[This Transcript is Unedited]
DEPARTMENT OF HEALTH AND HUMAN SERVICES
NATIONAL COMMITTEE ON VITAL AND HEALTH STATISTICS
February 21, 2008
Hilton Garden Inn Hotel
815 14th Street, NW
Washington, D.C.
Proceedings by:
CASET Associates, Ltd.
10201 Lee Highway
Fairfax, Virginia 22030
www.caset.net
TABLE OF CONTENTS
- Call to Order, Welcome and Introductions
- Committee Operations and Process Document
- Subcommittee Sessions
- ICD-10/11 Update: Christopher Chute
- AHIMA/AMIA Briefing – Linda Kloss, Keith Campbell
P R O C E E D I N G S (8:40 a.m.)
Agenda Item: Call to Order, Welcome and Introductions
DR. COHN: Will everyone please be seated? We are going to get started here.
Good morning. I want to call this meeting to order. This is the second day
of meetings for the National Committee on Vital and Health Statistics. The
National Committee is the statutory public advisory committee to the U.S.
Department of Health and Human Services on national health information policy.
I am Simon Cohn. I am the Associate Executive Director for Health Information
Policy for Kaiser Permanente and Chair of the committee.
I want to welcome committee members here this morning, HHS staff and others
here in person, as well as those listening in on the Internet. I want to remind
everyone to speak clearly and into the microphone.
With that, let’s have introductions around the table and then around the
room. If there is anyone on the phone, we will grab them for introductions
afterwards.
For those on the National Committee, I would ask if there are any conflicts
of interest related to any of the issues coming before us today, would you so
publicly indicate during your introductions. I want to begin by observing that
I have no conflicts of interest.
MS.. GREENBERG: Good morning. I am Marjorie Greenberg, National Center for
Health Statistics, CDC, and Executive Secretary to the committee.
MR. BLAIR: Jeff Blair, Loveless Clinic Foundation. The only potential
conflicts of interest that I might have is if it relates to health information
exchange networks.
DR. STEINDEL: Steve Steindel, Centers for Disease Control and Prevention,
liaison to the committee.
John Houston, University of Pittsburgh Medical Center, member of the
committee, and I have no conflicts.
DR. MC CALL: Carol McCall, member of the committee, no conflicts.
DR. STEINWACHS: Don Steinwachs, Johns Hopkins University, member of the
committee, no conflicts.
DR. STEUERLE: Gene Steuerle, the Orbit Institute, member of the committee,
no conflicts.
DR. OVERHAGE: Marc Overhage, Regenstrief Institute, Indiana Health
Information Exchange, member of the committee and no conflicts.
DR. FERRER: Jorge Ferrer, Veterans Health Administration, member of the
committee.
DR. FRANCIS: Leslie Francis, University of Utah, member of the committee
and no conflicts.
DR. FITZMAURICE: I am Michael Fitzmaurice, Agency for Health Care Research
and Quality, liaison to the full committee.
DR. CARR: Justine Carr, Beth Israel Deaconess Medical Center, member of the
committee, no conflicts.
MR. REYNOLDS: Harry Reynolds, Blue Cross Blue Shield of North Carolina,
member of the committee, no conflicts.
MR. ROTHSTEIN: Mark Rothstein, University of Louisville, member of the
committee, no conflicts.
MR. SCANLON: Good morning. Jim Scanlon, HHS, Executive Staff Director for
the full committee.
MS. PICKETT: Donna Pickett, National Center for Health Statistics, CDC and
staff to the Subcommittee on Standards and Security.
MS. JONES: Catherine Jones, National Center for Health Statistics and staff
to the Executive Subcommittee.
MS. DAWSON: Kirsten Dawson from America’s Health Insurance Plan.
MS. KLOSS: Linda Kloss, American Health Information Management Association.
MS. BOWMAN: Sue Bowman, American Health Information and Management
Association.
MR. RODE: Dan Rode, American Health Information and Management Association.
MS. JACKSON: Debbie Jackson, National Center for Health Statistics, CDC,
committee staff.
MR. BIRNBAUM: Adam Birnbaum, Blue Cross Blue Shield Association.
MR. CONNOLLY: David Connolly, Capital Associates.
MS. JAMISON: Missy Jamison, National Center for Health Statistics, CDC.
MS. KAHN: Hetty Kahn, National Center for Health Statistics, CDC, staff to
the Subcommittee on Privacy and Confidentiality.
MS. HORLICK: Gail Horlick, CDC Atlanta, staff to the Subcommittee on
Privacy and Confidentiality.
MS. SEDEKIS: Maurie Sedekis, American Medical Association.
DR. COHN: I want to welcome everyone, and congratulate the committee on
what I think was a very productive first day. We will talk about the agenda
here in just a minute. We are moving things around a little bit only in the
sense that our main action items —
DR. WARREN: Simon?
DR. COHN: I’m sorry. Who is on?
DR. WARREN: That’s okay, I’ll jump in. This is Judy Warren, member of the
committee from the University of Kansas, no conflicts.
DR. COHN: Judy, thank you for joining us. Anyone else on the phone?
DR. TANG: Yes. Paul Tang, Palo Alto Medical Foundation, member of the
committee, no conflicts.
DR. COHN: Great. Anyone else on the phone? Judy, thank you very much for
reminding me. We started out by offering you all the opportunity to introduce
yourselves.
I just wanted to talk about the agenda today. As I commented, one of the
main action items for the meeting, which was the letter on individual control
of sensitive health information, was deliberated and passed yesterday. It has
been referred to the Executive Committee for final wordsmithing, the addition
of a couple of sentences in the introductory area. That will be worked on as I
said by the Executive Subcommittee. Hopefully it will be posted within the next
several weeks after they have had a chance to review the final draft.
So with that, what we are going to be doing is moving the agenda around a
little bit. As a first item I do want to have us talk about the committee
operations and process document in Tab 5, recognizing that we have reviewed
that a couple of times previously. I am hoping that we are ready to consider it
and pass it as a base document for inclusion in our handbook as part of
operations policy.
From there, we are going to begin to talk about subcommittee and work group
activities operations and plans. I think Justine was very appropriate in some
of her comments yesterday about the fact that there are a lot of things
happening in subcommittees and work groups, and we need to somehow take a look
at them from a wider view, recognizing that there are only so many resources or
people on the National Committee, and we need to figure out what is the most
important to do. We need to figure out whether the committees are being staffed
appropriately to take on that role. I think Harry and Justine yesterday brought
up the issue of how many of these things are cross-cutting, and does there need
to be either additional formulations of work groups or subcommittees, ad hoc
work groups spun off to handle specific areas.
So as we talk through what the subcommittees and work groups are thinking
of doing, we should keep that in the back of our mind. Once again, we have many
ways of doing this. What we have tried to do is introduce a flexible structure
into the committee, with the idea being that we may need to augment in certain
cases work groups or subcommittees, if they are the right groups to lead on
something. In other cases it may be absolutely fine as it is. So I think that
is part of the conversation.
From there, by that time we will probably be ready for a break. After the
break this morning, we do have an update from Chris Chute, Professor of Medical
Informatics at Mayo, on the ICD 10/11 activity. This is an update to the
committee and briefing.
Then we are very pleased to have both Linda Kloss and Keith Campbell. Linda
is Executive Director from AHIMA. Keith Campbell will be representing AMIA,
talking about work that they have been doing looking at the overall issues of
clinical terminologies and clinical administrative data in the context of
health care, which will be another briefing and discussion.
The intent will be to adjourn by around 12:00 o’clock, just recognizing
that we are a little ahead of our agenda.
I see we have had one member arrive since we have done the introductions.
Would you like to introduce yourself?
DR. GREEN: Larry Green, member, no conflicts.
DR. COHN: Thank you. I guess I should ask before we move into the committee
operations document, does anybody have anything else that we need to have
included in the agenda today?
Agenda Item: Committee Operations and Process
Document
DR. COHN: With that, why don’t we turn to Tab 5. As I said, this is
something that we have been working on now at committee members’ requests for a
couple of meetings, and have received input from a number of you in terms of
additions, changes, deletions, and just broadly lays out how things get
identified as a work group item or subcommittee item, how we go through the
drafting, review of them, when it is appropriate to take something from the
work group or subcommittee for wider input. Sometimes that happens before the
full committee meeting. Once again, it is just meant to provide some rules of
the road and advice for subcommittees and work groups going forward.
I think Leslie asked for some modifications last time, which I think have
been incorporated in this. I guess I would ask for any comment or discussion in
relationship to this document at this point.
MR. HOUSTON: Just one comment, and I don’t know if it is appropriate to put
in this document some reference to the fact that I think the expectation is
that committee members need to be timely in their comments and their input on
activities that are going on and documents that are distributed to the
committee members.
As an example, this whole last privacy letter took far too long. A number
of things tried to address concerns, and other ones got put on the table later.
I think a lot of the concerns could have been addressed in a more timely manner
earlier on if they would have gotten on the table earlier.
