase within one year. The
delay should not be more than one year.
MS. GREENBERG: That the delay should not be more than one year.
DR. CARR: Right, that the delay should not be more than one year.
MS. GREENBERG: This sounds like the issue should be resolved in a year.
DR. CARR: Judy, have it the delay should be concluded within one year. That
is the proposal. I would now like a show of hands from the committee members
all in favor.
One, two, three, four, five, six, seven, eight, nine, that is a majority.
That is a simple majority, and so, the will of the committee then is to include
that delay should not be more than one year. Is there any discussion? So we
will add that back in and send the final version to the Executive Subcommittee.
DR. COHEN: Justine, just a point of order, you might want to find out the
people against and the people abstaining for the record, since it is an
DR. CARR: People against, Bill is against. People abstaining, I am
abstaining, Leslie is abstaining, okay. So the committee is in favor. We will
add that back in.
So that concludes the discussion of this letter. The final copy, hopefully
we can circulate today to the Executive Subcommittee.
DR. WARREN: Just a friendly amendment, the delay of the ICD-10 code set
implementation should be.
DR. SUAREZ: Yes, I think we need to fix the wording. But the idea is that,
if there is a need for any delay, that that delay should not be more than one
year. I think that is the main principle.
DR. CARR: So thank you, that concludes that. We now go onto our very
exciting program of this afternoon. Let’s start by going around the room, and
then I would like to ask Todd perhaps to give us some introductory comments. I
will begin, I am Justine Carr, from Steward Health Care System and chair of
DR. SUAREZ: My name is Walter Suarez. I am with Kaiser Permanente, a member
of the committee and co-chair of the standards subcommittee.
DR. MAYS: Vickie Mays, University of California Los Angeles. I am a member
of the full committee and a member of the subcommittee on population health and
DR. SCANLON: Bill Scanlon, National Health Policy Forums, member of the full
committee and the standards subcommittee.
DR. HORNBROOK: Mark Hornbrook, Kaiser Permanente, member of the full
MR. BURKE: Jack Burke, Harvard Pilgrim Health Care in Boston, member of the
full committee and member of the population health and privacy subcommittee.
DR. TANG: Paul Tang, Palo Alto Medical Foundation, member of the committee,
and quality and privacy subcommittees.
DR. COHEN: Bruce Cohen, Massachusetts Department of Public Health, member of
the full committee, and member of the population and quality subcommittees.
MR. SOONTHORNSIMA: Ob Soonthornsima, Blue Cross Blue Shield of Louisiana,
member of the full committee and the standards committee.
DR. FITZMAURICE: Michael Fitzmaurice, Agency for Healthcare Research and
Quality, liaison to the full committee, staff to the subcommittees on quality
DR. GREEN: Larry Green, University of Colorado, member of the full
DR. FRANCIS: Leslie Francis, University of Utah, member of the full
committee and co-chair of privacy.
MS. GREENBERG: Good afternoon. I’m Marjorie Greenberg from the National
Center for Health Statistics, CDC, and executive secretary to the committee.
And I want to welcome all of you and am very pleased to be hosting the session
DR. WARREN: I’m Judy Warren, University of Kansas School of Nursing, member
of the committee and co-chair of standards subcommittee, member of the quality
DR. CARR: So I too wish to welcome all of you. This is very exciting for us.
I think a year ago, Todd came and we spoke about many of our areas of
alignment, and are very excited about our working together. So Todd, I wonder
if I could ask you to give us sort of an overview of what we can expect today.
MR. PARK: Hello, everyone. It is fantastic to be back, and talking with you
about the power of data and innovation to improve American health care. It was
wonderful to speak with you all a year ago, and we are back with a lot of
exciting stuff to talk about.
As we had mentioned a year ago, the department, HHS, is engaged in a large
scale and growing effort to really maximize the ability for data information,
to help improve American health care, while rigorously protecting privacy,
confidentiality, really unleash the power of data information, to help catalyze
transformation in health and health care.
And the specific form that this has taken, the kind of flag around which all
of this activity is rallying is the Secretary’s health data initiative, which
is an effort to really turn HHS into what we are calling the NOAA of health
data, in this case, N-O-A-A, the National Oceanic and Atmospheric
Administration, which famously in certain circles has for decades not just
collected a lot of weather data, but has chosen to publish it in downloadable
They have made it available to anyone and everyone for free, which is then
directly fed a host of innovations, rapidly growing still in the world outside
of government, everything from weather newscasts to weather websites, mobile
weather apps, weather research, weather insurance, et cetera.
The government made a similar play in the 1980s when it liberated global
positioning system data, which of course now powers everything from Foursquare
on your iPhone to super tanker navigation systems in your car and everything in
between. So Healthy Initiative is the government’s latest effort to run this
open data, open innovation play, while obviously taking into account the very,
very specific considerations that are extremely important around privacy and
confidentiality, with respect to key transits of the data.
The whole idea is to encourage HHS and other sources of data, while
respecting privacy and confidently, to actually make data more accessible,
machine readable and used by a whole rapidly growing host of innovators across
the country, to leverage the data as fuel, mashed up with other information,
other capabilities, to build tools, services, programs, features, capabilities
that help consumers and patients to control their own health and health care by
giving them the information they need at their fingertips, of everything from
how to basically pick the right provider for their family, to what the latest
and greatest information about diabetes is, to clinical trials that could save
their life in the area. Tools, services, programs, capabilities that help
doctors and hospitals get ever better, continuously improving care, that help
promoters promote health and wellness, that help researchers advance the state
of the art in understanding any number of different deals, to help local
policymakers, public health officials make better decisions.
And the fundamental insight that we have really acted upon is what I
mentioned a year ago, Joy’s Law. It is a famous law from Paul Country(?), of
Silicone Valley. It is attributed to Bill Joy, who is the co-founder of Sun
Microsystems and a legendary figure in Silicon Valley, who once said, no matter
who you are, you have to remember that most of the smart people in the world
don’t work for you, which is of course true.
And if you really want to maximize social return on data, the idea here is
to not just have your own smart people work on the data and turn it into tools
inside, but to really have everyone who cares about health and health care
improvement, be able to respect privacy or data sets. Be able to access and use
that data to develop insights, tools and services that can help advance the
well-being of the American people. And what we have seen over the last couple
of years, that has helped the initiative rolling forward, has been living proof
of the truth of Joy’s Law.
What has happened as we, A) made brand new data available, B) actually maybe
less successfully but equally importantly, taking existing public data that was
in, say, books, PDFs and static websites, and turn it into forms useable by
third parties, i.e. machine readable downloadable data can actually ingested by
other websites. They are doing that, as well. And on top of that, promoting and
educating the world about the availability of this data. One interesting
insight that we had was about a 95 percent of innovators who could take our
data and turn it into useable products and services, who didn’t even know we
had the data. They didn’t even know that it was available to them. So we have
been engaged in a whole series of meet-ups and workshops and webinars and
whatnot, to get the word out about this data, both brand new data and data that
was previously public, but not turned into machine readable forms.
What we have seen is an absolute explosion rising tied in innovation across
the country, at a grassroots level, by non-profits, for-profits, students,
researchers, universities, entities large and small, who have an unbelievable
tide of activity, taken various swaths of data, and leverage it to build new
features, products, services, capabilities, et cetera, to help consumers, help
doctors, help employers, help journalists, help communities get the information
they need, make better decisions, and improve health and health care, which has
just been spectacular. And we, of course, actually I think they most notably
have seen this in our annual health datapaloozas, the health data initiatives
forums that the secretary has hosted. At the last health datapalooza, which was
June 9th, 2011, actually at the end of a process, in the American
Idol style process, where an independent panel of judges narrowed down a whole
field to ones that actually fit in the Natcher Center at NIH. We got the
opportunity to see 50 innovators present tools, services, capabilities that
they had built, that were live in the marketplace, had sustainable models of
operation, that were already collectively serving, just these 50, tens of
millions of people. It was just an astounding display of American mojo and
What we actually say now is that if your faith in American is wavering even
a smidgen, go to the Institute of Medicine’s website, a co-sponsor of the
health datapalooza. Look up June 2011 Health Data Forum, watch as many of these
50 videos as you possibly can, because it is the most awe-inspiring display of
American mojo that I have ever seen. And certainly put it best that day, we
said look, if you look at just these 50 innovations, it was just a small subset
of what is actually going on. Just these 50, no one organization, no 10
organizations, could even have dreamed this stuff up, let alone have actually
built it and deployed the scale that it is already helping millions and
millions of Americans.
We are, in the Secretary’s words, doubling down on health data initiatives.
So actually after the June health datapalooza, the Secretary issued an
executive order in August that asks every agency to publish a data access and
use improvement plan every six months. We actually just got our first draft of
plans in November, produced by senior health data leads that every university
now has, and they are spectacular. They are spectacular to the point where
actually, on our healthdata.gov site, which is the catalog we started about a
year ago of all of our fully open data, the amount of data on healthdata.gov is
roughly going to probably double in the next six months as a result. It is just
We are discovering that we have had data that we didn’t even know we had.
And then we just decided that, now even people within HHS realize. And think
about it, across all the agencies, of course it helps. So it’s actually helping
us to use our own data, on top of actually making it available, of course, for
lots of other folks to do useful stuff with it.
On top of that actually, it sounds like a very — detail, but I think it is
a critically important detail, we are engaged in a major upgrade of
healthdata.gov, because that is increasingly kind of the storefront, if you
will, that we are using to present this data to the world. The outside world
does not know how HHS is organized. They can’t navigate our 100 million
webpages to find the data sets they are looking for. They need one place to go.
And in an act of national service, Dave Forrest, who is the CTO of the
legendary Motley Fool website, has volunteered, we are paying him but very
little, to serve his country, join HHS and lead over the next few months a
massive upgrade to the healthdata.gov site. That is going to make it vastly
more usable, and he is going to turn it into much more of a site ultimately. He
has a vision for it, as an information utility of unparalleled power. We are
very grateful to David, who is here today, for leading that effort.
And actually, speaking of datapaloozas, we have announced that our third
annual health datapalooza is going to be held June 5th and
6th in the Washington Convention Center here in D.C. We had to pick
the convention center because it was the only site that could hold all the
awesomeness that his being generated. It is a two-day event this year.
And it is not going to just showcase the winners of the American Idol
process. We are going to have actually panels of patients, panels of doctors,
panels of community leaders, who actually judge the entries coming in for
health datapalooza. But there are also going to be genius bars set up, where
owners of data can actually set up shop and explain to any interested person
what this data set is and how you might think about using it, what it means,
what it doesn’t mean, so on and so forth. It should be incredibly cool. I can’t
think of actually a way to have more fun than to hang out at the health
datapalooza. I literally cannot.
And that brings us to our agenda for today. First of all, we are absolutely
delighted to be partnering with NCVHS on this great endeavor, to maximize
social return on health data. I got the report from earlier today that folks
here are very excited about the new NCVHS subcommittee on health data access
and use. Thank you so much for pursuing that. It is going to be extremely
exciting and an invaluable aid to us and others, as we seek to improve data
access and use, and improve social return on data.
And in another action beat, I am going to turn it over to the truly
incredibly exciting part of our presentation. Now, Brennan is here to talk with
you about a very significant move that CMS is making, a move to establish what
we are calling a data and information products line of service or line of
business, if you will. So to kind of put it in a nutshell, and I am saying this
on behalf of the Secretary, Deputy Secretary, Marilyn Tavenner, Michelle
Standard, who thought she was going to be here, but I am channeling Michelle.
We actually believe that over the long term, information and data parts from
CMS should be considered to be a line of business, on the same level as
Medicare and Medicaid, that is a national treasury, the data that CMS has, that
can be turned into a significant catalyst for health care transformation,
improvement of the well-being of not just Medicare or Medicaid beneficiaries,
but actually of the country as a whole.
And then, the fundamental idea is that, as opposed to thinking of data as a
byproduct of what CMS does, the idea in a nutshell is to think of the provision
of data innovation products in a responsible way respecting privacy, and
well-constructed and thoughtful approach, that the provision of data innovation
products is actually at the core of what the Center for Medicare and Medicaid
Services should be doing for the United States of America. And to make it a
priority, to make something that we are very proactive about, and to work in
close partnership with the health businesses broadly to make sure that the idea
does help produce maximum social return, in terms of health and health care
This is a really, really, really big deal, and we are so excited about it
that I can barely see straight. And it has gotten an enormous amount of support
at CMS and HHS and broadly across the government. So without further ado, I
should turn it over to the guy who I just can’t describe how much I love you,
Niall. I just cannot describe how much I love. I am going to turn it over the
MR. BRENNAN: Thank you very much for those kind words, Todd. Following Todd
is one of the more unique challenges in health policy yet, speaking challenges.
I would also add, thanks to Mark for joining me today, and thank you to the
committee for letting us present on our vision for where we want to go with
I would like to start with just a couple of introductory scene-setting
slides. None of these will be particularly earth-shattering discoveries for a
committee of data mavens like yourselves, but just to put things in context. So
obviously, CMS is the largest single payor for health care services in the
United States, with over 1.5 billion claims submitted annually. We directly
administer the Medicare program in all of its different facets and work closely
with states on the Medicaid side of the house.
In addition to all this data, there will be significant additional data
sources on the way in coming years. They have already started flowing to the
agency, electronic health record data, Medicare advantage plan encounter data,
and obviously beginning in 2014, additional Medicaid and health insurance
exchange data. But it is really not just about the claims data. We literally
receive billions of other non-claim data points in any given month or year.
There are eligibility verification checks, there is quality data, there is
visits to the website, there are calls to 1-800-Medicare, and it all represents
the very wide range of data that we feel we are only scratching the surface of,
both internally and externally.
Also, with the passage of the Affordable Care Act, actually prior to the
passage, we had already begun a transition from a passive payor to an active
purchaser of health care. We are expected to drive a new innovation in health
care. And obviously, the last bullet is very important. We take very seriously
our commitment to maintaining and respecting beneficiary privacy in all these
One of the first things that CMS has to consider any time it thinks about
releasing data is the complex and interlocking legal framework that confronts
us. And release of CMS data is governed by several different legal constructs,
the Social Security Act, the Privacy Act, HIPAA, ACA, SAMHSA and FISMA. And
some of them say you should do this, and some of them say you shouldn’t do
this. And so, it is kind of trying to thread needles that aren’t always lined
up in a straight line.