So I don’t know if there is a way to put something like that in this
document, but I think committee members need to voice concerns at the earliest
opportunity when things are brought forward. I really believe that that hurt
this last letter that we were trying to do. People sat, and all of a sudden
three-quarters of the way through the process, somebody raises an issue that
frankly could have been or should have been raised much earlier on in
deliberation about the letter.
So I just wish there could be something in here that talks to that point.
MS. GREENBERG: This might be here, midway down. It says, any NCVHS member
upon request may serve as a corresponding member of a subcommittee or work
group and receive draft documents.
MR. BLAIR: Microphone.
MS. GREENBERG: Oh, I’m sorry. About a third of the way down it mentions
that any NCVHS member upon request may serve as a corresponding member of a
subcommittee or work group and receive draft documents for review and other
communications of the group.
We could put another bullet there also after that, something about,
committee members should provide timely comments on the drafts as they are
circulated to facilitate action when brought for full committee discussion or
approval or something.
MR. HOUSTON: Again, I don’t think I am asking for — if a committee hasn’t
been distributed a document, obviously they can’t comment on it. But as things
are distributed for review, I think it behooves us to insure that we review
them, and if we have comments, provide them.
DR. CARR: I would like to just add to that. I think we talked yesterday
about the great benefit there was for the uses of data to have the reviewer
contingent. It was very helpful to say if you are not on the first tier, send
it on to the second tier.
But I also appreciate what you are saying, that we need to get the comments
at a time when we can act on them.
MR. ROTHSTEIN: We may also want to add something that says, as appropriate
efforts should be undertaken to include the full committee in the deliberative
process, which might mean open conference calls or briefings at a meeting along
the deliberative process, the way we tried to do that with the last letter, so
that people weren’t seeing it for the first time when it was coming up for a
vote.
DR. CARR: Right, I think maybe that is what I was trying to say. Maybe not
making it elective whether you join in but expected that you participate in one
of the early briefings so that we are never finding ourselves seeing something
for the first time on the plane.
DR. TANG: I don’t know whether you completed the discussion yesterday that
started out in the morning. I think this may have been something that John
Houston raised. I know we would like to have, and it is a good thing to strive
towards consensus in all of the deliberations and final recommendations. To the
extent that there is a super majority in a group of eight, it is seven folks
feeling one way and one person feeling another way, sometimes I wonder if it
would help the timeliness and the impact of the recommendation if we are able
to in those rare circumstances go ahead and go forward with the recommendation
with a super majority, not as a standard measure. It is almost like Margaret
was saying; sometimes you would do things with a minority opinion.
It may be worthwhile to consider whether that is an okay thing to do, both
for timeliness and for impact.
MS. GREENBERG: Well, the rules of the committee are that you just need a
majority to pass anything, though usually, you’re right, we certainly strive
for more than that.
That last dot point does say the opportunity exists for any letter brought
forward to include dissenting comments from the subcommittee or work group
members or other NCVHS members. So on occasion that does happen. I think it is
part of the process.
Was there something additional you wanted to add, Paul?
DR. TANG: I think that has always been available, but something that has
been so discouraged that it may end up holding up every decision to make sure
there is complete consensus. Then the recommendations coming forward may be
watered down so much that it is not as impactful.
It is almost philosophical. I can see how the process accommodates that. It
is more a discussion point, could there be a little loosening up of that kind
of approach, a philosophy of how we get things out of subcommittee or
committee.
DR. MC CALL: It sounds like the rules are in place. This is a matter of
using them when needed.
Back to the comments about having reviewers and/or previous sessions, I
think that both of those are good ideas. I think what I would recommend in
using those is that we use them judiciously. I think with the enhanced uses of
data, that is partially because it was very big and moving very fast. So it was
good to have another set of eyes. But for something that is smaller and more
technical, that may not be the case.
So as you guys decide and develop who the teams are, if they are ad hoc
groups or whatever, that I think will be an important decision, how to
structure some of that and what the needs are.
DR. STEINWACHS: Justine was raising yesterday about priority setting
processes. Our first two bullets talk about a process that starts things from
subcommittees, brings them here, but there is no bullet or little words here
that say that what the subcommittee wants to do or elaborate in the full
committee may not be possible if there are resource limits or a need to set
broader priorities.
I just thought as this document is trying to make us sensitive to things,
it may be in those first two bullets, or there needs to be a third bullet, I
don’t know, but something that recognizes that it may not be possible for
everything that we would like to move ahead to move ahead, or it may be needed.
We have had retreats at times for strategic planning. Another view is that we
try and use mechanisms to periodically place priorities and therefore identify
things that would move ahead and some things that would not.
It just might be helpful. It doesn’t require that everything fit into that,
but to have that third bullet.
MR. REYNOLDS: One thing Bill Scanlon said yesterday has resonated with me
quite a bit last night and today, and that is talking about the idea of values.
If you look at a lot of the documents we did, they were about information, they
were about data, they were about standards. They were about things where people
would have maybe differing opinions about how to think about it and exactly how
it should occur and what should be included and so on. But it might not really
get at the value system, about what people in their gut have to feel and have
to think and have to believe.
So as we look at subjects going forward, and especially, the more we deal
with subjects that go horizontally in NHIN and other things, we are starting to
get into some value situations more and more. I thought Bill really captured
that well.
So I think that is another filter. As we start a subject, knowing that it
is starting to lean on some values, we may think about it differently. We may
make sure that we adjudicate it different, think about it different, talk about
it different, just because it starts to get at the real goods. It is something
that is really at the base of how we all think and how we all feel.
So if you take the privacy letter, just stepping back from it, individuals
want certain things and doctors want certain things at the base. I want the
information I can to treat the patient, and the patient says, I want to tell
you what I want to tell you. Those are values that start to become involved,
not just, I am giving you this data.
So that is where I think we have got to be careful, especially as we are
dealing more and more with how this information is going to be used, because
the technology has almost become secondary. It is not, what are we doing with
each other. So that may be another filter that we add in as we pick subjects
and how we deal with them ongoing.
DR. FRANCIS: This is a comment first about tone, and then a suggestion.
This document reads as a dissent against someone being left out, particularly
the second sentence, and efforts to assure that members have the opportunity to
engage fully in all aspects of committee operations, and that all processes are
transparent.
I agree that is really important, but I don’t think it is the only thing we
are trying to do with this document. I think the other thing we are trying to
do is the cross-cutting question. What we want to try to do is, even if we have
different subcommittees, we want to be able to call on the expertise of people
from different areas.
So I guess the place where I would really push that is where it says any
NCVHS member upon request may serve as corresponding member, that looks like I
have got to call somebody up and say, hey, I want to be a corresponding member
of that. I think the onus should also be the other way, that some of it should
be about how when issues are taken up, there should be communication across the
committee so that peoples’ expertise can be appropriately brought into play.
We should make every effort to have issues be addressed by the
cross-cutting expertise of people on the full committee as early as possible in
the process.
So that was our yesterday conversation. This strikes me as our day before
yesterday conversation.
DR. FITZMAURICE: I am looking at the title, committee operations and
process. Yet down in that same sentence we are talking about all processes are
transparent. So either we have one process or many processes.
I might suggest, it could be fixed by either making processes up in the
title, or simply saying committee operations and let processes be referred to
in the next sentence.
MS. GREENBERG: Yes, I agree. I think as always Leslie is very perceptive.
Also, I think this document has taken on a life of its own beyond what its
original purpose was. It is a living document, and I am starting to wonder,
will we ever and should we ever finalize it. Even if we did finalize it, of
course, we would always update it or change it.
I think initially the purpose of the document was to clarify how to
subcommittees are expected to operate so we had some basic rules of the road
down. Since most of the work goes on in subcommittees and like any small group,
they take on their own personality of their members and their chairs, et
cetera, to just try to encourage some consistent rules across all the
subcommittees in the way items were processed.
Now, I think many of the things that were said today addressed broader
issues of how the committee functions, and strategic planning in the committee,
setting priorities, all of that, which are much broader than what was the
initial purpose of this document. But I can see that standing alone, this
document could seem to be kind of limited.
On the other hand, at this time when we are about to go through a major
transition, whether it makes sense to come up with a strategic plan for the
committee might be questionable or even go too far in that direction.
So I think I can beef it up a bit or modify it a bit to try to capture some
of what is here, but I don’t think we want like a ten-page document, either. I
tried to keep it to a page. I think it will have to be two pages or a page and
a half to get some of this in. But I am happy to do whatever you would like me
to do, but I need some guidance at this point, because we have gone beyond what
the original purpose was.
DR. COHN: I think we have people who are willing to give you advice.
MR. REYNOLDS: Thinking back to orientation, I think what is really
important is, with the change coming up, this is a great orientation document
on the structure of how we work.
MS. GREENBERG: That is what I am thinking. We should have something.
MR. REYNOLDS: The other things are about how we all work together and what
we are working on, but I think this will allow somebody to understand how they
can play as a new member, and how they can understand what it was and what they
might be able to do. But obviously anybody that is still on the committee still
needs to reach out to them.
I think if nothing else we need to pass this today or accept this today for
that purpose, because that will be needed soon.
MS. GREENBERG: If you can accept it today with the agreement that it would
be finalized with the executive subcommittee, then I think that would be fine,
unless you want to make it into a much bigger document, which I wouldn’t really
recommend.