As a result, traditionally speaking, we have generally provided beneficiary
identifiable data for traditional IRB-approved university research, to support
demonstrations that we sponsor at the agency, and for quality improvement
organizations. We have not traditionally utilized HIPAA provisions to make
disclosures to covered entities. And as Todd so eloquently and passionately
described, we believe that the result is that the health care system is not
benefiting from optimal use of CMS data.
We have been active in soliciting input from key data stakeholders over the
past several months. In fact, we had a fascinating and inspiring data summit,
we called it, just before Christmas, where we had a whole parade of both
existing and potential data users, and people who really know what is going on
when it comes to health data analysis.
And so, this is some of the feedback that we got. States really need timely
data for Medicare and Medicaid care coordination, but also for more general
research and population health purposes. All payors claims database, other
payment reform efforts, and quite frankly, the current process is not really as
customer friendly as it could be, regarding getting data out to states,
particularly when you view them as who are really a trusted partner with us in
administering the Medicaid program.
Providers need data on the beneficiaries they serve, to permit and enhance
care coordination and patient-centered care. We have made some important steps
here with recent decisions to give ACOs and ACO providers both quarterly
performance reports and monthly beneficiary level claims feeds. But we feel
that this may only be the tip of the iceberg, and we also need to make sure
that we have the necessary IT infrastructure and processes in place, to make
sure that we can do that in an efficient and effective manner.
And researchers, the ones that I have talked to say that CMS data costs too
much, is too old and that it comes with too many strings attached. And
specifically, we are increasingly aware that our current research data process
may not be designed to support this recent advent of broad-based research
inquiry, big data analytics, et cetera, et cetera, where we tend to look and
ask for very specific hypotheses, that may not necessarily be available on the
front end. And in doing so, we are sort of again potentially restricting the
ecosystem with the backend knowledge that might emerge.
The problem is not just restricted to external users. When you generate as
much data as CMS does, there can be internal challenges, too. CMS staff
struggle to a surprising degree to get access to CMS data in a timely manner.
It can be enrollment data, it can be spending data. And we also need to harness
data ourselves in ways that we never have before.
Obviously, this isn’t strictly linked to the transition from a pair of
claims to a value-based purchaser. But when you think of the Affordable Care
Act, when you think of value-based purchasing, when you think of ACOs, quality
resource use reports, the list goes on and on and on. You can’t do any of that
unless you have a very firm handle on your own data, and unless you are getting
the most recent data possible, and you can set baselines, identify problems,
measure progress, et cetera, et cetera, et cetera.
As I alluded to a little bit earlier, we have actually made very significant
progress, even in the last 12 months, in terms of data dissemination at CMS. We
are providing data to both MSSP and pioneer ACOs, both quality reports and
monthly beneficiary claims data. This is a real game changer, as far as I am
concerned, because for the first time in a fee for service environment, it
enables patient-centered care. Providers will not just know the interactions
that they are having with their patients, they will know the interactions that
those patients are having with all of the providers.
We also have the Medicare data sharing for performance measurement program,
otherwise attractively known as section 10332 of the Affordable Care Act, which
will provide 100 percent extracts of Medicare A, B and D data to qualified
entities for performance measurement purposes. So the really exciting part
about this particular program is that, in order to receive the Medicare data,
obviously there are rules.
One of the key rules is you are not going to get Medicare data unless you
have claims data from other sources, that you undertake to combine with the
Medicare data, and create our provider performance reports. So this creates a
framework and a potential to go from the now, where providers are receiving one
report from Humana, one report from Aetna, one report from United and nothing
at all from Medicare, to potentially receiving a single report, done in a
standard manner, covering all or most of their practice. We are very excited
about that particular provision. We are accepting applications, the program
went live on January 9th of this year.
Finally, and very, very importantly, we will touch on this a little bit
later, too, we are trying to forge ahead in the creation of additional
non-beneficiary identifiable data sets. Todd mentioned the health data
initiative. CMS is a very enthusiastic participant in the health data
initiative. We contributed, for the first time, hospital referral region,
Medicare spending and utilization data to the health indicators warehouse,
which was a major initiative under the health data initiative. We also have a
lot of information on quality measures, Medicare beneficiary, demographics,
disease burden, et cetera, et cetera.
You can now go to the health indicators warehouse, and for any given
hospital referral region in the United States, have access to a wealth of
aggregated information that, while it is not necessarily beneficiary level, it
is, I like to call it, conversation starter information. Why is imaging
spending in area X 50 percent more than area Y?
So all of these new approaches are really just what we feel are the leading
edge of a new wave of CMS data users. In order to ensure that it doesn’t engulf
us whole, we are going to need to ensure that future data release processes
will permit 100 percent extracts of data, across multiple years, on a routine
basis, enable analysis across multiple care settings, allow for the routine
creation of customized analytic files, and accommodate large increases in the
number of data users and the volume of data that they are demanding.
So what is the solution? We are going to get Todd to do it. Todd is going to
get me to do it. We are going to employ advanced analytics to create actionable
information products. We are going to establish new policies to support more
use and reuse of CMS data. We want to expand the pool of CMS data users, while
maintaining appropriate beneficiary protections. For example, we are exploring
the establishment of data enclaves or portals that would expand secure access
to different levels of CMS data for a wide range of users. Some people might go
into the enclave and have access to full beneficiary identifiable data. Other
folks might go in there just to get their ACO report or their provider
performance report. Other folks would get deidentified data.
MR. PARK: And one benefit, of course, being that the physical copy of the
data doesn’t move outside.
MR. BRENNAN: Exactly, because under current processes, we do a lot of
cutting and shipping. And while we have a DUA process to a large extent, we are
somewhat reliant on the good graces of data receivers, not to act in an
inappropriate manner with the data, because there is not a whole lot of
follow-up, auditing or monitoring that goes on with existing data users. So an
enclave would be a significant improvement in that regard. I missed the punch
line. And so finally, as Todd mentioned earlier, we are committed to
establishing a dedicated data and information product line of business at CMS.
How will this transformation affect CMS data users? We believe it will
result in data that is more timely, more accessible, more intelligent and more
flexible. Will it happen overnight? No. Are we committed to doing it?
Absolutely. I think it will be an iterative approach, and I firmly believe that
data users will begin seeing meaningful changes in the types of products that
they receive, and I think possibly just as importantly, their interactions with
the agency very soon.
And if we do that, we believe that we will support CMS in becoming a
data-driven value-based purchaser. We will make the health care marketplace
more transparent to help beneficiaries make the right decisions. We will help
providers move from maximizing the volume of services delivered to maximizing
the value delivered. We will support community and state efforts to identify
variations in health care delivery and take action that supports health and
health care improvement, the conversation starters that I talked about. And we
will help researchers of all kinds advance knowledge about how to improve
health care, again typing it back to Todd’s introductory remarks. We do not
have the firepower to fix all of the problems, so we have to leverage the
The next couple of slides are very important personally to me, and I think
for the whole concept and framework of the information line of business. We
have a lot of data at CMS. And we have said, we push it out to varying degrees
of success to some external users. But what many, many users really want is
information, not data. And so, the missing link to generate information is
analytics. We acknowledge that some people want raw data, but our data files
and layouts can be intimidating, and obtaining large amounts of raw data can be
expensive. So we want to explore ways in which we can provide users with the
information they need, without necessarily releasing beneficiary level data.
How do we unlock our data to develop insights and information for internal
and external users. And as I said earlier, without analytics based on data, we
don’t believe that we can establish baselines, identify interventions or
evaluate progress relative to our goals. And that is crucial for us over the
next couple of years.
This is just some excerpts and quotes from a publication by the Partnership
for Public Service called From Data to Decisions. I feel it is just very much
conceptually in line with where we are going. As they say, data is only the
starting point. It needs to be analyzed, turned into information and made
accessible to staff and executives, and be understandable to different
audiences. And it stresses that what really matters in this context is not
necessarily the latest whiz bang business intelligence tool. It is leadership
commitment to making decisions based on analytics and human resources, analytic
staff who can get into the data, make decisions, identify what is important,
what is not, and then get that information to the people who count.
The next couple of slides are just some simple, somewhat stylized admissions
of ways in which we are already trying to turn data into information. And I
don’t want this to appear at all grandiose. These are just readmission rates,
and to a certain extent, there is a slide before this slide, which is the
millionth of millionth of hospital claims that you have to comb through and
link in order to generate the admission rates. But these are data on Medicare
hospital readmissions from 2007 to 2010. We will have 2011 pretty soon. And one
way of presenting them to one type of data user is here is a table that you can
look at, or here is an Excel file that you can download and play with the
Other folks are more visual, and they want to get a quick sense of which
areas are doing best, which areas are doing worst. You can look at this and
very quickly determine where your problem areas might be. And again, this is
not just an interesting little map for research purpose. We have a $1 billion
Partnership for Patients program right now, being run out of CMS, that is
focused on reducing readmissions and hospital-acquired conditions. And so, we
strongly feel that this is among the information that they need to be using and
leveraging, in order to target the resources, and realize where they need to be
The previous slide was sort of a national snapshot of what is going on in
2010, who is best, who is worst. This slide amalgamates the four years of trend
data, and this is one of my favorite current slides that we have done. It tells
a couple of things from this slide. What it really tells you is between 2007
and 2010, despite a pretty broad-based national discourse about readmissions
being a bad thing and low hanging fruit, and something we really need to do
something about, readmission rates in the vast majority of hospital referral
regions in the Medicare program didn’t decline a whit. So it is a call to
action for the Partnership for Patients program.
And what it also tells you is that some places did get better, and some
places disturbingly got worse. And if you look at the little piece of blue down
on the bottom part of Texas, that is Harlingen, Texas, which is not necessarily
a poster child for health systems that operate in an optimal fashion. But
again, in terms of lessons learned and information, the QIO program had an
intervention in Harlingen over the past few years. And they had significant
reduction in their readmission rates. So we are pretty sure that there is a
link there. Can we try and disseminate those practices to other areas, or learn
from Harlingen? And likewise, there are areas that continue to get worse, which
is disturbing, and maybe places that we need to focus on sooner rather than
Another example of the type of information that we work with, an example of
non-claims information is we monitor the stock markets and the capital markets
because CMS makes billion dollar decisions every day. And about a year and a
half ago, we released a new SNF perspective payment rule. And pretty quickly
after introduction, it became apparent that there may have been some
miscalibration in the Rogue 4 system because SNF revenues rose dramatically by
And then, we flipped over to our stock market data, and we saw stock prices
for post-acute care companies rising rapidly. So we worked closely with the
folks at CM. They were aware, but again, we were able to bring multiple
information sources to a problem, and quickly get to a solution where
immediately, by the next rule cycle, we were able to adjust the payment rates.
So you see the dip where CMS proposed rule, and the stock prices and SNF
revenues gradually declined to historic normal levels.
Really what this slide does is just tell you in a slightly different way
what I have just told you, so we don’t need to spend so much time on it. So in
conclusion, we are very committed, as I said, to becoming a data driven
organization, to be meet both our existing and our new responsibilities. We are
transforming how we view and use data, both for internal and external use,
while maintaining our long-standing commitment to beneficiary privacy. We are
realigning our business practices and policies to better support data
information and development. And we are integrating data driven decision making
into our everyday work. I would be happy to answer any questions from the
DR. CARR: This is tremendously exciting. It really comes alive, all the
different displays and the very simple information that you are putting out
there. Marjorie, we have two reactors for the committee. We are looking for
Bruce and Bill to respond.
DR. COHEN: I certainly agree with your initial proposal that these data will
be of tremendous value. I am in a state health department, so my focus is going
to be less on the providers and more on state and community uses of the data. I
think in your target populations for these data, you left off consumers. You
certainly identified researchers, providers and state. I would certainly add
community and individual consumers.
I am going to focus on the population health aspects of what you have to say
at two levels. First, access to aggregated data, I think you have essentially
two targets of users. You have a variety of policymakers, planners,
community-based organizations, media, legislators, who need descriptive data
for surveillance and policy development for community needs assessment for a
variety of state and local activities.
The other thread that I will take up in a second is the researchers who need
access to individual level data, whether it is all that is confidential,
whether it is fully deidentified or partially deidentified, depending upon how
you use those terms. I am excited about making these data more available, and I
would strongly encourage you, as you liberate these data, to think about
existing resources that you can build on.
And my particular orientation and bias is for aggregated data, the current
existing and developing web-based data query systems. There are federal
web-based query systems that are very robust and get used all the time. No
wonder NCHS has vital stats, SEER stats, the EPHTN portals, WISQARS, there are
an enormous number of existing data sets that meet the needs of a variety of
users. So I think as CMS rolls out its product lines, it should take advantage
of the approach that many of these existing web-based data query systems use,
because they could address a lot of the general issues.
I think also about 25 states have their own web-based data query systems.
And some of them are quite narrowly focused on vital statics. The one,
Massachusetts, has data. We incorporate census data, we have hospital discharge
data, ED data, we incorporate data from our Department of Education around
graduation rates and employment rates. You need to consider the CMS data as one
piece of the puzzle when you are addressing public and community health. They
certainly need to be integrated into a larger context if providing these data
with other supporting data.
I would love to see CMS develop a program to work with states to help them
use and develop their existing capacity and their ongoing web-based query
systems, to incorporate these newly liberated data. And different states have
different levels. Massachusetts, our basic building block is the communities.
Some are county level, some are census track level, some are neighborhood
level. There are a variety of issues that need to be worked out. But there is
already an existing infrastructure in many states that I think is an incredible
opportunity to partner with these data as they become available. I would
certainly encourage to use those strategies.
And these web-based data query systems, not only from state governments, but
certainly the private sectors developing them, as well. And I would certainly
encourage thinking about partnership, public, private partnerships to develop
these web-based data query systems. They are used for community needs
assessment. The Affordable Care Act and IRS requirements, requiring hospitals
to do community needs assessments, is an enormous opportunity to put these data
in play, in conjunction with a variety of other data sets related to
understanding the public’s health and quality of community life. I think the
real opportunities as you move forward in developing your product lines, to
think about partnering with the existing data entities who are generating this
Let me go onto briefly to discuss access to individual level data. I think
your data enclave is spot on. Certainly, it is not a new concept. The research
data center at NCHS and through Census, is a perfect example, creating a data
enclave where there is confidential information, where people who need access
to individual level data generate queries and get the results without actually
having to hold the data. It actually benefits the user, as well as protects the
I would certainly, as you move forward in your data enclave strategy, again
integrate it with your existing federal sister agencies, and what they have
done in developing their data enclaves and providing ubiquitous access to
users. It would be great if I could go to a designated research data center and
get access to CMS data, as well as NCHS data, as well as all federal data. I
think again, this is an enormous opportunity.