DR. MC CALL: I wanted to add something to Leslie’s comments. It is about
how this document is going to be used, and the timing is very good.
Most of what is in here are promises of a work group or a subcommittee to a
member. The comments from John Houston are about the promises of the member to
the subcommittee. This is about roles, which are really about promises to each
other.
So I would add a few things in there, one being the specific that John
mentioned about the expectations of a member to a subcommittee, in terms of
staying engaged, but timeliness of response.
MS. GREENBERG: I think he was actually talking about members of the full
committee as well.
MR. HOUSTON: It really is both, because when it is internal to the
subcommittee, members need to be responsive. When we get documents that need
action, I think members of the committee then need to be responsive, whatever
the context.
DR. MC CALL: So that responsiveness can include maybe being a reviewer,
remaining engaged when materials come out for preview.
DR. STEINWACHS: Marjorie was mentioning my strategic plan here. I had the
strategic idea of a 100-page document, but I boiled it back down to one page.
Maybe on that second bullet, Marjorie, you could just add something for the
executive subcommittee review that reflects on the fact that the committee may
use strategic plans and other mechanisms for setting priorities and making
decisions or something.
MS. GREENBERG: There is something kind of missing there.
DR. STEINWACHS: That was all I was looking for, because we do do sometimes,
strategic planning. It wasn’t mentioned here in the committee processes,
either.
MS. GREENBERG: Yes, I think there is something missing here.
DR. COHN: I think we have had a lot of input. I think certainly just to try
to put things together, this clearly is a living document. If we pass it today
as an advisory, it is going to be modified from time to time. But I do think it
is good as a briefing and something to go back to.
I think we have heard — which I almost think is preamble — which is
really the responsibilities of the member to be engaged. I don’t think it is a
bullet. I think it is in that first couple of sentences at the beginning, which
is really where it fits.
I am also hearing — and I am using the wrong words, but there is at the
judgment of either the chair or the chair of the work group or subcommittee,
when things reach — and I have used the term complexity, and that is not
really what I mean, values laden, 12 pages long, we describe however much it
is. There is a difference between the Standards and Security two-page letter
that people don’t need to have large conference calls about, and a 12-page
privacy letter or a 50-page report or whatever, but it is a judgment issue.
There is a place at which point it becomes pretty clear to people or should
become clear that we need to start reaching out and pulling people very early
on to make sure that everyone gets our input. We just need to reflect that,
though I don’t know that that needs to be laid out and exactly how we do that
in here, but it is really more that there are judgment issues there.
Now, having said all of that, does that reflect a lot of your ideas? And do
people just want to pass this, with the idea that the executive subcommittee
will deal with the final modifications to be brought up regularly to the full
committee to reflect on and for revisions? Okay, does somebody want to make a
motion to that effect?
(Whereupon, the motion was moved and seconded.)
MS. GREENBERG: Let me just say I like some of what has been added today. I
think both from what Carol and Leslie said, this really is about now only roles
but responsibilities in some ways, too. There isn’t anywhere that it is written
down that for example every member is expected to serve on at least one
subcommittee. That I think could be incorporated into this first preamble. So
as the committee does its work groups and subcommittees, whatever, every member
is expected to serve on at least one. We also want to make sure that everybody
has the opportunity to participate and the processes are transparent.
So I think we could pick up both sides of that here, and that would make it
a nice document for the next class also. So I like that. If that is consistent
with what I heard, I’ll do that.
DR. COHN: I think that sounds consistent with the motion we are making. So
I have heard a —
DR. STEINWACHS: Moved and second.
DR. COHN: — moved and second. Any further discussion? All in favor, Aye?
(Chorus of Ayes.)
DR. COHN: Opposed? Abstentions? Okay, thank you. One more item off of our
table here in terms of action.
Agenda Item: Subcommittee Sessions
DR. COHN: What we want to do now is talk and reflect on what is going on in
the subcommittees and work groups in terms of both what happened at the
meetings and action items. Is Don okay to start with, Justine? Justine
commented yesterday, she thought we should make it griddable, all the
activities being planned, and it may cause us to take a step back. We aren’t
obviously going to adjudicate everything today; we only have about 40 minutes.
Don, do you want to talk about some of the plans for the Population
Subcommittee?
Agenda Item: Subcommittee on Populations
DR. STEINWACHS: I am happy to do so. Justine increased the pressure on the
Populations Subcommittee, and Harry too, when they put out that slide of the
productivity. There is privacy and security, we are pumping out those letters.
We are sitting back thinking too much, so there is too much thinking, too
little production. So we are going to get the production line here, and then
you can tell us to cut back.
Just to provide you with the full set of updates, as I have reflected
before, we did a letter on data linkages. We want to come back to the data
linkage issue either late 2008, early 2009, with the idea of every couple of
years having a half day or a day full of hearings and trying to measure, are we
making progress in enhancing the use of data that is already in the government,
are there issues arising that we want to be addressable by us or the Secretary.
So that kind of monitoring strategy is what we decided, and that is in the
longer range books, but not too far away.
We had considerable discussion of the medical home hearings. We have set
aside a day and a half prior to the next NCVHS meeting. That would be the 19th
and 20 of May. We went over a rough agenda and a statement of goals and
objectives, and trying to get at issues of how is the medical home being
measured, is there consensus around measuring. A lot of the measures are really
structural, so there is also the question, what does the medical home actually
do, and will we have data when you look at available sources, NCHS, maps, other
sources.
Ultimately there is the outcome issue, will it be observable over time as
medical homes come into play, that there is a difference in outcomes. There are
certainly a lot of parties out there who are active in this, NCQA, NQF, the
clinical societies, so the hearings would be an opportunity to bring them
together, with the perspective of our capacity as a country to measure this, to
monitor it.
That also fits in — and Bill Scanlon is not here — to issues that came
out of some discussion at the executive subcommittee about health reform and
the anticipation that in 2009 there will be things coming on the agenda, and
what is the information that we have available through health statistics and
federal and state resources that helps people plan and adjust and titrate what
health reform might be.
So the medical home fits as one piece of that, but Bill challenged Simon to
think bigger and harder. I remember he was the czar of reform data, for data
reform.
DR. COHN: With him not being here, the job gets bigger and bigger.
DR. STEINWACHS: Larry, do you want to add anything on the medical home?
DR. COHN: Can I just ask, we have A’s, Q’s and M’s. What does that stand
for? It might help us understand better, since Don, you are talking to this.
DR. CARR: What I am trying to do here is saying, what is populations doing,
and think about are there any other committees that would be related to it, and
who is the customer and do we need to broaden the scope.
So the A’s are active, we have a hearing planned. The Q is quiescent; every
two years we would take another look at that.
DR. STEINWACHS: Or periodic.
MS. GREENBERG: P is better.
DR. STEINWACHS: We never want to be quiescent.
DR. CARR: No. But data linkage, is that something that links to privacy or
something that we would be putting there as well.
Then health reform data, is that an active thing or a planning thing?
DR. STEINWACHS: I’d put it more in the planning.
DR. CARR: P is for planning.
DR. STEINWACHS: Yes. Jim Scanlon also assured us that we need to think
realistically about what we can do.
DR. CARR: Did you mention surge capacity?
DR. STEINWACHS: I was moving on to surge capacity. Ready for surge?
DR. CARR: Yes.
DR. STEINWACHS: As I described already, the Population Subcommittee is
trying to surge on productivity.
We had a considerable discussion. Doug Benning, who works in Jim’s office,
has been working with our subcommittee, and particularly with Bill Scanlon and
Kevin Vigilante on trying to pull together what was learned at hearings that
were held 53 weeks ago. Bill assured us that this would have been more timely,
but there was a secondary data activity or something that intervened and slowed
down this process.
But we got into a very productive discussion about, who is really the
customer for what we learned. The hearing was broad, but there were a number of
principles and issues that came up.
So the suggestion, which I thought was excellent, there is now in the DHHS
an Assistant Secretary for Preparedness, is that the right title, Jim?
MR. SCANLON: Preparedness and Response.
DR. STEINWACHS: Preparedness and Response, and who has an office of about
500 people. They are beginning to tackle the data issues. They have contracted
with AHRQ, we understand, to work with them over the next couple of years on
this.
So we were coming back with a suggestion that what might be appropriate in
drafting a letter which we are working on is that the person who received the
latter might be the Assistant Secretary, most appropriately, that this doesn’t
have the kind of policy issues. We can see right now that — this is open for
discussion certainly, but maybe the idea that maybe the real customer or
recipient for this would be the Assistant Secretary.
I thought that was very useful, because that also helped to set the tone of
the letter in a way in which we have something to offer. I’m not so certain we
have something to offer that is worthy of the Secretary’s attention.
I’m not sure that we are going to do anything more on surge preparedness. I
think that is a question mark. We are trying to wrap this up. There could be,
I’m not setting that aside. Certainly it might be after AHRQ and others do work
on the data needs for preparedness. It might be something for NCVHS and for the
Population Subcommittee to come back to, and that probably will be a couple of
years or so away.