In Massachusetts, we are working really hard to develop our all payor claims
database. It is being done by our sister agency, the Division of Health Care,
Finance and Policy, that primarily collects and interfaces with the commercial
entities. As a state agency partner, we regretfully heard that CMS current
rules wouldn’t allow us access to any of the deidentified Medicare data once
our state gets it. I think the current regulations, as you mentioned, really
limit access to the data that are being incorporated.
I think there is just an enormous amount of energy going on in developing
state APCDs, that should go across the commercial and the Medicare, and
incorporate the Medicaid data. I think you really need to loosen those rules
around providing those data, and working with states so those data can be used
by a variety of players. I think this has been a huge black hole in trying to
use Medicare data for community planning in our Department of Public Health. It
has been virtually impossible to get easy access to these data. I am really
happy that you are rethinking your strategy for providing access to primarily
deidentified data for more general policy development.
So congratulations, I am really looking forward to it. There are lots of
folks out there who do state policy, who do community needs assessment, who
will benefit enormously by having these data more ubiquitously available.
DR. CARR: Let me just do a check about what is the best model. I just looked
at the agenda, and we have it a little bit differently. I am sorry about that.
Did we want to do more presentation and then have a longer reactor time, Todd?
MR. PARK: It’s up to you entirely, it’s up to the committee.
DR. CARR: Why don’t we go onto the next presentation then. I know a couple
of folks have an early departure, so I wanted to make sure that you had the
DR. PARK: So should we move to the data users and perspectives panel? I
think what we will do now is move to the data users and perspectives panel, and
then bring it back to reactions from the committee. And so, Kerry Hicks is
MR. HICKS: Thank you for allowing me to participate in today’s user and
perspectives panel. My name is Kerry Hicks. I am founder and chairman of
HealthGrades, a company built on the core principles of information
transparency, provider accountability and consumer empowerment. We help
consumers make informed decisions about either a doctor or a hospital, and to a
large degree, rely extensively on CMS data to build the tools necessary for
consumers to begin making those critically important decisions.
As both Todd and Niall indicated, the federal government collects and
oversees a massive amount of health care data. This data can provide immense
societal benefit, when made accessible to responsible organizations and managed
appropriately within the current privacy and security parameters. There is a
plethora of organizations in this country, I would argue both for profit and
not for profit, which have the capability of creating enormous benefit to the
health care marketplace, and ultimately to the public good, if given reasonable
access, timely access, to the government’s vast treasure trove of data.
A bit about HealthGrades, again just two minutes on this. HealthGrades.com
is the leading consumer health care website used by consumers to find, to
research, to select and connect with a health care provider. We currently have
over 200 million unique users, coming again to select a health care provider to
our website. In January, we had about a million visits per day. We expect that
number to grow by about 30 percent per year, consistent with our past growth.
Today, we employ over 600 individuals in the major operations centers in
Denver, Atlanta, New York and Madison. We work with over 800 hospitals in the
U.S. As an underscored point, there is an enormous difference in provider
quality. And ongoing analysis of the Medicare data is essential to both quality
assessment and improvement.
According to our most recent study on health care quality, using risk
adjusted Medicare data, on average there is a 73 percent lower chance of dying
if you are admitted to the worst performing hospitals, as opposed to the best
performing hospitals, across 17 common procedures and diagnoses. If all
Medicare patients from 2008 to 2010 had been treated at top performing
hospitals, approximately a quarter of a million lives could potentially have
HealthGrades is an example, I believe, of a successful public private
partnership. We rely on access to a number of governmental data sets. Consumers
accessing provider related information at HealthGrades will find risk adjusted
performance data on all 5000 hospitals each year, for 30 common procedures and
diagnoses. Hospital measure for patients satisfaction based upon the HCAP data,
hospital patient safety information from 13 indicators of patient safety
developed by AHRQ. Pertinent information on 750,000 physicians, gathered from
hundreds of disparate sources, including data from all 50 states, and providers
themselves can access and maintain their data through our physician portal.
We do enormous research on again this traffic that the committee might
benefit at least from some of our learnings. And that is, 82 percent of
visitors coming to HealthGrades are seeking information on physicians.
Seventy-two percent of those are looking for a new doctor, 74 percent of the
visitors will consider two or more physicians. Once a physician selection has
been made, we know consumers act on that information. Fifty-four percent of
visitors coming to HealthGrades will schedule an appointment with a doctor.
Ninety-five percent of those scheduling an appointment will do so within a
week, and 38 percent of those scheduling an appointment are doing so that very
There is a tremendous amount of consumer demand for any data that could
provide comparative, meaningful information. Many organizations are gradually
filling the demand where once there was a void. And if you look at just an
overall macro trend, it terms of the rate of growth, in terms of internet
usage, health care information is growing at four times the rate as the overall
internet as a whole. So you can see again there is a tremendous need, out
certainly in the marketplace.
We have partnered with the Centers for Medicaid and Medicare Services, and
we enjoy that relationship. And we hope it continues to play a very large role
in the process, through the development and implementations of standardized
data platforms and delivery mechanisms that can make the data more accessible.
I applaud several recent efforts on this front, including the President’s
Open Government Data initiatives and the roles HHS and CMS have played to make
the data more accessible, including the creation of HealthData.gov, the annual
health data initiative conference that we have participated in every year, the
data user roundtable held in December that allow organizations to share their
challenges accessing and using Medicare data and providing a forum to offer
recommendations on how to improve.
As CMS explores new ways to collect, package and make data even more
accessible, I would like to offer some observations and thoughts briefly on
what might be most useful to stakeholders of this process. The data use
agreement is now indicated is in place for some data sets, such as the
Healthcare Cost and Utilization Project, the HCUP family of data. But that data
use agreement restricts access to only those organizations considered to be
conducting quote pure research, which tend to be institutions of higher
education and not for profit groups.
If other organizations were allowed to access the HCUP data, such as for
profits who have the ability to invest resources, exploring ways to make the
data more useful to the general public, this data would indeed be very useful
to all consumers. It could, for example, provide confidence of performance
information on ambulatory surgery centers, which currently does not exist in
We also believe that data should not be limited to purely academic
endeavors. Reviewing the current data use policies is one way obviously to
address this. Lack of physician’s specific information, even at a basic level,
meaning physician identifiers, procedures performed by and et cetera, make it
difficult to provide consumers with any meaningful, clinical relevant
information on physicians. The Affordable Care Act addresses some of these
issues through data collection and reporting requirements, but we are still a
few years away from being able to access much of this information. Indeed, what
Niall indicated, the 10332 statute certainly lifts that veil somewhat. We would
like to see the kind of broader application of the 10332 policy, again, making
that data more available to more organizations.
In conclusion, CMS implementation of a specialized unit, overseeing data
supply, demand and governance will improve on the great work that has already
been done. The application of market principles, meaning a willing buyer and a
willing seller, coupled with cost and quality, transparency, will inevitably
lead to an efficient marketplace. I would argue, as Niall did, as well, we are
still years away from that. But I think we get closer to it each and every day.
The direction announced by CMS today is aligned with HealthGrades’ guiding
principles of transparency, provider accountability and consumer empowerment.
Health care quality can be used to improve quality and contain costs through
effective knowledge transfer, based upon greater data availability. More
efficient data sharing mechanisms, public private partnerships, will engage to
empower consumers thorough more easily accessed and meaningful information.
The last decade has seen a tremendous evolution in terms of data
availabilities. It is up to us to determine how to make best use of this data.
The next decade, I believe, we will see exponential increases in uses of the
data that couldn’t be envisioned five years ago. And CMS and HHS are
strategically applying resources to plan for the growth in this area, and
indeed fostering innovation. I applaud their efforts, this committee’s efforts
and the direction that is being taken. Thank you very much for the opportunity
to share my observations and thoughts.
DR. CARR: I have a question. On the Thomson Reuters and HealthGrades, have
you ever looked at whether you come out with the same assessments of the
providers that you look at?
MR. HICKS: We look at the correlation, and there is not tremendous
correlation because they measure different things than what we do. They will
take into account, for instance, perception in the marketplace with respect to
quality, where we look at either mortality or a major complication, and only
one endpoint in those 30 most high volume procedures and diagnoses.
DR. CARR: This presentation is very timely based on a hearing we had two
days ago, three days ago, on measures that matter to consumers, and moved
really from the long-term what is that magical measure to what do we have
today. I am interested in knowing how you use consumers in developing what you
put on the website.
MR. HICKS: That is a great question. So when we get a million visits a day,
we can essentially survey any part of those groups in either an AB test or a
strict just in a survey. And we will come up with an answer pretty quickly. I
can assure you, having done this for about 14 years, whatever I would predict
would be 100 percent inaccurate or wrong. But how we ultimately develop, what
are those indicators.
We know, for instance, that 99 percent of the consumers coming to look at
hospital quality will stop at the star ratings. We have a methodology, we have
all the detail, but it is akin to Morningstar and the financial services. Once
you get the answer, those of us around this table that are interested in how
you got to the answer, 99 percent of the people kind of don’t care. They
ultimately want what is the conclusion. They will take outcomes, generally over
And you think about again a lot of us assume or presume probably enormous
health care literacy in the marketplace, which frankly doesn’t occur. And
meaning time to read perfusion, find acute myocardial infraction. If you had a
billion dollar budget and an ad campaign, you could probably move the needle
there, but not very far. I would argue, so I think that is one.
And I think the other variable is most consumers actively avoid or ignore
provider information, up until the time that they need it. So their level of
interest is very shallow, and then all of a sudden, it becomes enormous. So
that timing of when you deliver information is a critical component, in terms
of what we do and ultimately how we measure. But a direct answer to your
question is either through survey or through A&B testing.
DR. CARR: Any other clarifying questions or should we continue on? Next.
DR. KELLY: It is really a delight to be here this afternoon. My name is
Brian Kelly, I run informatics at Aetna. What I am going to really do is just
make a couple of brief points. First, I am astutely delighted about what CMS is
doing here. My talk really is entitled why data sharing is so important to
making quality healthcare more affordable and accessible. And I know that data
sharing is very important because at Aetna, I actually oversee our data
governing. And not a week goes by that I don’t get one or two requests for our
data from some external, academic organization, pharmaceutical company, ACO or
the list goes on. So every week, we are looking at who should we supply data
to, why is it important, and I can tell you it is very, very important.
I am an absolute believer that sharing data is critical to how we transform
health care. And that being able to add the 50 plus million members of data
that resides in the CMS data sets will just hugely move the needle. This is how
we are going to enable big data, which is going to transform health care.
But to do this, we really need a new paradigm around data sharing. And many
of the aspects of that new paradigm will require public policies and a lot of
consumer education. If consumers really knew how we could really help solve the
affordability and quality of health care across the United States, they would,
I believe, have a different opinion on how data sharing could be done.
Obviously, we have to always do data sharing in a way that respects privacy, an
individual’s right for privacy. But I believe, and I am going to show you a
couple of examples of how we can do that today.
One minute on Aetna, and I only do this because I think it is really
important to understand that we are the commercial version of what Medicare is
becoming. We are a health plan. We serve about 18 and a half million members. I
have 73 million different people in my integrated database at Aetna, and that
has huge value, and I can transform health care, and do really neat things for
those members because of that data.
You can see the numbers. What is interesting on the bottom is we have 10
million members that have a personal health record, many that use that personal
health record. And that personal health record is automatically supplied with
all of the data we get on a member. Their claims data is there, lab data, their
pharmacy data, their HRA data, and now more and more, their biometric data. And
it is powered by a clinical decision support system that tells them what their
preventive services have been done or have not been done, and what their gaps
in care, based on very sophisticated things. We are using data all the time.
We interact with over a million health care professionals every year. And it
is this data sharing between this member and this doctor, and these systems of
care that is so important. The other thing we do, and I didn’t realize this
before I got to Aetna a few years ago, is we have 3,000 nurses and over 100
doctors at Aetna. And those 3000 nurses touch those 18 million members every
week. And what they are doing is they are using this data to actually solve
gaps in care and make health care better.
The last thing is that we are transforming in many ways into a health
information technology company. We have invested over $2 billion in assets in
this over the last few years. We have acquired Medicity, the US largest Health
Information Insurance Exchange. As I mentioned before, on this integrated data
set, we use a very sophisticated set of clinical decision support, what we call
CareEngine, to identify gaps in care, and help route those gaps in care to both
doctors and members. We have one of the largest data warehouses, and we have
more health IT people focused on health care than Microsoft, Google and Oracle
combined. We are using this data and we are making this happen every day.
If you go to the next slide, I just want to talk a little bit about how fast
the world is changing. Two years ago, we did not have any accountable care
organizations. We now have 10 under contract, we have over 20 letters of
intent, and we have a pipeline of 50 other hospitals that we are in serious
And these systems of care, which I believe will transform health care, will
align incentives and the member and the doctor and the hospitals and the health
plan, to all do the right thing and improve health care. They require massive
data sharing. And it is not just the claims, the labs, the pharmacy, the HRA,
the biometrics that we currently have. It becomes more and more granular data.
It becomes that clinical data that is in the DMR. And more and more, it is
going to require access to genetic information, and the lab tests around
genetic tests to really make health care better.
When I went to medical school 25 years ago, I remember there was really one
kind of breast cancer. Basically, you knew the person had breast cancer, you
did the staging, you figured out where it had extended. Now, based on receptor
testing, there are many, many different types of breast cancer, and they are
treated very differently with different drugs. And if you don’t get it right,
you are harming the patient, you are exposing them to drugs that may not work
in them, you are delaying drugs that might work in them, and you are driving
huge inconvenience, side effects and cost to the health care system. So the
sharing of data, and this whole new taxonomy, is really, really important, this
whole concept of information common is critical.
What are we doing at Aetna to try to help this? Our basic philosophy is that
we are data hounds. We try to get every piece of data that we have, and
essentially what we do is we try to tie that together. So for these 10 ACOs
that were already under contract, and the 20 other letters of intent, what we
are doing is we are supplying our technology infrastructure, this $2 billion of
investment, and now it is changing that information. We are sharing that
information with the member, through a personal health record, we are doing
very sophisticated decision support on it, that we are routing directly to the
physician, caring for that population in these ACOs. So they know if I have
1500 members empaneled to me, they know exactly where they are on those
critical performance methods that are actually going to drive long-term
We then use that same data to drive our care management nurses in what they
do and what the hospital care management nurses do. We are, in many instances,
embedding our own nurses in physician practices, who have access to this common
set of data, who can really help coordinate that care and make a difference.
That allows us to do population management, and that allows us to do risk
I want talk just for a minute on how this data can really transform clinical
research. For the last two years, we have been one of the largest data
providers to the FDA’s Mini-Sentinel Project. Some of you may know what that
is, but essentially it is a drug safety program. And the way Mini-Sentinel
works, and it is a wonderful paradigm for how the government and the private
sector can work together, and do things in a very inexpensive way that adds
huge value, and protects patient privacy.
So what happens in the Mini-Sentinel Project is smart folks at the FDA will
be looking at a new drug that was just released. And they will say I wonder if
this new drug causes a heart attack or a stroke or liver failure or any
condition. What they will do is they work with us, and we set up our
infrastructure, so all they do is they send us a query that they have already
programmed. And the query says Aetna, will you please look at the ICD-9 codes
or the diagnosis codes for this set of conditions? And would you look for these
side effects, do you see these conditions happening for a person on this drug?
In 24 hours, we can take that query, run it across our 18 and a half million
members, determine every member in that 18 and a half million that is on that
drug, and see if we see any evidence of the side effect they are looking for.
It costs us almost nothing, very easy to do. Within 24 hours, all we send back
to FDA is a list of saying we found 2000 patients on the drug you are
interested in, and here are the incidences of side effects you’re using. A
wonderful paradigm, we can do a lot more.
I really think that this opportunity of what CMS is doing here is absolutely
fantastic. My recommendations, first one is clone Todd Park, get more of him
out there. I think that would actually work well. I do think that looking at
all of the Medicare policies, because one thing I will mention is that when we
run most of our queries for research right now, we are limiting it to our
commercial data set. And the reason is that our lawyers tell us that our data
use agreements, our Medicare Advantage programs, do not allow us to do the
I assure you that you should hold us to the same high standard you hold
yourselves to in how we use that data. But I think looking at that type of
sharing would absolutely graze that. If you think about it, I am most
restricted in my ability to impact the over 65 members of our country through
partnerships with pharmaceutical companies and other research entities, because
that is the one data set I don’t have access to.
Please do anything you can to educate consumers and policymakers on why it
is important for data sharing. Over the years, I have seen a couple of
wonderful children’s cancer centers and children’s hospitals have brochures
that tell people that when they bring their children there with a rare disease
why it is so important for their children and their families to participate in
research. I think we can do this with data in a way that really does respect
privacy and advances the industry.
And then, combining public and private data facilitates comparative
effectiveness research is absolutely critical. I really do thank you for your
time. I am delighted with where this is going. Anything you can do to further
it would be fantastic. Thank you.
DR. CARR: Excellent, thank you. We will keep going then.
MR. GEIGER: I am Harley Geiger and I am policy counsel with the Center for
Democracy and Technology. We are not data users. We are a policy-focused
organization, although we have testified several times before NCVHS, other
federal agencies and Congress on health IT policy, particularly as it relates
to privacy and security. And today, I am here to talk about the underlying data
architecture of many secondary use programs, including the programs that are
supported by CMS data, but also looking beyond CMS. We have talked about this a
bit already, particularly through the comments of Dr. Cohen.
First of all, I want to be clear that CDT strongly supports secondary use
programs for all of the reasons that Todd mentioned earlier and that have been
discussed here today. They can enable broad-based improvements in health care
and they are critical to health care reform. But as HHS knows and has already
been discussed, the health care system and the very usefulness of secondary use
programs depends on patients providing full and accurate information to their
providers, which in turn depends on the public, maintaining trust in the
privacy and confidentiality of their data.
Now, the current trend for most secondary use programs is to pull all of the
data into a centralized location that is maintained by the government, where it
is analyzed and then released to data users. And unfortunately, the pattern
that we have seen is that there is a new NASA centralized database created for
almost each new data need. And regulations are locking plans into this model.
Examples include CMS’ own proposed risk adjustment program, state all-payor
claims databases, and the recent Office of Personnel Management health claims
data warehouse. Many of these, in fact, are analyzing the data for similar
Unfortunately, centralized architecture can exacerbate privacy and security
problems because it necessitates the creation of multiple copies of the
patients’ data that is then held in multiple locations. So this can increase
the risk and severity of data breaches, due to the volume of data in each
database, and the fact that there are copies in multiple databases. It erodes
public trust, as we have seen in fact with the risk adjustment program, and
with the OPM data warehouse. There is a lot of stirring of public opposition
against both of those programs.
It can also be expensive to maintain multiple large centralized databases.
It is inefficient and burdensome for plans to maintain multiple feeds with
multiple agencies, especially if those agencies are conducting similar
analyses, but require different data formats. In our experience also, data that
is held by the government often ends up being used for purposes other than the
purposes for which data was collected. But because the government already has a
copy of the data, the public has less opportunity to provide input into how
that data is being used in the future.
And lastly, it is a poor long-term strategy to continually board a new
massive centralized database for each data need because they are already so
many data needs out there. And that need is simply going to explode in the
future. So CDT hopes that HHS will continue to consider decentralized systems
as alternatives where appropriate, not necessarily as a replacement for a
centralized system in all circumstances.
And it is also important to note that privacy and security policies wrapping
the architecture, regardless of whether it is centralized or decentralized, are
very critical to maintaining the effectiveness of the program, although that is
not the focus of this presentation. In a decentralized system, government
researchers of course can keep the results of their analyses, but they never
obtain a raw copy of the patient data.
As I list here, a decentralized alternative can reduce the risk of data
breach by reducing the number of databases holding copies of the data, and also
reducing the volume of data in these databases. It is more in line with public
privacy expectations to keep data at its source. And a lot of plans have
massive proprietary concerns with sharing raw copies of their data with too
many entities. This minimizes data transfer, because you are really only
transferring the results, not raw copies of the data. And as was mentioned
earlier, it leverages a lot of existing systems, many of which are query-based.
There are two types of distributed systems that we recommend HHS continue to
seriously consider for secondary use programs. A distributed query system where
researchers and agencies write the code, they submit it to the data holders.
The data holders then analyze their own data, and the data holders submit the
results to the researchers or to the government agency. And this is sufficient,
we think, for many common research purposes.
However, it has been brought up several times before that because you are
allowing the plants to analyze their own data, it may not be appropriate for
those categories of uses that can confer if competitive advantage or
disadvantage to the plans. And so, for that category of secondary use program,
we recommend what I am going to discuss in the next slide.
A distributed access system, and a key difference from a distributed query
system is that the researcher or agency then accesses the underlying data,
while it is still held with the data source. And then, executes the code or the
query themselves, so the data is still behind the firewall of the data source.
But the data source itself is not analyzing the data and returning the results.
To accomplish this, the data source can set aside a copy of the data in a
secure environment that they themselves operate. The agencies can access this
via secure interface, which should be flexible enough to handle many queries,
and that can also support role-based access controls. And then, the agencies
and researchers keep the results of their analysis as opposed to copies of the
data. This can be also supported by auto controls to ensure accountability.
And the advantage here is, as I said, keeping the data at the source, using
many privacy, security and propriety concerns. But also, the health plan itself
does not have to run the queries, which can be burdensome on any small plans.
And also because the agencies themselves run the analysis, there is less chance
of fraud or inaccuracy for uses that can confer competitive advantage or
disadvantage. Although there are unique challenges associated with the
distributed access system, primarily as they relate to network reliability. A
weak node can hamper analysis across nodes. This can be perhaps mitigated by
having windows where the high availability ed server is open for analysis. And
these challenges have to be addressed before deploying any distributed system
on a broad population scale basis. So to be clear, we are not saying that this
system is ready for primetime, but it is something that we think ought to be
And so, that is our recommendation. One, please do not lock plans into a
centralized model in the future, the way that CMS’ proposed rule on risk
readjustment does. Instead, word your regulations to leave open the possibility
of decentralized systems in the future. And second, continue to test
decentralized systems. There are many that are promising right now, including
Mini-Sentinel and the multi-payor claims database. But these must be tested on
a population scale basis for the sorts of secondary use programs that we see
currently being conducted by all-payor claims databases, and CMS itself. And
that is it, thank you very much.
DR. CARR: Thank you very much. Next speaker?
MR. DAVENHALL: I am Bill Davenhall with a company whose name is Esri, but it
is spelled E-S-R-I. For many people in this room, you will probably recognize
perhaps not the name, but you will recognize the maps that are created with the
software that our company develops. So my first order of business is to tell
you that most of you are our customers. All 50 state health departments, the
CDC, HHS and every federal department is a user of this technology. And so,
what I am going to try to do today is explain very quickly what this analogy is
supposed to do for you.
First of all, this is a view of the National Health and Social Data
Ecosystem. As you can imagine, this technology is about linking data that is
geographically relevant. Now, what you see here are the things that the people
in the health and the social world think about. Where are things, what are the
hazards, what are the social resources, the health resources. And for any other
industry, if you wanted to call it an industry, like transportation or utility,
they build the same models. This technology is able to link all that data to
health models. And health models can be linked to transportation models and
utility models and water models and all sorts of models that you can imagine.
So it is a way to think about how your world is connected. It is really a
technology, it just doesn’t produce maps. It literally provides the framework
that allows you to think differently about what you are working on. So I bet
all of you, when Niall showed that beautiful map, you probably still in your
own mind can remember where the darkest spots were, can’t you? I know I did.
That is the power of that technology is cutting through all of the millions of
pieces of data that Niall had to sort through, to get to that point where you
have in your image. You are probably going to ask him later can you see that
map again, because you want to look at it more closely, to see whether you live
there or you have relatives that live there or something like that.
I brought prepared remarks, and I think they may have been available to you,
so I am not going to review those. But I do want to point out some things that
probably are important. My remarks are really coming from 40 years of working
with your data, both the National Center for Health Statistics and CMS data.
And I would say, as a graduate student who walked into his office and
cubbyhole, and found 25 shipping boxes found with the Rainbow series, I thought
I had died and gone to Heaven. And I carried those things around in a moving
van for like almost 30 years before my wife made me throw them away, because
she said aren’t they all on the internet? And I think probably most of them
are. So that is how I got introduced to the National Center for Health
Statistics. They taught me how you can creatively use color to remember things.
You all know what I am talking about, I suspect.
I am going to start where I wanted to end up, and then I am going to explain
this slide a little better. There is a great laden opportunity in the wider use
of CMS data assets, both for clinical administrative purposes that meet
national, as well as local health policy, and human and social service deliver
imperatives. The line is getting finer every day, between where does health
begin and where does social begin. And I would say it is time to sort of put
that to bed and say you know, it is all on the same continuum.
Any health or social provider serving a diverse population should have the
ability to compare what they are doing to what the most popular governmental
health and social programs are doing. Presently, most of us work in murkiness
because the most useful CMS data is neither easily accessible nor practically
useful at the lowest geographical levels. We call that finely grained, and I
would say accountability demands finely grained data, with the provision that
it must remain confidential.
So I believe that CMS holds an incredible key to launching a new generation
of health data analysts in America, analysts that will be equipped in the right
invest data, at the right geographical level and when they need it, creating an
enhanced capacity to serve up data and encourage its intelligence use. So Niall
is right on about data access and use. They are two very different things. Data
accessibility is one thing and usability is quite different, but they have to
be linked together. One without the other is not where we need to go.
Finally, I believe that CMS represents our best hope of helping the entire
health and social ecosystem understand how our health system actually works, or
for those of certain vents, doesn’t work, for individual consumers,
communities, providers and policy makers. So on this slide, what I want you
realize is that this little bit of data that we are talking about here today,
we know that there is CMS data and then there’s 100 other kinds of data. It is
all trying to serve as many people across many places for many different
purposes. So utility becomes a big factor, leveraging the limited data
resources we have becomes a critical economic driver.
When you take a look at the CMS data users view, which is where I have been
for many, many years, I was answering these kinds of questions. And I would say
probably most of my years were spent in the research community, so I was always
asking who, what, where and why, and how much. And this is the technology now
that I would say everybody in government is learning to leverage this
technology, so they can cut through the haze and get right to what is most
How do you make this data actionable? How do you get it into that position,
so Niall could show you that picture of the United States? And he does that
with a sense of not just urgency, but a sense that it is correct and right, the
underlying data is accurate. I have this slide because we have a big data dam
that is going on in this country. And it has to do, if you don’t get the data
right when you start, it doesn’t get any better. You just clean it, but it
doesn’t actually get any better. And so a directly claim coming from a doctor’s
office or hospital is just passing it onto the Aetna’s, and the Aetna’s have to
clean it. And they are going to pass it onto National Center and they may have
to clean it, and everybody is cleaning the same data, when it should have been
corrected at the first get-go, at the point of service.
This slide is just to tell you that some organizations are getting their
data sets what I call GIS ready. Getting that piece of accuracy right, and they
are going to go through the dam a lot quicker, and they are not going to be on
the rocks. I have a lot of boats there that are in troubled water. You don’t
want to be an agency who has a lot of data in troubled water, because it just
means that the data is accessible, but is not very useable.
This is my recommendation, and I really am just echoing Niall. They have
already said they are going to do this, so this is really great. It is to
create this data access and use program, add value to data before you serve it
up, and assist users in consuming it intelligently. So there is an incumbent
responsibility, I think, that both CMS and people like National Center for
Health Statistics do, in explaining what we have. And I know the National
Center has done a lot of that over the years. So it shouldn’t be a big stretch
for CMS to also get into that same business, where they help us as end users
use that data intelligently.
This is an example. This is a recent piece of work that was done by Loma
Linda University Medical Center. For those that don’t know, that is in Southern
California. It is where they did the first pediatric heart transplant, Dr.