We talked about the 21st century report about the 60th anniversary
celebration for this committee in 2009, the confluence. This is like the stars
or the moon in Venus or something coming together at the appropriate times.
We are going to start some telephone calls to start talking about a
specific proposal for what we could do on that. It is two parts. One, which I
think is simpler, but that may just be because I don’t fully understand it, is
to be able to provide a report on progress on the recommendations on the 21st
century report that were made six, seven years ago. You said it came out during
the last anniversary.
MS. GREENBERG: There was an interim report at the 50th anniversary.
DR. STEINWACHS: So it is about five, six years since the actual report
itself. One was to provide a progress report, how well are we doing. That may
need some hearings, too, in order to try to understand how well we are doing
and are there other issues or problems that are arising in achieving that
The second, which I thought was very important and something that I would
hope more broadly involves members of the committee, is an effort to update the
vision, to the extent it needs that.
A ky missing piece in my mind is the role of electronic health records, of
personal health records, that are not part of the discussion there or the
vision. So whether that needs to be a small piece or a big piece, but it seems
to me that it needs to reflect progress and it needs to reflect what this
committee and the country is about.
So that could benefit from input around the table. This vision is in a
sense our vision, when you get down to it. So you might want to think about
some processes that might separate the updating somewhat as a function from
what we ought to undertake more collectively, and looking at the vision and
saying what is it that needs to be modified, and to what extent do the
recommendations also need to be changed and modified that reflect the fact that
we are now talking about a different vision for the world ten years from now
maybe than we were five years ago or so.
So that tells you where we are. I was trying to think if there was anything
I missed. Marjorie, is there anything I missed?
MS. GREENBERG: I don’t think that you missed, but I just wanted to make a
few comments. One is that I do think the medical home issue could be quite
cross cutting, could certainly engage folks on the standards subcommittee. Part
of the way you would collect this information might be in the HIPAA
transactions as well as in surreys, et cetera.
That hearing is going to be immediately receiving the May full committee
meeting, so I think anyone who isn’t on the Population Subcommittee or Quality
Work Group but is interested certainly would be welcome to come. I would
encourage some representation from the Standards Subcommittee. Justine, are you
still on Standards in your free time?
DR. CARR: I think so.
MS. GREENBERG: We also talked a bit yesterday, and I was talking a little
bit just this morning before we started about how a number of the topics as
well as topics that came out of the data uses report could have some synergy
with this whole 21st century vision for health statistics. Larry and I were
engaging on how the medical home discussion investigation can relate to that.
Also, I think Justine and I were discussing the issue of de-identified
data, is very much related to the whole health statistics agenda. That is one
of the reasons you use de-identified data, to produce some kind of statistics,
as opposed to address an individual patient or something. And data stewardship.
There are data stewardship principles, it wasn’t called that, but throughout
the 21st century vision report.
So I think it could be kind of a hub or a coordinating mechanism for a
number of the issues that are floating around there on that chart. So that is
just something to consider. We were talking at the end of the day yesterday
about having a lot of different little things and not an integrating mechanism
or way to see how they all interrelate and to coordinate them. I might be
making too much of it, but it could be a way to do that.
Depending upon when we have — we are not really discussing it this
morning, but it was discussed in the executive subcommittee — when we have
this 60th anniversary celebration, I would say the earliest would be June 2009,
the latest would be June 2010, which would be the year after, but that is what
we did for the 50th.
We don’t have to have something completely wrapped up with a bow. The wy we
used those two interim reports at the 50th was to get input as well, and bring
in some kind of key thinkers and former chairs and all that.
So I think we do need to at least have a recommendation for the next chair
and the next version of the committee as to what our target date is for the
60th, because we are going to have the National Academy and all that on that.
So I think that should be a goal for May.
DR. COHN: We talked yesterday about having customers and deadlines. So
maybe this all will get the moons and the stars aligning.
DR. STEINWACHS: These are very important events to create deadlines.
DR. COHN: Well, yes. There is nothing much worse than having a big meeting
and not being prepared. So it does help encourage people meeting deadlines and
all of that.
DR. STEINWACHS: I think deadlines are useful, partly because it helps — if
you want to think about total committee resource allocation, it puts it in much
sharper perspective than if you just say we have these things moving, and they
can move at any speed. That depends on the resources you can put in.
DR. COHN: I think that Marjorie has done us a service by thinking more
globally about all of these things and putting them together. I do think there
is a fair amount of synergy between the various items you have discussed. It
does seem like there should be some comprehensive planning around what we are
talking about here.
You are talking on the one hand about evaluating current, doing vision. Now
we have thrown de-identified data and medical home as part of the whole thing.
MS. GREENBERG: And data stewardship.
DR. COHN: And data stewardship.
MS. GREENBERG: They do relate.
DR. COHN: They do. And an updated vision and coming out with a strong view
of the future on health statistics and vital statistics is very close to our
court here, so it makes sense.
DR. STEINWACHS: Some of these things are very timely too when you think
about the challenges in the vital statistics.
DR. COHN: As well as data for health reform.
DR. FRANCIS: Maybe it is a question that was implicit in what you just
said, Simon. I had understood, and I may be wrong about this, that part of why
people were interested in a medical home and then de-identified data out of
that was to have the cost and quality data that would be — this is from the
Quality Work Group — that would be critical across populations if people want
to get into health system reform in a big-time way, which I take it is going to
be on anyone’s agenda after the next election, whatever happens in the next
election.
So my question is, I take it part of the reason there is an A by medical
home is that there ought to be an A by health reform data in both Populations
in Quality. Maybe I’m wrong about that, but that is what I took it to be the
linkage between those two meetings, having only gotten to one.
As somebody who is on the privacy subcommittee, I would just make the
comment that I think one of the enormously interesting challenges is going to
be how to do that with our appropriate thinking about privacy in the mix,
because that could scuttle — if that doesn’t get into the mix, that could
scuttle important initiatives.
DR. GREEN: I want to grab the moment to unite what Marjorie was just saying
with an anticipation of what Dr. Chute may be telling us here in a few minutes.
It seems to me it would behoove the committee, particularly with this huge
change in membership that is coming. I have been grieving, anticipatory grief
because of the maturity, the longitudinality, the seasoning that exists here. I
decided last night it was time to start celebrating it, in this perspective.
I think Marjorie’s thinking — we really need to listen hard to what she is
saying. She has got this incredible perspective on what this committee has been
about for a long time. What she is really telling us, I believe, is that there
is the normal change, there is the normal management of data sets, taxonomies
and classifications. But every now and then in history there is more of a
revolutionary moment, where the old ways of doing things need to collapse and
new ways have to be born.
It may be good timing that we are refreshing the committee, we are getting
a new group in here. Perhaps I am just dead wrong on this, but I don’t think we
are talking about refining old systems. We are at the point of redesigning the
data and information architectures that underlay the largest business in the
United States.
This is not small time stuff. This is not getting from version seven to
version eight. This is a remake.
The medical home idea’s importance in my view is not intrinsic; it is
extrinsic. The medical homes getting its legs here in Washington is a signal,
not noise, it is a signal that people recognize that the redesign of the system
is now joined and underway.
So the data systems are going to have to be realigned with the redesigning
system. The informatics competencies and genetics competencies are not coming,
they are here. Now they are just going to be further elaborate.
So this grid that you are making up there, I’m not sure that the columns
will necessarily be the columns that we would want. That 60th anniversary
meeting could very well be a great moment of consolidation and launching.
I would like to see the new committee think about devoting considerable
time to this idea because it is such a propitious opportunity.
DR. COHN: Larry, I’m glad you are going to be a part of the activity.
DR. OVERHAGE: Ditto that. Ditto what Larry said.
DR. TANG: Could I add one comment?
DR. COHN: Sure, Paul.
DR. TANG: Larry spoke so eloquently, I can’t resist trying to put up
another tag line for it. I can’t see any of the A’s or M’s or P’s or what is
going on up there, but in some sense maybe we are preparing the landscape for
data driven reform.
So it is a little bit of a twist on what was said before about the data and
health reform, but I think Larry is talking about how do can we redesign the
system based on our accumulated knowledge and experience of what doesn’t work.
So maybe we need more of a data driven reform.
DR. COHN: Paul, thank you. I think what I am hearing here, and I recognize
we are not going to do the strategic planning on all of this at this session
today, but I think we have created a very nice context and envelope to put this
in.
I think I will charge the executive subcommittee along with Marjorie and
Jim to come back in May to come up with a comprehensive proposal about all of
this stuff and how it works. I think, Don, you are right. There are a variety
of things. Some of them require probably more than just Population Subcommittee
membership. Others might be fine. It is one of those issues of thinking
through. As Harry commented yesterday, some of these things just go right
across the entire membership, and we need to get everybody’s knowledge and
input on all this.
So thank you. Marjorie, I want to thank you for handling four or five of
those items and consolidating them into an overall banner, which I think is
very helpful.
So with that, we are just giving more and more work to the executive
subcommittee, but it was good.
Justine, do you want to talk about quality?
Agenda Item: Quality Work Group
DR. CARR: Sure. This is helping me. Paul, I’m sorry you can see it, but you
would love it, I’m sure. The columns are the current committees. I think the
point is just to say, does it make sense to try to line up the committees. I
think what we are seeing is that we may be looking at things differently.
In terms of quality, two things. Yesterday’s work group looked at the 2004
Quality Work Group report. It had 23 candidate recommendations. I would say
three of them are done and 20 are a work in progress. These recommendations
span the spectrum of the committee from Populations to Standards to Quality to
Privacy, all of the above.
We did a first walkthrough yesterday to see where we are. We talked about
the fact that updating that would best be done in small pieces so that it is
not so overwhelming. But there is more work that needs to be done, and we got a
lot of great input on it yesterday. So again, we are at another one of these
points of looking back. In many ways, as we update the 21st century report,
this maybe is a part of it as well.
In terms of other topics that have come through Quality, we talked at the
executive subcommittee about the issue of quality versus research that came up
during the uses of data report, and more recently has come up in the public
press. Our sense for now is that that is in the Q, quiescent zone that we are
going to hold on that because there is other activity going on and we will have
to see whether there is a need to do that.
Other issues that came up from the uses of report are data stewardship and
de-identified data, as we have talked about. These are not just Quality, they
go across. So I am happy to hear them incorporated, as Marjorie suggested, into
this top couple of rows that are the 21st century update, data stewardship,
de-identified data, medical home, health reform data. We are sequestering all
of those.
When we say data stewardship and de-identified data, in the report we
talked about holding hearings on de-identified data. Is this work that we would
not initiate in this interim period? Is it something that is for a later time?
DR. COHN: Are you asking me, Justine?
DR. CARR: Sure.
DR. COHN: What I am thinking I am hearing as the next step, given that all
of this fit into an overall context that we need to come up with a plan about
how we address them in context as oppose to out of context.
So I guess the answer is, we need to come up with a plan about exactly what
we are doing on all of these, that doesn’t mean that everybody is spending the
next year solidly in hearings. I think that becomes part of the context.
DR. CARR: The other thing that we had talked about at the executive retreat
was the customer. That is why I made one of these columns, who is the customer,
because I think as we do this work, we ought to know who it is, and that would
presumably inform our time line as well. So feel free to jump in on customers
for some of these.
DR. COHN: Justine, are you done with the Quality report now?
DR. CARR: Yes. I think the topics that the Quality Work Group has discussed
are up on this list, and unless anybody else wanted to add anything — but I
think our main activity now will be working toward the medical home hearings in
May, and the second would be updating this report.
MS. GREENBERG: I just wanted to mention in regards to the 2004 report —
and it really fits into this broader discussion we have been having — Carol
made a very — particularly since it was almost six o’clock, I was unable to be
perceptive about anything, but she made a very perceptive comment that all
those items in the 2004 report.
First of all, progress has made on them, as Justine said, on a number of
them, but many of them still haven’t been resolved or are still issues when
trying to use data for quality or other purposes. In a way, they didn’t have a
home. But there was a meta issue of — you can say it better than I could, but
who was minding the shop. There was no vision or there was no structure for
addressing all these.
Do you want to say anything more about that? I thought that really nailed
what the issue was with that report and why it hadn’t just moved on or
whatever.
DR. MC CALL: I’ll describe it and also, as you pointed out, it leads into
what we will be talking about later this morning, I think.
The report laid out the things from 2004 and said this is what we said we
were going to do. There were all examples of two or three different kinds of
things. It was either create a mechanism for, and maybe it was capturing a new
piece of data or a new metric or modify a mechanism for how people code this or
modify certain diagnoses, or how do you get functional status.
But every single one of them was an example of nothing to do with quality
per se, but what I had called at the time building blocks of quality. So the
metaphor was, if quality were a paragraph or a sentence, what we were building
were the words or sometimes the individual alphabet or letters, that then could
once created, you could begin the task of writing something. So this was very
foundational work.
So that was one thing. So I thought, that was then. When you think about
what Larry said, and he is right, we will face unprecedented change in terms of
the magnitude and the speed and the diversity of data that we are going to see.
So when you think about where is the home, where is the overall process
whereby these things get created all the time, that seems to be the meta-issue.
We as a group certainly can’t do that. We lack the subject matter, the
expertise, and even if we had it, we lack the raw capacity. It needs to be a
much more living, breathing, dynamic thing.
So that seems to be the meta-issue, but I also think it is actually was is
in these materials here today. I think to the list that Justine is putting up
there, those seem to be the meta-issues. There is an early version in the 21st
century document. It doesn’t talk about some of them quite at that length, but
the seeds are there. It links to data stewardship and health reform data,
although Paul had a different phrase which was better, and the whole
de-identified. So this is going to be the century of getting data under
control.
DR. TANG: Can I piggyback on whenever my turn is?
DR. COHN: Paul, you are on next, and then I want to move to our next
subcommittee.
DR. TANG: I very much want to piggyback on what Carol did on top of Larry.
I think this is what you are proposing, Simon, but I maybe will make the
proposal that we look for an ad hoc or unifying project.
It seems like what I am virtually seeing of what Justine put up is, what is
called for is to establish a data foundation for health care reform. Carol
talked about letters, Larry talked about the vision, but I think we need to put
it all together. I think that is what you were saying, Simon.
But I think if we are ready to kick off a project, and we would love to
have this thing done by the time the new Administration takes over, it would be
the way I had envisioned it. We do need a home at least to come up with a
framework for thinking about it, if it isn’t actually an entity for carrying
out the actions. But AHEC 2.0 or whatever might come about could help move that
along. But I think we are quite uniquely positioned to try to bring all these
elements together, but not as individual pieces, but rather as something that
establishes the foundation for the act of transforming the delivery system in
the way that we create new evidence for care.
I guess I am trying to put on the table what about going ahead and kicking
off such a project, or at least starting the process to get that done. If we
wait for May, then three more months have gone by.
DR. COHN: Jim, I think I will let you respond to that one, and then I will
jump in. Paul, thank you.
MR. SCANLON: Well, I think I made these same points to the executive
subcommittee, and I think it is worth discussing today. We are in for — not
only are we going to have almost half the committee turn over in May, but we
are also going to have an election in November, someone told me. So I think
this is probably — the timing right now is probably not the best to start some
kicking off of new initiatives for which there won’t be an audience.
With health reform you have to be very careful. There are various flavors
of health reform. Some of them are state based. Some of you come from states
where this has been tried, others are more national based. You choose the wrong
way and you are irrelevant. A lot of other folks are interested in these areas.
So the timing here is such, it is a sensitive sort of timing. So are ideas
about integrating a lot of these things. I think a year from now probably will
be a better time to do the integration piece. To the extent that the landscape
changes on health and health policy, it would be obvious where the gaps are.
So at any rate, I guess my advice to you is, for the remainder of this
year, it really should be consolidating previous work, continuing work when you
promised you would do it, and thinking about what this framework might look
like, not initiating it. Otherwise you are falling in the valley of death where
the previous Administration can’t use it and the new Administration will
disavow it because it was something that they didn’t start.
I think you might want to refer to information and data for policy rather
than reform. I think you are almost sentencing yourself to a bad experience if
you start it that way. Improving information and data for health policy, we
always need that. It could be at the statistical level, it could also be at the
international health record and so on.
So that would be the way I think I would advise you. You certainly could
start with some exploratory looks at these, but don’t declare a framework or a
big initiative. This is the valley at the moment, and it is just not a good
time in terms of audiences.
DR. COHN: Gene, would you like to make a final comment on this one, wrap it
up, and then we will figure out what we are going to do with the remaining
couple of minutes.
DR. STEUERLE: When I came on this committee several years ago now, I was
the only economist then. I was asked to look at trying to figure out how
incentives affected all these results. I guess I am no longer the only
economist, but I am still struggling with that issue.
I see the issue just permeating every discussion we often have, where we
see sometimes a way to move forward, almost a way to move forward, but we often
see that there is lacking an incentive structure to make something happen.
I still am never quite sure how to integrate that into these documents and
into what we do, but I often see — sometimes it is a lack of incentives within
the bureaucracy, sometimes it is a lack of incentives in the payment structure
for the way we pay for Medicare or Medicaid. Sometimes it may be a lack of
incentive structure just to make sure that there is an adequate amount of money
put aside for research and for new types of research.
We struggled in our committee in Populations with an incentive structure
for somebody to care about the data linkage. We had people from every major
agency around town, from IRS to Social Security to HCFA and everybody else,
claiming that there was some really good things they could do with data
linkage, but we couldn’t quite figure out how to break through what might be a
privacy issue or a lawyer somewhere who decided they weren’t quite sure how
they could break through how to exchange data with other committees.
I don’t know where to go with this, but I do see this issue as still —
maybe it is because I am an economist — as still driving a lot of what we are
talking about. We get close to where we think the next step can be made.