Bailey, and also the first proton accelerator. They are trying to step up to
the Accountable Care Act, and this is a picture of the emergency department
utilization, where they decided that the diagnostic information suggested they
did not need to be in the emergency department. And this represents loss
charges to that hospital, over $30 million worth of loss charges. And wherever
you can see the brightest darkest color, that is where most of the loss charges
The granularity of this is that census tract in block-group levels, and they
know a lot about the people they serve. They almost know nothing about the
people they don’t serve. They could know a whole lot more if CMS takes a route,
that begins to share that data at a more finely grained manner, so they have
something to work with. So it is just one illustration of the process of moving
from what we have been working with for many years, which is highly aggregated
data, to something less aggregated, deaggregated, but yet still remains and
meets all of this confidentiality provisions. If there are any questions, thank
you very much.
DR. CARR: Great maps and great illustrations. But let’s go on to our next
speaker and continue on.
DR. ROSENTHAL: I am Joshua Rosenthal, co-founder of the startup RowdMap. A
couple of months ago, I was at MIT, giving a presentation to a roomful, a
couple of hundred people, MIT students, doctors. A couple of months ago, I am
at MIT giving a presentation to a couple of hundred of groups of people. And
they want to build the next best health care company, particularly around data
and analytics. And MIT students, there are doctors, there are designers, there
are Pure Play Tech guys there. And a couple of weeks after that, I am at
Harvard giving a lecture to essentially the same sorts of folks. These are
undergrad students, in computer science and business and in health care. And
these two groups of people have the same question, where is the action, where
is the opportunity, where is the most meaningful stuff happening in the health
care system right now. Is it New York or Silicon Valley essentially? And I said
you are not going to believe this, you need to sit down, but it is actually
happening in D.C. And it is not nanotechnology, it is not biotechnology, it is
actually transforming the payor system from fee for service to pay for
performance. And they went nuts and got into all sorts of fistfights. No, they
didn’t do that.
But the point of the story is that, that is actually what is going on. You
guys have done fantastic work in moving from fee for service, which is I get
paid more to do more stuff with unwarranted variation, everything that comes
along with that. Part of our background was around the Dartmouth atlas, I speak
with some authority around that, to pay for performance.
I get paid more if I do better things, better clinical outcomes that are
patient-experienced. That is fantastic, and it also means you are transforming
the big fish in the eco system, the insurance companies from risk brokering
adjustors to participants in the health care system. I don’t just mean through
ACO, I mean through their identity, the M&A activity. If you scan the
newspaper, you have seen in the past six months, half a dozen billion dollar
cash plus deals where they are requiring not just M&A populations, but
providers and means to control those providers.
And that is really important because you, and I assume you have done this
intentionally, have changed what it means to be a profitable life. You are no
longer in the DM world where profitable life is someone who is at the low end
of the critical risk strat. I think I am going to be healthy, so I am
profitable. A profitable life is now someone in Medicare in an M&A program,
where I am controlled by pay per performance. So you have aligned incentives
It does mean that you have some things to think about very carefully. And as
the next slide loads up, you will actually see what I am talking about. So you
are hearing all sorts of problems, and it’s security, scalability, man hours,
custom files, et cetera, et cetera, et cetera. Those are all symptomatic
problems. Your key problem is that you have a systematic business change. You
have done that intentionally, it is fantastic. But it breaks your historic
infrastructure. So it costing you money to do all of these things and you have
these issues. But you are telling providers, plans and everyone that they need
to be data driven. And if that is the case, you need to make that data
accessible. And when I say data, I mean information, interpreted data in a
meaningful way. And that is system wide CMS data. Today, it is research and
quality control, which is different than STAR and reimbursement and provider
payment, and tomorrow Medicaid rolling out and other things.
These are two sorts of separate issues that you have to deal with. One is
research quality control. Here, you have a small number of large data sets that
is identifiable reidentification concerns. Your concerns are security, privacy
and automation, and internal scalability. And you also have just as important,
and geographically cross walkable by contract, this is the fine grain stuff we
are hearing about. STAR and reimbursement, here you have a large number of
small files, which are public. But your issues are coherency, usability and
access, not just access, but access in terms of interpretive access.
And so, my suggestion would be, and I guess you are already doing this, so I
will just echo that this is a fantastic idea. I couldn’t have conceived to do
anything better. Solve today’s problems with the technical foundation for
tomorrow’s business paradigm. And that is some sort of system, we can talk
about distributive access versus query versus centralized, but there is another
way to look at it. A classic approach is a data extract system and tool. This
is simple, it works, it is proven, reliable, and you have heard a couple of
examples about that. It is centralized or has an abstraction layer with
distribution. But the point is, I can pull together meaningful information on
the fly with the UI tool creates it. And this works, everyone does it, solves
security, privacy, automation, and uses something like this.
And if you pull into the non-beneficiary STAR stuff into the data structure
in taxonomy, you can actually link all the stuff where plans make money, not
just the fine grain piece. You can actually link the profit drivers to it,
because you created that around your STAR metrics geographically. You can use
something new, slightly different approach to that is what you call a cloud. It
makes you seem cool, but there are also some technical benefits. I won’t talk
about that, you can ask the real experts about that.
All of these systems just have a few things in common, and it is worth just
thinking about this a little bit, because the devil is in the details here.
There is some sort of system or tool, where users can create things on the fly
with an extract, with a GUI, or have a permission that is credentialed, and I
have a certain level of access and an enclave. And it allows me to pick the
grain, what I want to do, et cetera, et cetera. And you want to be smart about
it, so you’re not reprocessing everything, so you can be smart. You can
basically do new processes, at APCD level, three types, new, update, cancel.
There are different ways of doing it.
But when you build this, you can actually implement quality control metrics,
and this would be incredibly helpful with your data right now. Basic check some
type stuff, as well as universal constants, meaning there should never be more
than 500 contracts here that have this stuff, that will filter out a lot of the
slop that you find.
And most importantly is taxonomy, is their hierarchies. The stuff can’t be
completely flat. When Niall says interpretive meaning, what he means is a
taxonomy. And when I say taxonomy, I am not talking about conceptual diagram or
a diagram or a meta data model. I mean baking in your business questions, into
the structure of the data. And Niall showed you a couple of examples of that,
and that is really, really important. And when I say business question, I don’t
mean plan-making profit. I mean things that you guys think are important and
you incentivize correspondingly, improved clinical outcome, improved patient
experience, et cetera.
There are a bunch of slides and they are showing you very specific examples
of how we have done that on our own, because we had to, because it wasn’t done
for us. And you can flip down and look at it later. And then, some sort of
learning center, some sort of distribution, where I can actually share and
learn in internet time with other folks. Not just share code and not just have
documentation, what does this thing mean?
We are not to spot errors. In 2009, a contract meant something, in 2010, it
means something else in your taxonomy. But actually share interpretive
frameworks. And here, on one hand, you need greater granularity as Bill talked
about. On the other hand, you have this sort of meta interpretation.
I want to talk about what it means, when I pull these metrics and I see
Aetna versus Cigna. Having very sorts of different clinical rates by your
performance metrics, in areas with very specific races, versus low access
providers, high access providers and the Dartmouth stuff we pulled in.
And what does it mean when they actually implement HealthGrades, and it has
very different effects on very different groups of people. And I can do all of
that stuff right now with public data that you already had, at least on the
Medicare population. And so, having the hooks to crosswalk that back, is
incredibly important. And you will see an example of a learning center around
that. I would say having someone outside the health care space is fantastic.
Also, extending that data into controlled views, data explorers. In the Pure
Play tech world with some of these MIT kids, the things that are winning the
awards are actually them looking at health care data with Google, with
ReadWriteWeb, with TABLO, things that you guys aren’t intentionally doing, but
they are using your data. And there are tens of thousands of people using this.
So building a good consumer-grade portal to actually get into, and there are
some slides showing what that would look like, and pushing the data, at least
the interpretive data, out into the ecosystem like little seeds and pollen.
That is what has worked so well in the Pure Play techs. I would encourage you
to take it a step further. It’s like YouTube for data, you can play around with
it. That is the easiest way to think about it. And finally, same system for
public data, okay, fine.
So in a nutshell, you have databases, this is the beneficiary stuff, small
number of large databases. You have files, large number of small files, which
have this interpretive difficulty. And you need to put into some sort of
structure for some sort of meaning, for standard taxonomy, not meta data, but
around business. You need a learning center where I can go and figure out what
the heck is going on. You need some sort of simple system where I can get
access to it.
And then, if you are innovate individual, I have got a number of start-ups
which have been successful, not as funded, but actually acquired by plans, by
technology companies and by other folks, as well as private equity companies. I
am speaking with at least what has worked for us. And we can do that because we
have built these systems that Aetna and company uses, and around 85 percent of
your commercial Medicare population through it.
But for the Harvard and MIT kids, they can’t do that, that’s really tough.
They need to build up there where it says this, asking him to build all the way
down into the data is just incredibly difficult. And my background is PhD
Fulbright in Sorbonne’s Applied School for Advanced Studies, and I assure you
these kids are far, far smarter than me. So it is not an intelligence issue or
a different species. But if you want to explode innovation, you need to do
that, to make that applicable for them.
And if you do that, you have already done the hard work. You have already
created the standardized way of looking at stuff. And I assume you have done
this intentionally, because it is just brilliant, there is no other way to do
it. The standardized way of looking at stuff in your population, that ties
payors to providers, which is why you see all of the M&A activity.
So I’m a payor, I want to make a data-driven decision. How am I doing
compared to my peers, because that is really important. Where should I focus,
what should I do, which interventions, intervention being a product like
HealthGrades. I am a data-driven intervention vendor. How am I doing compared
to my peers, you asked a peer question about HealthGrades. Which members does
my intervention work for most, meaning when I deploy this thing, does it
improve satisfaction these metrics?
And these are out of all the health people we know, and stuff you guys see
thousands of these great ideas coming out. When working with these kids at MIT
and Harvard, 99 percent of them fail. And if you look through very carefully in
the Health 2.0 stuff, you see 99 percent of those companies fail within a few
months. Why is that? Because they haven’t tied it to a meaningful business
question, right? They haven’t tied it to profit or profitability, and then they
put a 299 iPad app and no one really uses it, and et cetera, et cetera.
You guys have built the infrastructure to allow them to do that. They can
take that silly iPad out that says how fat I am today, and actually say when
you deploy this with Aetna, you actually improve in these areas, these specific
metrics, and this is worth X amount of dollars in reimbursement, this is worth
Y amount of dollars in Medicare retention, et cetera. And then, you can do
meaningful research. What do successful payors look like in terms of their
impact, geographically, access, et cetera? What does successful interventions
look like? What is the nexus between the two, what are the characteristics of
key drivers to improvement, and how can policy accelerate these key drivers.
You have done this hard work, it’s jaw-dropping. After our last successful
ideas(?) that we are going to get out of health care, but you guys kind of
shamed us into staying in it, because if you are doing this work, we can’t go
walking away now, right? So you have done that. All you need to do is put a
couple of more chips on the table, to really capitalize on this. That is why
it’s exception when you say you are going to double down, that is beyond our
You need to do two things. You basically need to build that infrastructure
data information products and services, because you are asking everyone else to
be data-driven, so you need to give them the ability to do that meaningful,
that changes your infrastructure. You can’t just give them a transaction list,
like you go to a menu and say order this and it costs this much. If you are
going to be pay for performance, then you need to have some interpretive
framework on that. And you need to make it meaningful, which means taxonomy,
and please look at the examples of what I am talking about. And you need to get
people to use it, which means not only getting people to use it with kind of
fairs and challenges, but getting a decent looking portal, and helping them tie
their wonderful ideas to the specific drivers of profit in health care. Because
when the Pure Play tech people come in from MIT and elsewhere, they don’t
understand health care is littered with perverse incentives, and that is why
they fail. You need to help educate them through whatever you are building.
Thank you for your time and attention, appreciate it.
DR. CARR: Amazing tour through some amazing minds. It really opens our minds
to hear this.
DR. W. SCANLON: I can’t replicate Todd’s demonstration of enthusiasm. But
let me tell you how enthusiastic I am about this and how positive it is. I
started in health services research about 40 years ago, and worked very closely
with HCFA. And I don’t think you can imagine the data that was available then
to guide policy. It was some combination of things on paper that were analyzed
with colored pencils, and some computer output that took gosh knows how long
sort of to create. And so, the fact that we are here today, talking about the
potential that you are actualizing is very enthusiastic. Some might say well,
it is 2012, so shouldn’t we be here sort of anyway?
But that has been true for a long time. We should have been somewhere
further along and we weren’t, and the fact that we may be close or at the point
now is incredibly positive. And I say all that, even though I spent these last
40 years having a much better experience on average than most people. I worked
on government contracts that I had access to HCFA data. I worked at GEO and we
had access to HCFA and CMS data. We weren’t CMS employees, but we might as well
have been. We were basically sort of in their face all the time and using all
their data, so we could do anything that they were able to do.
At the same time, what we were doing was hard, and it wasn’t sort of capable
of meeting the needs of the time. I can remember in this last sort of 10 years,
we faced the issue of the SGR with respect to physicians, and the big question
has been are physicians dropping out of Medicare. And we always asked ourselves
can we answer this on a timely basis. Because if you appear before the
Congress, and you tell them three years ago, this is what happened, they say we
didn’t come here for a history lesson.
And so, we were saying is there anywhere in the CMS data flows that we can
tap in, and try and get some measures of physician participation, and it
virtually turned out to be impossible. So the notion that you are bringing data
sort of closer to real time, very close to real time, is such a positive thing,
sort of from a policy perspective, that I am incredibly enthusiastic.
It is particularly important, because I think just at the time we are at
now, we have got a compelling need to improve our health care system. And I
think that was brought home, or at least brought home for me very clearly, sort
of over this last year, as we were discussing the deficit and we recognized the
role of health care sort of in the deficit. And there is this feeling that we
have to do something, and the risk is that we don’t do something that is
well-informed. And that is a risk of great magnitude, because we are talking
about affecting tens of thousands, if not millions of people sort of
So being well-informed about the change is critical because I don’t think of
it as just trying to cut costs. I think of it as trying to improve efficiency.
This is the economist sort of in me, which says we may reduce the cost, but we
also want to do that in a way that doesn’t harm access and doesn’t sort of harm
quality. In fact, maybe in the process, we can improve sort of on both, and
that really should be our objective. But that involves a level of
sophistication that we haven’t had the capacity for before, because we were
really lacking sort of the data.