It is a question of what is the incentive structure in place to do — if
you want to cut the Gordian knot between for instance the loss of data that is
causing harm to our population because we can’t get it, and the concern over
privacy, because if we gather more of that data we might actually violate
someone’s privacy; what is the incentive within the bureaucracy to make sure
these data linkages happen, not just that we identify good data linkages, but
that somebody has an incentive for it to happen. What is the incentive for
people to start reporting their data when they file with Medicare in a way that
insures — for instance, a simple one. Medicare doesn’t have a differential
charge for data for simple things like lab tests that are electronically
reported. We can debate whether we adopted the perfect system, which we will
never have, but we could have an incentive for people to do this.
So I am just struggling for how to make sure that these incentive issues
don’t get lost, because it seems to me that is always the point at which we get
stopped. I’m not answering the question, but —
DR. MC CALL: Well, not as an answer, but just as an idea, up there on the
list is — where did it go? It was health reform data, but it got called
something else. Health policy data.
MS. GREENBERG: Data for health policy.
DR. MC CALL: But I would add something up there. There is going to be a
unique window during reform where there can be carrots and sticks put in place
that says, A, you have to do certain things to play, number one, and if you
don’t do it you can’t play. If you fail to do it, you’re out. Or if you do it
kinda-sorta, you pay a penalty.
I don’t know what the right answer it specifically, but there is a window
that I believe is going to open up. When it opens up, I think there needs to be
a point of view about what the opportunities are to kick this into high gear
around incentives, to put that in there.
DR. COHN: We are beginning to run out of time. I would ask — we are due to
take a break in a couple of minutes. Maybe I could ask Mark to talk about
privacy, and congratulate him on getting a letter passed yesterday.
Agenda Item: Subcommittee on Privacy and
Confidentiality
MR. ROTHSTEIN: This will be very brief, in fact. At our subcommittee
meeting yesterday, we decided to have a conference call to take up the issue of
what our subcommittee’s agenda will be for the coming few months. We didn’t get
to that in any detail. We spent most of our time polishing the centennial
letter of the NCVHS.
DR. CARR: Do you have a privacy letter that you are developing now?
MR. ROTHSTEIN: No. The one that we passed yesterday was the only thing that
we were working on.
DR. COHN: Harry and Jeff, do you want to talk briefly about Standards and
Security?
Agenda Item: Subcommittee on Standards and
Security
MR. REYNOLDS: Yes, just a couple of quick things. Four of the six people on
the committee are at a term point, so we did a little planning so we could pass
on a legacy to whatever the next committee makeup is.
If you notice, ever since it came out we had the most letters. Justine,
first she got us yesterday, then she made everything up there mundane. That is
what the M stands for. We will probably for May be bringing back to small
letters, to be the first two letters. We have got to be the first two letters
of the next 100. But we are working through those right now on NCPPD and a few
other things, so you may be seeing those. But those will be more maintenance
type letters, not moving the ball forward based on a lot of the discussion that
you have heard, other than allowing things that are going on how to continue to
grow the right way.
So we have prepared a document which we will pass on to everyone about what
are the hot issues that we always have to keep an eye on because of what we do
with some of the key standards, and then what are the subjects to develop. We
went through that, made a list. We have a lot of those mapped into some of the
things up there. So that document is ready and will be passed through the
executive subcommittee on to whatever goes forward. So we don’t want to take on
any new subjects or act like we do anything new with such a change coming up in
our group.
MS. GREENBERG: You are talking about the matrix?
MR. REYNOLDS: Yes.
MS. GREENBERG: If it had like a one-page description too, to introduce it,
I think that would help.
DR. COHN: I’ll just comment about NHII, but we talked about yesterday about
reinvigorating. I think we would all observe that it has been the style of at
least this current chair that a lot of the work that is normally NHII has been
handled by ad hoc work groups. So as we look at what NHII has done over the
last while, I think functional requirements for the NHIN as well as secondary
use cases are basically — once again, it is a style issue. I’m sure in future
we will consider exactly how all that works as the next chair or the next group
comes on.
Now, I want to thank all of you. I think this has been a very interesting
conversation. Paul, I do reminded — being an emergency physician, I do
appreciate your sense of time. I am however also reminded by Jim that timing
may be everything, or may be 50 percent of all of this stuff.
So the good news is that we have the medical home hearings coming up. I
think out of this conversation, rather than us just talking about 21st century
vision, we are actually going to start moving forward and trying to figure out
an overall project plan, how it should all come together. I think Jim’s comment
is really not one that we don’t work on it, because everything that we do takes
12 to 18 months to reach fruition, except in truly exceptional circumstances.
But it is preparatory, and it is beginning to put it in place. If we don’t have
80 percent of the work done in the next six to eight months, 12 months or
whatever, getting from there to the final report will be impossible in a
reasonable time line.
So hopefully this will come together. We will be talking to all of you
about the time lines and pieces. Then I think we will begin to do some
planning, but it will be something that you will all receive and have an
opportunity to provide comment on way in advance of the May meeting, I think.
So thank you all.
It is now 9:56. We are running a couple of minutes late, but pretty much on
time. Why don’t we take a ten-minute break, and then we will come back to our
final session, talking about terminologies.
(Brief recess.)
Agenda Item: ICD 10/11 – Update
DR. COHN: Why don’t we get started again? Would everyone please be seated.
We are very pleased to have as a final session conversations about
terminology and classifications and a national and international update. I
mentioned earlier, we are starting off with Dr. Chris Chute, an old friend and
colleague from the Mayo Clinic. Chris is Professor of Biomedical Informatics
and Chair of their department. Chris has over the last 15 or 20 years been one
of the thought leaders in the country and internationally in relationship to
terminologies, classifications, what in the heck we are going to do with all of
this stuff. I am also reminded of such terms as shallow information models and
other things like this from yet another time.
I guess I should publicly disclose that back in the ’90s, Chris and I
co-wrote a number of articles in this area. I am very excited to have his
leadership into this ICD-11 effort. I am hoping that he will update us on that
as well as what is going on
Now, Chris will be followed by Linda Kloss from AHIMA and Keith Campbell
representing AMIA, talking about other terminology issues and conversations.
Chris, once we are done with the first conversation, we would love to have you
sit up there as part of the conversation as we dwell more on conversations and
next steps.
DR. CHUTE: Thanks very much, Dr. Cohn. It is my privilege to present to
you.
Let me just review for you some of the use cases for which ICD-11 is
intended. This is not exactly a surprise or a shock to you. Morbidity and
mortality of course is a longstanding and traditional use case for public
health. However, it is being broadened to include this notion of a scientific
consensus of a high level phenotype.
It is very clear that the historical notion of the ICDs as a, to put it
carelessly, a haphazard collection of terms and concepts into a more systematic
collection of terms and concepts is very high on our agenda. We anticipate of
course that it will continue to be used for patient characterization in
clinical research as well as primary care.
The goals then are to evolve — now, this is jargon for you — an
ontologically coherent classification. Please don’t ask me what that means.
MS. GREENBERG: You don’t know.
DR. CHUTE: I don’t know, either. Simplistically it means we will use the
existing technology, the classifiers that have been evolved in the computer
science community, description logics, all the machinery that Keith actually
knows more deeply than I do, he did his thesis on it, to insure the inherent
consistency and completeness of this classification system. It will be authored
using modern tools and environment, essentially.
Importantly, and I will expand on this, it will be linked to underpinning
terminologies such as SNOMED and the gene ontology to make more explicit the
relationships between these higher level rubrics and classifications with
underpinning clinical and medical concepts and genomic concepts, I might add.
So the rubrics within them, and Simon, you will recognize this term, you
were a co-conspirator in a proposal we had put forward many years ago, of
defining the underlying classification categories in terms of aggregation
logics. That is a notion I will expand upon, but simplistically it means that
these terminology elements and components will define logically what the
classifications mean and what the rubrics actually mean.
Those of you that are familiar with the ICDs recognize that there are no
human language definitions associated with the terms or concepts. It is just
the surface forms and perhaps its implicit meanings from a limited hierarchy.
This is intended to be addressed not only in English, but in other natural
languages, and how we hope to achieve that is something I will share with you.
Clearly we are in the genomic era, as I will expand upon, and we need to
maintain longitudinal consistency, given that the heritage and history and
dependency of many sectors of public health and vital statistics are dependent
upon a longitudinal sameness, if you will. We must be respectful of that. While
having an elegant fashioning of the core contents of ICD-11, we need to define
a mechanism to insure this longitudinal consistency.
This is a walk-on slide. This is not in your packet, but Marjorie reminded
me, to point out that there is a larger structure associated with this
activity. This is not a back yard project. It is directly a child of the
revisions of WHO with strong collaboration and cross membership of the family
of international classifications at WHO, including their update and revision
committee. There is associated with it a formal modeling and informatics
committee chaired by Mark Musand at Stanford, and topic advisory groups which
are made up of international experts drawn from countries around the world who
in turn have corresponding work groups to create a community that defines the
core of the revision process, but not the entirely, as I hope will become
clear.
Many of you have seen this slide before. It is one of my handy back-pocket
notions of familiar points along the continuum. Simplistically we all recognize
that classifications can be adopted into groups using grouper software. The
premise here is that an aggregation logic can do the corresponding task from
underlying nomenclature or terminology concepts to achieve higher level
classifications algorithmically. In a word, that is what aggregation logic
really is.
This slide debuted I believe in the Standards Subcommittee of NCVHS in
August of 2002. Many of you were there. It is the premise that if we have this
collection of patient events, findings and interventions, these are elements
within a record, we can aggregate them into reimbursement and management codes
which I would argue is the current primary use case of ICD in this country. But
we recognize that that aggregation may be suboptimal for decision support or
error detection or public health surveillance even. If we do this
algorithmically, once we define an aggregation mechanism for one use case, it
becomes relatively straightforward to define corresponding algorithms for other
use cases and other high level classifications, and that is the premise behind
aggregation logics.
We must recognize that we are in a fundamentally transformative notion of
defining diseases and even signs and symptoms. The genomic transformation is
something that I need not remind you about, but it has huge impact on our
understanding of classification.
Many of you know, I babble on in my lectures about Linnaeus not quite
getting health care right because he didn’t understand the germ theory of
disease, and hence he classified rabies as a psychiatric disorder, which
arguably at some level, given that one goes mad tragically in the final phases,
there is some pertinence to that. But it is clearly a viral disease, and the
fact that Linnaeus did not recognize that is a significant flaw in his Generum
Aborum, which I’m sure you have all read, of 1768.
Our understanding of health care today is highly syndromic. If you look at
something like congestive heart failure, which is essentially a final common
pathway of multiple diseases, and we fail to recognize the genomic
underpinnings and manifestation and etiology of many diseases, this is starting
to change. But as it changes, it will transform the way we understand and
classify disease. While it would be hubris to say that ICD-11 by 2015 will
incorporate all that knowledge, at least the infrastructure and machinery to
begin to accommodate that knowledge will be in place explicitly within this
environment.
This is another pocket slide. There is no data whatsoever on this slide,
this is whimsical, but it is the notion of activity within the bioinformatics
world and activity within the medical informatics world separated by what I
call the chasm of semantic despair, where these communities essentially don’t
speak to each other, this community speaking gene ontology more or less and
this community at best speaking SNOMED, without necessarily cross talking or
linking or recognizing. So these are among the sociologic challenges that we
are confronting.
The notion similarly that if we have highly aggregated definitions and fine
detail down to the level of single nucleotide polymorphisms for particular
enzymes, here a methyl transferase, related to a clinical notion of disease, be
it allergic rhinitis or asthma, deriving from anatomy and disease categories,
with cross linking activity, say immunology, that might bridge these things,
and speculative associations at the genomic polymorphism level, that is
effectively when a miracle occurs and we can move on.
This is the traditional hierarchy of ICD. It is essentially a strict
hierarchy by design, where the notion of mutually exclusive categories, so you
don’t double count mortality or morbidity, is explicitly represented.
What we seek to do is change these nodes into more complex objects, if you
will, where each of these formerly dots, now a square, that is important,
incorporate notions of definitions, synonymy and indexing terms explicitly in a
computable way. To do that we begin to have cross relationships and arcing
between and among these concepts.
I have often complained that if you have gastric cancer you are not allowed
to be a G.I. disease in ICD, you are a cancer. So while these mutually
exclusive categories, while important for the use case of population
statistics, get in the way when we try to understand disease relationships and
connections. So the raw ICD-11 will be very enriched with these kinds of cross
linkages explicitly.
However, we recognize that using that for the longitudinal traditional use
cases would be highly problematic. So we are serializing this cloud of semantic
relationships into a more traditional — this is actually ICD-10 that is
displayed here, but the premise is that there would be derivatives that would
look historically like the tabular versions that would be algorithmically
derived from the core mother of the ICD. As I said before, if we do that once,
we can do it many times, to have specific derivatives perhaps for mortality
coding, which would be relatively small, morbidity coding which would be much
larger, and other use case codings that would still derive from the common
core, yet give the appearance, feel and usability of the traditional ICD for
application.
So our task then is very complex, a distributed editing of complex
information — this will sound familiar to Keith, who did his thesis on a
variation on this — with the requirement that it be intuitive and simple to
use, no cost to users, low profile application.
Where I am going with this is, this is a slide prepared — I’m sorry, this
is Barry Hunt speaking, I forgot that it is voice on this slide. In any event,
he had the notion of Hi-Ki, which is a variation of Wiki. If we look at that,
there will be a requirement for a robust back end of appropriate concepts,
description logic capability, aggregation logic capability, leveraging
industrial and community standards, and of course
integrating with the existing terminologies.
That is a tall order for an editing and authoring environment, yet that is
exactly what we are trying to achieve. It implies that we have a suite of
development tools that include a lightweight distributed editor, description
logic capability so that we can bring in the modern classifier and ontologic
principles of disease, a robust back end for persistence and storage, and work
flow management to insure that this is all scalable across a community that we
would like to work with.
So there is now a consortium of open source distributed authoring led by
ourselves, Apelon, the National Cancer Institute and Stanford, who are all
contributing to common tooling. That involves LexGrid, which is a tool that is
used widely throughout HL7 and the CA Big community for a common model of
information, the Wiki environments that I will tell a little more about, and
then of course the collaborative protege environment from Stanford.
The LexGrid is now an ANSI and an ISO standard in terms of having a common
model and representation of underlying information. It is quite appropriate for
this particular use case, since it can support formal and logical definition,
has multiple import and export capabilities. It is an open source resource
naturally, supports multiple languages with 16 bit characters, and in the ICD
use case this is going to be terribly important. It permits intrinsically cross
linking between code systems. It can accommodate logic rules, and of course it
has the capacity to serve as an underlying central archive with distributed
nodes and synchronization.
The Wiki environment. The whole premise of Wiki as many of you know, with
Wikipedia it has become the largest and most successful social computing
project in history, quite to everybody’s surprise. This was not expected, but
it is still a remarkable resource that is actually from a computer science
perspective quite capable. It is able to sustain over 12,000 hits per second.
We think it unlikely ICD-11 development is going to have 12,000 hits per
second, but if it does — that is a shocking statement, I know.
It is also important to recognize Wiki alone is fairly staid. If you are
fairly with the semantic Web community and semantic Web technologies, they are
the folks that bring you fun things like XML, which many of you have heard of.
You may not have hard of RDF, the Resource Data Framework, or OWL, the Ontology
Web Languages, which are at increasing levels of sophistication on these
underlying representations, that permit a more formalization in a semantic
context.
The relevance here is that Wiki and semantic Web technology get married.
What we are talking about is a semantic Wiki, where the representation of the
underlying Wiki technology is manifest with RDF, Resource Data Framework, sugar
on top simplistically, which enables a rigorous formalization of an ontology
and a classification in Wiki machinery, so you get the best of both worlds.
That is actually quite pertinent and quite exciting. It is development out of
Europe essentially the beginning of last year that has transformed this.
This is our current prototype, again working in partnership with Apelon and
National Cancer Institute. This prototype is being used to edit the Biomed GT
at the National Cancer Institute today. As you see, it is very Wiki-like, but
nevertheless is under the hood a very structured organization of content. This
is ICD-10CM. The contents now of course include the usual suspects of ICD-10CM,
ICD-10AM and soon the German and Canadian modifications as well as a starting
point. Clearly we are not starting with a blank sheet of paper, but we are
learning from the existing clinical modifications of the nations that
contribute in the family of international classifications.
This won’t work unless we have a domain hierarchy. This is not Wikipedia,
where my mother can change any article she might wish, but there be a clear
hierarchy of the revision steering committee, the domain and topic working
groups, accredited experts, accredited persons by experts, and of course the
registered public.
What is the premise here? The task we are confronting of creating human
language definitions, complex relationship assertion, completeness criteria,
exclusion criteria and other formal definitional characteristics clearly exceed
the workforce that WHO can bring to the table.
We are hoping, and I use that word advisedly, that the same kind of
enthusiasm that has greeted the Wikipedia phenomenon and engaging the world
community in the process of writing a formalized body of scientific phenotype
linked to underlying terminology would engage the interest of many in the
health and public health communities, who would want to have their imprint on
this process so that the amount of input we would achieve we are anticipating
to be significant.
Our preliminary indications are, that is a valid assumption, but it will be
done in the context of a formal framework to insure the integrity and
consistency of the work emerges.
These will in turn after we go through the Wiki process create formal
change sets with integration between the Wiki world and the Protege world.
Protege is a formal tool developed at Stanford University by Mark Musand’s
group, an open source tool, that does the formal ontological reasoning in a
robust fashion. So that will be the back end touch-up environment to insure the
ontological consistency. We are not going to get that out of Wiki, but we will
get it as we get out of that cycle of going into the environment.
So the Protege OWL environment will be the final touch-up into a revision
or a snapshot of where we might exist, and this whole circle will continue to
iterate, we expect on a monthly basis, to insure a rapid incremental evolution
of the tools and resources as ICD-11 evolves.
This is just the change ontology as manifest in the Protege environment.
Let me move on to something that is perhaps more intriguing. That is the
issue of our formalized associations with the IHTSDO, the SNOMED people. It is
very clear that as SNOMED evolves, they will require high level nodes that
aggregate their granular data, if you think about it, with the notion that they
are mutually exclusive, exhaustive and so on and so forth, because the use case
to aggregate SNOMED data is incontrovertible.