You have covered all of these areas in your presentation and the CMS
presentation, as well as sort of in the others, I think at three levels at need
for information. One is at its most aggregate, which is a policy and program
level, sort of for both policy development to sort of in program management.
And then, I think as a second level, there is the clinical sort of side of
things. And the ACOs were a great example because the notion of managing in an
ACO without information, it’s not a health plan. The trade-off here in terms of
trying to keep people sort of interested and attracted to them is freedom of
choice. And that means that the people in the ACO do not have sort of
sufficient information, so we have to think about how do we compensate for
that. And so, I think there is critical sort of information there.
And the third thing, which is the area which I think where we have to work
the hardest, and this committee has actually worked on, in terms of its
meaningful measures efforts, is how do we influence the patient in a positive
way, in terms of giving them the information that they are going to find
useful. And be able to sort of affect their behavior, sort of in a positive
At the top level, this aggregate level for program and policy management,
again sort of the fact that you brought things into real time, you are
realizing sort of much more about sort of all the data that is within CMS, this
notion that it is not just claims, but it is cost reports, it is assessments,
it is other sources of information. And the fact that those can be merged and
matched, and there is real power in doing that, in trying to understand things
better. That is an incredible positive thing.
I think we need to consider about how do we go further in that. In Bill
Davenhall’s presentation, his idea that everything influences everything else,
I think is very sort of relevant. The slide about the readmissions, those are
critical. It’s not just the program and the initiative that we have, but we are
also talking about penalties we are not talking about. And we are going to put
penalties in place for hospitals. One of the issues which is an hypothesis that
I have heard entertained multiple times is, what does socio-economic status
have to do with readmissions. What are we doing in terms of the safety net in
this country, if we don’t take that into account when we think about which sort
We are living off and most Medicare policies are modeled on the 1983 DRG-PPS
system. It has worked, but as we start to think about sort of being more
efficient, we may need to think about how do we refine. This is a challenge
because now we are talking about not necessarily that you do have available.
And we have to think about sort of how do we access sort of that additional
information. This committee had a hearing before the beginning of sort of the
health reform discussion, about information for health reform.
And one of the things that was discussed at this hearing was how difficult
data access is within the federal government. That going across agencies is not
a no-brainer, and each time there was an agreement to do something, there was a
start from step one. And it would take sort of a significant amount of time,
sort of to get to a point where there could be the kind of sharing that was
going to be useful. Moving forward on that is also important, and hopefully you
can maybe get OMB to be one of your partners on this, to sort of push that sort
As we release data, to deal with these more aggregate issues, I think it is
important to think about what can you do to be of technical assistance to the
users. I sort of both recognize when I was a user, and hopefully we were doing
responsible things. But sometimes, sort of after a lot of effort, we would
discover oh, my goodness, we are going in the wrong direction. We didn’t
understand that, we didn’t have that sense of what was really sort of within
this information. And the results you get comport with your biases, then you’re
assuming the analysis is right. You have to be sort of cautious about that.
If you can think about it as a point of your line of business, not just the
information being conveyed, but what kind of assistance you can give, that
would be sort of, I think, very positive, as well as maybe thinking about sort
of how you convey information and the types of information you convey, as being
sort of more failsafe than less.
I am not going to talk about what you are doing with respect to the ACOs on
the clinical level, because I think that seemingly you are doing exactly the
right thing, in terms that you have got to provide the support for these new
types of delivery innovations, when they are not natural to the organizations
that are involved. Again, we are not talking about a health plan, which may
have the information. We are talking about sort of independent provider
I think one of our big challenges, and there is a lot of interest in this,
is how do we information that is useful to patients. And Justine brought up
this issue of consistency of information. I have seen presentations about the
group insurance commission in Massachusetts and their physician rating sort of
efforts. And the fact that a single database was given to a set of payors, and
they rated physicians, and all came back with different answers about the same
physician. And that may all be legitimate, but it is the transparency, I think,
is a minimum that we have to ask for. Sort of why did you get this result
versus someone else’s result. If you are having a line of business, you need to
think about one of the prices you may be asking sort of users for is
transparency. That we know sort of how the data are being used, so that people
will be able to evaluate it.
Another thing I think we should be thinking about is, and you already have,
sort of if they have data, you would like to have it, too. And I think this is
a very big thing because again, sort of going back to the whole world is
interconnected, we shouldn’t be thinking about Medicare or Medicaid in
isolation. We should be thinking about them sort of in the markets that they
work in, and sort of what is happening sort of in those markets. For that, I
would salute Aetna. You didn’t mention this, but Aetna, along with three other
large insurers, has contributed to this health care costs institute, giving
them the claims data, and Kaiser Permanente is sort of one of the others, and
United and Humana are the other two. One of our problems in policy has been we
have had no clue what is happening sort of in the private sector. We have known
a lot about Medicare, but not enough. But we have had no sense of sort of what
is happening sort of on the private side.
So this notion that we really need to share information, so that we can get
a sort of more detailed and sort of refined picture of different sort of
markets, because being an economist, I have to say I think markets matter. And
so, we want to know what happens sort of locally, it is absolutely critical. It
raises some questions about these models in terms of not sharing the data
itself or having it centralized. We are going to have to have some data to
migrate, in order to be able to pool sort of information.
The other thing that has been a handicap in the past, in terms of letting
sort of others work the data, in terms of understanding the context for it, is
that when one has data from one source with individuals, and you want to match
information about their environment, that can be as revealing as their
identification code. And so, if becomes a bit difficult to say well, I need to
study this in context, but I can’t tell you anything about the context. And we
have to think about sort of strategies to overcome that.
And I guess the last thing I would say is I am as enthusiastic as Todd. I
haven’t sort of probably conveyed as expressively as would, but this is an
incredible movement forward.
DR. CARR: So let me open it up to other members of the committee. Do we have
someone on the line?
DR. TANG: This is for Todd. And one, I share Todd’s enthusiasm about what
this represents as a means to transforming the health systems. And I think to
But I had a couple of questions. One has to do with maintenance. So we
mentioned that in the state summit, there were a lot of databases, and not
everybody knew what databases were. Potentially even the owner of the database
didn’t know what the fully existed. So one question has to do with maintenance,
and that certainly happens in our organization where we have these reports, and
some of it doesn’t seem to be around anymore.
The other question I have is related to that. In this phase of the health
data initiative, I think its active data that the government currently has. Is
there a thought of moving to the next stage, almost like meaningful use, and
having shared consideration of what data could be collected, that they should
contribute to the value. So for example, in the old days, a doctor-finding
website might say, where did you graduate, where did you get your training and
are you board certified? Nowadays, it includes a video of your views on life
and health, et cetera. And it’s just that the data collected and are meaningful
to the consumer of that data, change over time. Would the government be open to
considering adding new fields that may increase the value for the consumers of
MR. PARK: Terrific, and my colleagues should actually also weigh in, as
well. So Paul, the first question was about maintenance, is that right?
DR. TANG: Correct. So let’s say for contractual reasons, you require your
contractors to supply certain data. Does it get maintained every year, more
often? So for example, if you want to use, let’s say, data on social services,
the updates to what services are available and what tests that certainly could
change over time, are these databases specifically maintained.
MR. PARK: With respect to maintenance, I think it is useful to think of the
HHS data universe as a highly diverse eco system. I think that the routines for
maintaining and updating data are as varied as types of data themselves. And
so, if you think about what you mean by data, it is everything from the latest
and greatest scientific knowledge at the National Institutes of Health,
Medicare claims data, FDA recall data, community health performance data, et
cetera. So it is really quite different from place to place to place.
What I think that we are actually seeing that is enormously beneficial is
think of the different data sets within HHS each as data sets are owned by a
particular business owner, whether it be someone at Medicare or someone at NCHS
or someone at National Medicine(?) Center. What we are actually seeing is that
it is incredibly valuable to have those data owners interact with people in the
outside world who actually use the data. Because in some cases, data updated
every five years could be perfectly fine. In other cases, it is completely
useless. In other cases, it is something in between.
The thing we have to move toward is a world where the data set owner has a
lot more intelligence about how their data is actually being used and how it
could be used, so they could actually dial it into their plan for how to
actually maintain and update the data, and so that the data set owner is
actually armed with more information, to go back and make the business case
internally for more funding, because they could actually improve public health
at XYZ dramatically, improve health reports to XYZ in a dramatic way, fi they
actually just got the data out on say a monthly basis or a quarterly basis with
a six-month lag, as opposed to that sort of with a three-year lag.
This has already actually been happening, which is fantastic. And I think
that one of the key to-dos, I think, of all of us that care deeply for the
health ecosystem is actually to treat it as a public space which we need to get
the suppliers and the users more closely intertwined and talking to each other
a lot more, whether it be at physical datapaloozas or via what data force will
build in terms of HealthData.gov 2.0, so on and so forth. I think the more
dialogue that happens between suppliers and users, the better the maintenance
of supply will be, and the more targeted and social ROI investments in data
will actually become. I think that is critically important and it is already
I think with respect to the collection of new data, so first of all,
absolutely. Again, it is hard to generalize because the kinds of data that the
government has and is liberating are so diverse. But are these data sets or
data programs iterating? Absolutely, absolutely. Just look at Hospital Compare,
which is evolving to actually add more metrics, with respect to the hospital
conditions. It is actually learning a lot about what is working there and not
working there, and iterating beyond that, so on and so forth.
I actually can’t think of a data set to suggest that is actually static and
will never change ever again. But I think they will all live and breathe and
grow in all kinds of ways to try to add more value. And again, just to kind of
return to the punch line, very important for those investments in data set
evolution to be informed by how best can this data be used to generate value,
especially in the context of a health system that is now halleluiah, moving
toward a world where it is really getting increasingly focused on, and is
actually incentive to and rewarded for and supported to improve care and
improve health and lower cost.
So again, dialogue between the suppliers and the users critically,
critically important. And actually, as a final point, again in the kind of
context of kind of steadily improving and involving data, HHS is just one
supplier. The healthy initiative is not an HHS initiative. It is an American
initiative. We are very happy to be an anchor tenant, we are very happy to be
an anchor supplier. We are very happy to be in the somewhat unfamiliar position
of actually being on the cutting edge of doing something.
But what is actually happening is that other people are joining the party,
other data suppliers are actually joining the party. So the state of New York
actually just announced that it is launching its own healthy initiative. And it
has launched something called the Metrics website, where it is publishing more
and more data sets that are machine readable and downloadable and accessible by
the American public. The state of Louisiana is actually beginning to make a
move in this direction, as well. Private sector companies like Gallup
Healthways are moving in this direction, as well.
Maybe one of the most interesting examples is an example that I talked about
maybe a year ago called Blue Button. Blue Button is an initiative where the
Department of Veterans’ Affairs, the VA, the Department of Defense and CMS
decided about a year and a half ago to do something that seemed incredibly
simple, but was actually in retrospect pretty radical, which was allow
veterans, members of the military, and seniors served by Medicare to actually
download an electronic copy of their own information. And it is actually
something that we weren’t exactly sure how population it would be, because we
said well, how many people really want their own data.
Secretary Shinseki, who was the secretary of the VA, in fact was told that
if the program were wildly successful, that 25,000 veterans would ultimately
choose to download their data. Well, basically no marketing, and the vast
majority of seniors, members of the military and veterans still do not know
Blue Button is out there. But to date, 750,000 veterans, members of the
military and — have actually chosen to access and download a copy of their own
data, on average multiple times each.
And there are anthropological studies now, which have actually shown that
your veterans, what are they doing? Well, they are actually printing it out,
for example, and they are taking it with them to their docs. It turns out that
half of the care provided to American veterans is provided not by the VA. And
what are all the meds they are on, when was the last time they saw a doctor,
what were the diagnoses, what did they say.
Now they are actually passing it along, and the veterans have literally
said, because I did that, my doctor realized I was on a med that I didn’t
remember I was on. And thereby, changed the medication regime away from a
course that actually would have put the veteran in the ER, for example. People
were uploading their personal health records. People were doing all kinds of
stuff with it, which was actually very exciting.
But the maybe even more interesting part of the Blue Button story is that,
after we did Blue Button, word started traveling in certain circles. And we
actually got calls from data holders in the private sector. And the question we
got from them was, are you allowed to do that? And we said, can you clarify the
question? And they said, are you allowed under HIPAA to give patients and
consumers a downloadable, electronic copy of their own information? We said
yes, absolutely, absolutely you are. But I think no sub-regulatory guidance or
detailed memo would have possibly been as persuasive and definitive as
Medicare, the VA, the DOD just doing it, and actually increasing that.
So what is happening now is that, more and more private organizations, like
Aetna, United, Walgreens, Mechassin(?), actually Louisiana, Vermont, et cetera,
have either Blue Buttoned or committed to Blue Button their data. So we think
that actually, by the end of the summer, well north of 50 million Americans,
probably actually closer to 100 million Americans, will actually have access to
Blue Button data. At the end of the day, a bunch of those folks will be
Medicare beneficiaries and veterans and members of the military. But there will
be many, many, many others who access their own data through private
I think at the end of the day, the government is only one holder of health
data. And health data is sort of really about sort of everyone who cares about
health and healthcare for America in responsible ways, improving accessibility
to and use of data, so that we can actually power the kind of health and health
care improvement that this country really, really needs, and which we are
actually beginning to see happen.
DR. TANG: That was a perfect answer and an exciting one. I don’t think if
this is even possible, but I think I am even more enthusiastic than Todd.
Speaking as a provider who not only wants to consume some of the data, that is
out there, but be partnering along with our patients on involving the data set
so it has richer and richer data that is meaningful, and then can be reflected.
It is just totally exciting, thank you so much.
DR. CARR: Okay. So I have Bob, Larry and Walter.
DR. KAPLAN: I am Bob Kaplan from the NIH, and actually I have been in
government a relatively short time. But during that short time, I have had the
opportunity to hear Todd speak several times. Actually, one digression that the
NIH directors meet on Thursday mornings and we invited Todd to come. He is
veering his head. So what happened was, Todd was giving a little talk to the
NIH directors, and his cell phone rang. And he sort of ducked away from the
podium and he said this is my wife. And he said my wife and I are having a
baby. She is on her way to the hospital. And he said, and I have to move
through this next dozen or two slides quickly. And actually, all the physicians
and the NIH directors group seemed much more nervous than Todd was.
My question is, first of all, I think everybody loves all the stuff you are
doing, and it is going in a really interesting direction. But I wanted to raise
another question about the connection between suppliers and users for the
information. So I served on an IOM committee last year that was interested in,
among other things, data needs. And one of the concerns was that they refer to
a problem that they called indicatoritis. And the concern was that people who
are making decisions in health care, primarily directors of public health
departments, are so overwhelmed by data and by indicators, and by the lack of
harmonization of indicators, that they don’t quite know what to do.
And so, if you look at some of our big data needs, a committee like this for
example, how are we doing as a country in advancing the health status of the
population. And this committee concluded that, well, what we really need is we
need a good harmonized summary measure of population health, like a quality or
something like that, in addition to all of the stuff you are doing. I don’t
think it’s either or, but there was concern about are we missing harmonization
or opportunities. Are we sort of feeding indicatoritis without getting at the
basic things we need.
MR. PARK: This is a phenomenal point. I would say you’re absolutely right,
it is not either or. In fact, I think one of the most important goals of data
liberation is to accelerate R&D on what the truly meaningful measures are.
And also, to recognize that meaning, as you know better than anybody, what a
meaningful metric is. In certain cases, it is the same for every American, and
actually in many of the other cases, it is not. And so, I think that the notion
of actually democratizing access to our data, so many, many, many other smart
people, besides just us, can really turbo charge R&D, and what really is a
meaningful measure, and to whom and when. I think it is critically important.
One of my favorite examples of this actually is the section 10332 provision,
that is an unbelievable breakthrough. Don Berwick says it is one of the five
most important things he helped to do at CMS. The whole idea of actually
allowing provider identifiable metric claims data to actually be accessed by
parties outside of Medicare, mashed up with other data that produced truly
comprehensive quality reports and performance reports, that are shared with
providers and then actually vetted and shared with the public.
One of the benefits of this is that, on top of actually allowing critical
mass of data to be put together, to then produce new transparency performance
for providers, which by the way, is not just helpful for consumers, it is very
helpful for providers. So I, as a doctor, know where I stand relative to other
people and how I might improve.
The other provision of 1052, which is I think amazing, is the qualifying
entities are A) supposed to use like this inventory of measures. But then B) if
they get local support, are allowed to come up with new measures, that they
will actually then propagate. And you can imagine then a significant quickening
of our understanding of what really works and what really makes sense to a
doctor, what really makes sense to a patient, because I think all of us would
agree, quality measurement for all of the brilliant work that has been done, is
still in its infancy. And I think a lot of the reason why it hasn’t progressed
as far and as fast is because access to the underlying molecules that you need
to do that kind of research, it hasn’t been as broad and as deep as it should
So I couldn’t agree with you more. I think that the notion of actually
really finding meaning in the data and finding a way to communicate that, in a
way that a doctor, a public health official, a patient can really understand
and use, critically important. And I think that actually data liberation, done
in the right frame, as we are talking about, can actually significantly
accelerate progress on that front.
MR. SOONTHORNSIMA: I will be brief. Thank you very much for those comments.
To follow up on what Todd and Bob were just talking about around data
liberation. I think specifically around performance measure, there is a slide
that you had, Niall, when you touch on those activities that are well underway
of recent progress, well underway such as providing data to HCOs, and the
second bullet point was around Medicare data sharing for performance measure.
On that particular point, and I heard around the room, particularly from
Aetna, that marrying that data from multiple sources, multiple payors, private
and public, would be a lot more meaningful for performance measures. That is
stated. What are then some of the qualifications, you said in your bullet point
that the data would be provided to qualified entities. So what are some of the
parameters, in order to accelerate, last point, and advance some of these
initiatives that are happening across the state, particularly in Louisiana and
other parts of the country?
MR. BRENNAN: So specific to that program, very high level, you need to bring
data from other sources to the table. You need to demonstrate experience in
combining claims data from different sources accurately, into a single
database, either using a centralized or distributed approach, if that is your
choice. You have to have experience in calculating claims and or claims plus
clinical quality measures. Experience in public recording, and that is pretty
You also have to undertake to allow providers that confidentiality review
and appeal the reports before you make them public. And also, I think a key
distinction here, and one that we tried to clarify in the final rule, a
qualified entity is not a monolithic organization. So all of these skills,
because they are pretty highly technical skills, don’t have to be under the one
roof of one company. Qualified entities can meet these criteria through
partnerships and contracting and collaborating, to get the necessary expertise
in the door.
MR. SOONTHORNSIMA: And when you talk about performance measure, are you
talking about at the facility level?
MR. BRENNAN: The individual provider level, technically providers and
suppliers to services. So it could be physicians, hospitals, skilled nursing
facilities, home health agencies. And again, as Todd mentioned, quality
measurement is still, despite a lot of activity over the past 10 years, still
in many respects is in its infancy. So we don’t have a lot of measures for
certain areas, for certain types of providers. We have made a lot of progress
on providers, a little progress on physicians, and then some of the other
areas, not so much. And so, because of this alternative measure process, it
gives people the tools to innovate, in partnership in many respects, with the
providers themselves, to develop measures that are useful to everyone.
DR. GREEN: Without exception, I thought that your presentations were not
only interesting, but they were fun. And I wanted to thank you all for coming.
I would like to make a quick observation, but I mostly wanted to direct a
question to any of you that would be willing to address it. The observation
probably links most to Bill Davenhall’s presentation, and it also cuts back to
the initial slide set. This committee has learned over the last couple of years
that there is an evolving contagion across the country, at the community level
of communities, that are feeling like they need to accept responsibility for
their own health.
And in the users groups that were identified, I didn’t see community as a
user. And yet, Bill was talking about linking data that are geographically
relevant, and to get the accountability requires finally green data sets. That
means this has to be driven down to local levels. And so, my observation is
that we came to the conclusion that the country is missing a critical
infrastructure. When you guys have succeeded at all of this wonderful work, how
does it actually get sorted out at the community level, and where does that
occur? Where is the organizational framework for that? My question is about
workforce. There is a lot of analytic work to do here. Who is producing the
analytic workforce that is going to be needed?
MR. DAVENHALL: I will jump in to that. Most of that is being done through
community health advocacy groups, which DHHS has sponsored for many, many
years, really empowering the local community. Most of them work, as I said, in
murkiness. They haven’t had necessarily the tools and they certainly haven’t
had all the data they need. So they are actually being forced to deal with
national policy implemented at the local level in the absence of data.
And I would say when Todd first started to talk about the Blue Button, I
said I want a Blue Button for communities. That is what communities want. They
want their own Blue Button for their community. They want to know how their
health as a whole community sort of speaks to the population health issue. But
right now, to assemble that kind of Blue Button for a community, I would have
to think about that for a while. That is not an easy go to create that.
MR. PARK: Just to build on that, actually interestingly, the entire health
data initiative got started as the community health data initiative. Actually,
it was initially focused on community empowerment and community health data.
And so, one of the early products was this help indicators warehouse, where we,
for the first time, amalgamated across every single HHS agency and other
sources, like EPA and USDA, 1200 metrics of national, state, hospital region
and county level public health performance, health care system performance, and
determinates of health performance, like access to healthy food.
And that was actually where Medicare, for the first time, debuted a whole
raft of prevalence of disease, prevention, quality, utilization of health care
service data at the HR level. And it is available at HealthIndicators.gov, and
it is available both as a website and via a webservices API, where people can
actually access and integrate the data into lots of other tools, which they
Now, one very interesting, speaking of the dialogue between suppliers and
users, now they have actually made that data much more accessible and exposed
to a lot more people. We are getting a lot of feedback about the many, many,
many ways we can actually make it better. So actually, people asked for new
indicators. So actually in March, there will be 53 new indicators, now derived
from Medicaid, community health center population, new prevalence of disease
metrics that Medicare has actually produced, Lyme disease, which I am going to
check that out immediately.
A action second point, said you know what, I know why you did HRR. But you
know what, I don’t think of my community as an HRR. And in fact, the Los
Angeles HRR, it is like a small country. It has got 20 million people in it, so
it is not incredibly helpful. So one of the ideas that we have been exploring,
actually based on the community Blue Button idea, based on Bill’s comment, is
called Choose Your Own Adventure, Choose Your Own Community Tool, where
essentially what will happen is hypothetically, you could pick ZIP Codes and
say that is my community.
And then, essentially CMS would auto calculate metrics, and it would
suppress them of course if there were privacy problems, because there are too
few data points. It would suppress if the statistics weren’t actually valid.
But they seem to recognize the fact that my community and your community may
not be different. We may live in the same place, but we may have different
definitions of community because we are both considering different things. And
so that, to me, is just freaking awesome. And Medicare can do that because it’s
got all of this micro data that it can actually draw upon.
The third thing people have said is they have said it is great that you have
this data, but it is from 2008. And so, he had the best zinger, and it wasn’t
aimed at us, it was just aimed at health care in general. He said health care
data these days, and he was talking in the past, it’s like having a speedometer
on your car, and having it tell you how fast someone else was driving down this
highway three years ago. Which I guess at some level, it might be useful, but
really not very useful if what you are trying to do is optimize performance.
So CMS is actually going to be loading up very soon, ’09 and ’10 data, into
the health indicator warehouse. And one of the things we are exploring, and I
have no idea if this is actually doable, but I have asked Niall and Niall is an
incredible dude and he does impossible things. I said well, what if we could
actually present community health performance data, it yields agency services,
quality, prevention, et cetera, quarterly, with a quarter lag. And Niall said,
you know, let’s take a look at that.
And I think the caveat of that would be, of course, that you haven’t had 100
percent run out and all of the things aren’t final yet. But I think people
understand that. We live in a world where US government publishes GDP growth
stats and unemployment rates every quarter, and then revises them later. And
everyone understands, no one actually gets indignant when the rates get revised
because they understand it is preliminary. They would much rather know what
unemployment was last quarter in some reasonable timeframe, then have to wait
for 18 months to actually get it right.
But again, the common denominator on all of that, it’s actually going back
to one of the themes I was articulating, is that we know this now, because data
users came to us and said, if you did this, here is what we could do to improve
health care. And so, one of the things I think this committee could do, one of
the things I think we actually want to do in general is just get more of that
to happen. And we will respond, especially now that we have a data and
information products line of business, a specialized unit, that is going to
focus on doing nothing but increase the access to utility of our data to
improve health and health care.
DR. KELLY: One thing that I really do think will change things a lot in the
next couple of years is the emergency of the ACOs, because ACOs almost always
are local structures. In the ACO model, you begin to line incentives correctly.
That local structure now has an incentive to keep people healthy. I will tell
you, I spend my first two years at Aetna as the head doc for our large
commercial clients. So I can tell you those employers that have large
populations in those areas where there are ACOs are going to be very interested
I can also tell you that there are a lot of retailers that are out there
right now, that are extremely interested in figuring out how do they play in
the community to foster health. And if we get the incentives aligned at the
provider level, just to keep people healthy, they are going to want to work
with the community to lessen the incidence of diabetes and obesity and
DR. GREEN: How will they do that?
DR. KELLY: I think they are going to get very creative, because the best
thing you can do is align incentives, and let very smart doctors and very smart
providers figure it out. They will do it in a million different ways, but they
will do it because they will have data to say you know what? I am actually
going to not only improve the quality of my community, but I am going to get
paid more if I have a healthier population because they are going to drive
lower risk costs over time. So if you align the incentives and you empower
people at the local community, they will engage the ecosystem in ways that they
haven’t, because the incentives have to be aligned.
DR. ROSENTHAL: That might be where the marketplace comes into play, as well.
There is probably a dozen entrepreneurial groups that wanted to do something
like this, to take a product out for these. Now that the financial incentives
are aligned at the community level, they want to build something like this. I
am thinking of three or four right off the top of my head. They kind of do it
because they didn’t have access to the data.
I think one of the risks, and I don’t mean this is an real risk sense, is
that you are going to have to be very smart about where you have official data
products and information products and services, versus where you push things
out and let the market take over. You can do that on your own, but to your
point, you are already being incredibly intelligent about this. You are not
going to be able to do it as smart as MIT Harvard kids, and that is where you
need to let them do it. And there are a couple of dozen groups that are playing
around, but wanting to bring something like that.
Now, why are they doing that? Why didn’t they do it before? Because for the
first time, the money is there, meaning incentives, so they are going to do it.
And for the first time, the data is actually there to allow them to do that. So
creating those conditions is absolutely imperative.
And on that note, if you ask how do we know what we need to do, you can
always get user feedback, you can always have groups of people in a room like
this. But what do we know and what do I know. If you actually build the
infrastructure, you can do data driven product development, or data driven
information development, meaning you look at what are the files people are
using in the enclave. You look at what are the views they are talking about,
you look at what are the discussions, and they leave the comments. You can
definitely have the live stuff, it’s not either or. And this is tried and true
in financial services and ecom and everything else outside of health care. So
you set up the incentives right, which you have done, that is why I said you
have done the bulk of the heavy work. You liberate the data so people can come
in, smarter people than anyone here, no offense, myself twice included. And you
let them go at it.
DR. GREEN: I take it that none of you think that there are any workforce
DR. PARK: There are absolutely workforce issues. But supply, in this case,
follows demand. And so, one of the things that happened at the last health
datapalooza is the University of Michigan announced the launch of the nation’s
first consumer health informatics program, which they promptly stole Chuck
Friedman from ONC to go run. But it is a joint venture of their School of
Information, which is the rebranded School of Library Sciences, and the School
of Public Health. And that will absolutely become the way of the future.
And their only problem is that the 30 people that they actually are going to
graduate A) they are going to be 20 times that number that apply, and B) those
people are going to be fought over viciously by all of these entrepreneurial
groups and new outfits that are actually trying to build the services to
actually help docs and hospitals succeed as ACOs. There is going to be, there
already is, rapidly rising demand for this kind of expertise.
DR. CARR: I would just add, wearing my other hat at Steward Health Care,
where we are one of the 32 pioneer ACOs, almost overnight, the dialogue changed
and we have already created community health workers that are culturally
compatible, to understand why the care that was given, now the receptor arm,
why wasn’t it working or why did they not follow up on the appointment, why did
they not take the medicine, why did they not let homecare come into their home.
And it is very exciting to see, and I think in many ways, it is analogous,
you put up the map or you put the picture, and suddenly it’s like, oh, why
didn’t we think of that. So Bob, I know you want to speak, but we have Walter
and Judy, and then we have public comments. So I don’t know if folks will be
around for a couple of minutes at the end, but I just want to make sure that we
give fair amount of time. So it is Walter and then Judy, and then I think at a
quarter of 4:00, we have two public comments.
DR. SUAREZ: Thank you very much for those wonderful presentations. I am
baffled by the fact that we are not only seeing a liberation of the data, but
we are also seeing a growth of the number of data collections and databases
that are being created, which is a challenge, too. And that is the question I
have, it goes out of my just simple count, we have HIEs now that are going to
begin to collect data, or are collecting already.
We have certainly CMS collecting not just Medicare data, but also collecting
potentially multi-payor claims data. We have states collecting data across the
board. OPM, the Officer of Personnel Management, trying to create a federal
employee claims base data. I think Bill mentioned the Health Care Costs
Institute, where more payors are trying to get this.
One factor is all of this data still is a payor-based or a payor-driven
data, or a lot of the detail and counter level data is coming from that source,
which incidentally in light of the discussions that we had in the last two days
on ICD coding, it is based on ICD coding. And it is based on ICD-9 coding at
this point, so it will be interesting to see when we go to the ICD-10 code, the
significant benefits that that will bring.
I was looking at the maps that you were showing, and I was dreaming of how,
in Google, you can go today and begin to zoom into a particular area, and go
even down to the house or the street, and how we could try to do that kind of
same analysis on a particular condition or a particular topic in health, and go
from the country to a state to a city to a block, and even to a particular
But my question is then, with all of this plethora of data sources and data
collection efforts that are happening, how can we ensure that there is some
consistency and harmonization really across them, so that when someone is doing
an analysis on some topic, using some of that data, the results are not going
to be widely different. And because again, Kaiser is an example. We are 10
different states, and the district we participate in a number of these data
efforts. You can pull out data from each of those sources and come out with
completely different stories about the same topic.
So how can we create that consistency in the source of the data, so that the
analysis of the data, which is at end of purpose, will be coming up with
consistent information and not misleading information?
MR. BRENNAN: So first of all, I bet Josh has some opinions to share on this.
But it is sort of the classic, again, example of how data liberation or data
sharing can actually help. Because right now, what you have is a lot of people
analyzing their data in silos, and making individual decisions while I am going
to attribute physicians based on a 30 percent threshold. But I am going to
attribute physicians based on a 35 percent threshold, or I am going to say a
gold star physician meets the clinical criteria, nine times out of 10.
And I am going to say a gold star physician meets the criteria seven times
out of 10. And that is how you end up in a situation that Bill referenced in
Massachusetts, where the group insurance committee gave combined data to
different insurers, and ended up with one insurer said a physician was good and
one insurer said the same physician was mediocre, and a third insurer said the
physician was bad.
When you start to promote data sharing, I think that inherently promotes
standardization and harmonization. Now, I am not saying it is going to happen
by magic or it is going to happen overnight, and that is where also things
like, I am thinking of George Thomas’ stuff, Todd, better data tagging, like
better sort of Medicare dual-eligible beneficiary like there is just one
definition of what a dual-eligible is. And we know the data points that go into
making them a dual-eligible, whereas right now, about eight people have eight
different opinions over what constitutes a dual-eligible, even at CMS, let
alone outside of the building. I don’t know if anybody else wants to add to
DR. ROSENTHAL: It might be worth looking, I put together some of Hendesey’s
and my slides, addressing that particular issue, in doing it in two ways. But I
am calling taxonomy, which means structuring the data in a specific way, and
then sharing, which is clarifying it. And by that, obviously on one hand, data
standardization becomes more important and more difficult. But by liberating
it, it is going to naturally happen, provided you have some basic structure.
Everyone does taxonomy, you either do it well or you do it poorly, that is
the question. And by that, I mean, please look at the slides. You will see you
are going to have to get better about some of the basics, like consistency. If
something is missing in a cell, you can’t have free-hand typing three sentences
and things like that. If contract entity means something in 2009, guess what,
it means something different in 2010, you need to label it. So there are some
basic things you are going to find around that. But by sharing that data with a
community, putting in a center where you can capture that, the people doing
that work will be your users largely.
But you do have to display your taxonomy, because you do have a relationship
with the payor, the contract provider, you have various structures embedded in
those files. They aren’t available in the documentation. And by showing
pictures of that to the users, that will provide the framework to allow them to
do that work, provided you are actually able to allow them to share that. So on
one hand, you are going to have to do a little bit better job around increasing
the transparency of what you already done, because you have already done that
difficult work and giving them the electronic means to share and do that work
for you. And that is not unlike you find in other industries, by the way.
DR. KELLY: I absolutely agree with all of that. I also think, though, that
the incentives in the new models, and the engagement of the consumer, I found
if the consumer is a doc is actually a powerful QA tool. If they download a
copy of their Blue Button and they are not on these drugs, they have never seen
this diagnosis before, they are usually pretty vocal about it to me. So that,
and actually the fact that, with the pioneer grants, with all these ACOs, we
are jointly defining very specifically how quality will be measured, and how
they will be paid. So I can tell you that people will now pay a lot more
attention, at least those initial measures. So we are not going to get there
overnight, but I actually think the consumer and these new tandem models will
drive better data policy.
DR. ROSENTHAL: One other quick thing, these conversations are not unrelated.
What are the meaningful quality metrics and what is the data consistently?
Those are both around user creation transparency and taxonomy, to say I take
this molecule and this molecule, and I combine it.
DR. WARREN: Mine is kind of very detailed. So when you talk about you can
look at the data according to provider, what kind of providers can I find out
MR. BRENNAN: Are you talking specifically about 10332, the Medicare data
sharing program that we have referenced, or sort of our more general ability to
identify providers and our data?
DR. WARREN: Well, both. It was the former, but I want the answer to the
second one, too.
MR. BRENNAN: Qualified entities will get 100 percent extracts of A, B and D
data. So theoretically, that is going to enable them to measure any type of
provider that we do business with, because the provider IDs will be available
on the files.
DR. WARREN: And you’re talking about NPI?
MR. BRENNAN: Yes, NPI will be on the file, and also some of the more legacy
provider identifiers, like UPIN and the like. So you should be able to identify
every provider. I know that there are some issues that were also dealing with,
in parallel below other activities at CMS, regarding sort of like a gold record
for providers, and ensuring that somebody really just has one NPI. But
generally speaking, we believe the provider data is pretty solid.
DR. WARREN: So if I wanted to go into this database and find out for a
community, what was the care provided by hospitals, long-term care, homecare,
physicians, clinical psychologists, nurse practitioners, physical therapists,
can I get down to that granular level?
MR. BRENNAN: So again, I want to make sure we are talking about the right
thing. If you are approved as a qualified entity, and you have private sector
claims data, and you combine it with Medicare claims data, yes, you can. If you
are talking about the Health Indicators Warehouse, which is at the hospital
referral region level, or future iterations of that, which may be define your
own community, you won’t be able to identify individual hospitals. But what you
will be able to do is say, our long-term care hospital utilization rate is
three times the national average, why?
DR. CARR: I would like to now open it up for public comment, and the first
person from the state of Louisiana Bureau of Policy Research and Program
Development, Lucas Tramontuzzi.
MR. TRAMONTUZZI: Thank you very much. I will be brief because I know
everyone has flights to take. My name is Lucas Tramontuzzi. I am the chief data
officer for the Louisiana Department of Health and Hospitals. I would like to
thank this committee for the opportunity to briefly share one state’s
perspective about how we use data from CMS and from HHS generally speaking, to
try to provide value for consumers and citizens.
The Department of Health and Hospitals contains both our Medicaid agency, as
well as our Office of Public Health, Office of Aging, Office of Citizens with
Disability and our Behavioral Health, in order to serve 1.1 million Medicaid
recipients of our 4.5 million residents. And that number, 1.1, we anticipate to
rise closer to 2 by 2014. And so, as a result, and going back to Bill’s point,
I saw exactly where those dark spots were on the map, and they were all on
Louisiana. And so, it goes to show that we are ranked 49th for a
reason, and it is a combination of uncoordinated care, we have patients with
chronic diseases, with not the right management and engagement. And so, it is a
We have a workforce issue. We do not have the horses in our analytic stable
to try to address this as a department or as a state. And so, the one thing
that we have to do as a state is really reach out to your private partners,
public partners, in order to work together to figure out how are we going to
And so, one of our challenges when it comes to data is how do we make sure
we have the authorization in order to be able to work together, to make these
differences. There are a couple of requests that I have made CMS, and that I
have posted to this committee. We need to minimize the operational barriers to
access the data. For example, the usability of parts A, B and D, the CCW versus
the TAP files. We need to get the standard formats, the historic versus more
timely data, it is an issue. For each year that data is delayed, it just makes
it difficult to help build out things like the tumor registries and other data
sources that we need as a state, not just the department, but as the state in
order to provide better care.
Content, certainly we are all uber concerned about ensuring confidentiality
of our residence. There is technology in place, we can use double hex
encryption. There are ways to do it and we shouldn’t allow that to be a
barrier. But we really need to have conversations very quickly about how do we
best do it, so again, we are not having these massive data dumps into places
and trying to do it after the fact. We need to really think more forward about
how could we do it, so then that way we could start linking it appropriately
and start doing some amazing things with it.
The next is the standardized and the practice of the painting the data. We
states have gotten very good about sharing the ways in which we are able to get
it. Hey, this is my application, this is what I wrote. Here, you try writing
the same thing and see if you can get the data. And sure enough, it doesn’t
work out. And so, we really need to work together to figure out how do we make
it more predictable to do that.
The other thing is, it is very, very difficult just to try to anticipate
tomorrow’s research questions. So when we request the data, we have to request
it for a specific use. But if something comes up tomorrow, we can’t use the
data to research that question. And so, we need to be allowed some flexibility
in what we can do once we get it. Again, we want to make sure it gets used
appropriately. Hold us to those high standards, and prosecute us when we don’t
meet those standards. But we need some flexibility.
The third part is emergency management. The MDS data is invaluable because
it is our only source of really knowing how is in a nursing home. Two hundred
and eighty two of the 285 nursing homes in Louisiana report to MDS. When a
hurricane comes in, we need to turn on our systems. We are not allowed to bring
that patient level data into our management system. So when a nursing home
evacuates, we don’t know who is in that nursing home. Unless we ask the nursing
homes to actually type in each recipient that they are moving. It is a waste of
They need to be working very quickly to get those patients out of there,
instead of trying to report to us, again, the state, not the department. The
data is already there. Help us leverage that, so then that way, we can put
emphasis on those time-crunching moments into the right activities.
Another part is just coordination with Social Security. It is nearly
impossible for us to manage the eligibility between Bendix and the EBD as we
manage where these recipients are going. We need better cooperation with the
other federal agencies, because again, we are just losing time and energy, and
trying to match who is going to qualify for whatever the appropriate program
The last time I ask is your support, and we support Todd and his initiatives
to formulate these data commons for researchers and others, to leverage in a
way that is meaningful. Not only for the community, because again what we see
in Louisiana is that it is at that level that we are going to get the impact.
Also, allow us to work with you, to develop a single Blue Button, a single
place for our residents to get all of their data. It is great that you can do
it for CMS, it is great that you can do it for Blue Cross or for the other
ones. But as a parent, my children’s data is with one provider, my wife’s in
another, I am in another. By the time we get to CMS, our data is all over the
We want to work as a state to be able to unify in one place, so then that
way, from the user standpoint, I don’t have to fish for it. It is all right
there for me, as I move across the system from private insurance, eventually
into Medicare, if it still exits. But again, it would be nice to have it in one
place, and to allow states to work with you, to kind of bring it together. And
with that, again I would like to remind the committee, please be agents of
change. States are so reliant on you to help guide. You really are in a unique
position to help move us all along. And again, I thank you for this
DR. CARR: We have one additional public comment, from New York City
Department of Health, Pat Lynch calling in. Do we have Pat Lynch or designee
calling in? Okay.
MS. GREENBERG: I just wanted to make sure people also realize that they can
DR. CARR: While Marjorie is looking for that, is there anyone else in the
room who would like to make a comment?
MS. GREENBERG: To CMS.data@CMS.hhs.gov until March 16th.
DR. WARREN: I have a question. Given the nature of our discussion and debate
about ICD-10, what would be the impact to all of the ability for you to mash up
your data and do reports, if we delayed implementation of ICD-10? And I am
especially interested, as you start describing communities and populations.
MR. BRENNAN: Well, I would say we are reasonably comfortable with the way we
use and analyze ICD-9 data. So if ICD-10 data were delayed, it wouldn’t
necessarily affect many of our current and projected plans. What it would
effect would be potentially the content of some of those products, because
ICD-10 has so much more granular detail.
DR. SUAREZ: Have there been any analysis on your part then on the benefits
of having, and perhaps even examples, data quoted on ICD-10, to generate some
of the reports, and some of the information that you are producing? Has there
been any examples or analysis done about that?
MR. BRENNAN: To be honest, that is a little bit outside of my personal
bailiwick. CMS or I don’t know if other health plan reps, or Todd wants to
comment. Again, I obviously know that ICD-10 is considered by many to be a much
richer source of coding information. But let’s not forget that ICD-9 is five or
however many thousand ways of identifying clinical conditions, too.
DR. CARR: Okay, I think with that, I would like to make just a couple of
closing comments. So NCVHS has been in existence for over 60 years, coming up
on 63 years, I think. And as you look over the history, there are kind of
moments in time that ignite excitement and change and imagination. I think in
2002, 10 years ago, we published Shaping Health Statistics Vision for the
21st Century. And interestingly, at that time, we identified that
the population’s health was about health and disease, but functional status and
But we also identified community attributes, context place and time. And it
is very exciting that now 10 years later, when we look at the way we are now
taking the data available to us, and putting it on maps in communities, mixing
up medical data with community data, financial data, employment data, we are
actually approximating what was the vision for the 21st century, 10
So I want to thank you tremendously for your very exciting, illuminating and
stimulating presentations. You have really inspired us by your leadership and
sparked our imagination by your creativity, and we look forward to ongoing
collaboration and participation with you, at any time, in any way that you see
fit. So thank you very much.
And with that, we will adjourn the National Committee on Vital and Health
Statistics meeting. Thank you very much.
(Whereupon, the meeting adjourned at 4:00 p.m.)