If that happens, that sounds a lot like ICD, if you think about it.
Similarly, I have asserted and others have agreed that ICD-11 will require
linkage to a lower level terminology to integrate into these definitions of
what these rubrics actually mean, using description logic mechanisms and other
formalizations of creating these branches and arcs. If we are not careful, if
we build our own terminology, that starts to sound a lot like SNOMED, which
could be a problem.
So a potential future state, if we are not careful, is that the green space
— and my apologies to you that are color challenged here — the green space is
the classification space and the red space is the terminological space. There
would be a requirement for a green space in SNOMED. If we are not careful, we
will end up creating what I call a ghost SNOMED or a ghost ICD within SNOMED
that are conflicting, perhaps competing, and not necessarily a desirable
circumstance, and would make the challenge of terminological unification quite
problematic.
Where we would like to see this future — and as Marjorie knows, I spent a
good five minutes writing this PowerPoint slide, so I expect applause after
this happens — is, if we are very, very clever, these environments will merge
to the point that the alternate future becomes a joint effort, where by
definition the green layer within SNOMED is ICD-11. The dependencies of ICD-11
on an underpinning terminology are intrinsically based on the SNOMED
environment from which we can create these linearized derivatives for
longitudinal consistency and familiar tabular form as we move forward.
I must assert that the advantages to the organizations I think are obvious.
We avoid the notion of ghost emulations or competing or conflicting content.
Both organizations can leverage the expertise of others. They retain
independent intellectual property rights, so the ability to distribute the
green space, albeit without its full set of definitions, but nevertheless the
green space on a global basis without IP impairment, would be retained.
The most important point of course is that this mapping question that has
bedeviled the relationships between the ICDs and SNOMED for a decade now become
moot, because by definition ICD-11 is defined by these underlying relationships
and connection, hence making the mapping not only moot, but algorithmic if you
want to achieve it, because it is intrinsic with the publication of the
classification itself, how that aggregation would actually work.
There are caveats. Let’s be clear here. I don’t want to misrepresent the
reality. No agreements have been finalized. I can’t say that more plainly. It
is being considered by the IHT board. It is being considered by WHO Council. I
might add that there are members on the IHT board that are rabidly enthusiastic
about this idea. I can add that the leadership at WHO is rabidly enthusiastic
about that idea, but that does not always mean success, as many of you know. So
let’s be clear, no agreements have been finalized.
Intellectual property sharing is expected, but may be an issue, since those
countries that have not joined the IHTSDO organization may feel disadvantaged
if they don’t have access to the full terminological underpinning. This is a
challenge that WHO is anxious to avoid.
The whole notion of tooling. I have asserted that much of this vision is
dependent upon sophisticated tooling that will enable the ontological
consistency, whatever that means, as well as the distributed contribution of
the public and the community in a hierarchical organized fashion. It would be
elegant if we were to share tooling that would achieve that kind of goal. As
many of you know, IHT reincorporated in Copenhagen, Denmark last summer, and is
exploring the question of where the tooling would go. There has been
established a joint harmonization board across the two organizations.
On a related question of whither ICD-10CM and what would the ICD-11 project
do to the future and relevancy of ICD-10CM, NCHS and WHO at a meeting in
Otarara, Japan, began the discussion of exactly what that means. Basically the
answer is, everybody is going to play well together.
Simplistically, NCHS has agreed in principle to migrate ICD-10CM
development to the ICD-11 platform, with the explicit goal that as ICD-11
evolves and its content emerges, that the clinical modification of ICD-10 in
the United States could co-evolve, so that by the end of the period, which is
roughly expected to be the year 2015, the content of ICD-10CM and the content
of at least the tabular forms of ICD-11 would be congruent. So the notion of a
tumultuous transition from 10CM to 11 would be avoided, insuring that there be
harmony, and the argument then that we should wait for ICD-11 becomes less
compelling, given the consistency across the two versions as it evolves.
So NCHS would work closely with the revision committee to insure the
consistency, and that transitioning to ICD-10CM might provide then a mechanism
for tightening these linkages in a longitudinal fashion with an empirical and
practical experience from electronic health records, leveraging some of the
technology that would be used in the ICD-11 development.
That, I think, Simon, is what I had to say.
DR. COHN: Chris, thank you. I want to spend a minute or two framing why
this is important, which is a different conversation from ontologies and other
things that I fully respect in terms of being able to discuss in great detail.
We live in a world — and I think Justine has described this earlier — in
a hybrid world, a world of transition. We talk a lot about transition, but the
transition I am referencing right now is the world of moving from paper
environments to electronic health record systems, in fully computerized health
care environments.
Right now, we know that health care environments account for about 15
percent of health care, depending on which survey you look at. And we certainly
know based on what we have been seeing that it is unlikely that we are going to
suddenly turn around in 2009 or 2010 and suddenly discover full implementation.
So the question is how to deal with that from a data centric fashion,
preserving data, knowing that in a world of electronic health records you can
deal with a lot more coding and classification and access to information than
you can in the paper environment. If people are having to page through an ICD-9
book or a CPT book or whatever and deal with paper, there are limits. You have
got 10,000 codes, and in ICD it is 15, and you have got 7,000 or 8,000 CPT
codes, you have got hundreds of thousands of SNOMED codes, for example. We are
not counting LOINC or anything else.
So what you do with a computer is very different than what you can do in the
paper environment.
So the question is, how do you leverage, how do you move forward in a
hybrid environment where some people have very expressive data, can produce
SNOMED data with great clinical granularity, others are left in a world where
getting the right ICD code for this condition takes a lot of work, and on a
good day they are doing a good job on all of it.
There has been the promise of tight mappings between administrative and
clinical terminologies for many years, and it has bedeviled many of us. Part of
it is the fact that ICD changes, SNOMED changes, CPT changes, LOINC changes,
everything changing simultaneously. So yes, you can have a mapping that might
work on one day, but a month later, three months later, six months later, it is
hard to keep everything connected and together.
I personally am very taken by this vision, because it offers us the
opportunity to potentially successfully make the transition, what I would
describe as across one of the chasms, which is that transition from the 15
percent electronic health record implementation to the 90 to 100 percent
implementation, in a way that makes some sense.
One of the few things that I have heard recently — as I said, I would be
happy for good mappings and industry standard mappings, but it has been much
more of a promise than a reality.
So this is just one person. I have been living in a world using SNOMED
within Kaiser Permanente for many years; we were an early implementer. I also
live in a world of ICD and CPT, because clearly that is important things that
need to be reported, and also how reimbursement works.
So it is an interesting framework, and I think a very impressive one. So
Chris, I want to thank you for thinking this one up. But knowing that a lot of
people have got to play, SNOMED people, ICD needs to play, but it sounds like
we are developing a very usable foundation.
Now, obviously this is my own personal feelings. I just think this is a
very important activity, and I think I have said so in other public
environments.
Justine, did you want to comment, or do you have a question?
DR. CARR: Yes, I second all that you have said. I think another key
challenge is that even if you have the mappings done, it is not saying what you
want to say. It is a limited vision. Some of the things that you pointed out in
terms of, is gastric cancer a G.I. issue or a cancer issue.
I think right now, certainly ICD-9 doesn’t tell the detail that you might
want to tell in your problem list in the electronic health record. So there is
a barrier to using it, because you are forced to say something not exactly what
you wanted to say, which you could say much more clearly in free text. So I
think this is tremendous on all fronts.
DR. STEINWACHS: Just out of curiosity, Simon, my experience in the old
paper records systems was that frequently clinicians in ambulatory care used
the most aggregate kinds of codes. So that opportunity to get greater
specificity, to put more digits on the ICD, frequently doesn’t occur.
I was wondering whether or not as part of this vision of what electronics
systems could do, would have some encouragement for clinicians to give the
highest level of specificity that they currently understand about a person’s
disorder, or whether or not the tendency may still be, particularly in
ambulatory care more than in-patient settings to use the fewest digits you need
to put down a broad diagnostic category that meets the billing needs, but then
doesn’t always meet some of the quality monitoring needs.
DR. FERRER: Dr. Chute, do you foresee considerable maintenance efficiency
gain by this new —
DR. CARR: Say that again?
DR. FERRER: I’m sorry. Do you foresee considerable maintenance efficiency
gain with this new Wikipedia-like restructuring?
DR. CHUTE: It makes the question of public input and public for a much more
broad based and potentially global. So notions of correction update can be
systematically harvested from such a tool. The tool would be publicly
available. The question then is, what would the update and revision committee
that will persist, I’m sure, do with that kind of information.
So I think it would be hugely more efficient on the maintenance issue to
gather potential issues and new insights, particularly in the genomic space,
which we know will continue to accelerate over the next 20 years.
DR. TANG: First, I certainly regret that I have been deprived of the visual
experience that Dr. Chute has probably created there. I look forward to
hopefully getting some slides so in retrospect I can enjoy it.
The future ICD-11/SNOMED marriage sounds like nirvana, particularly after
you exorcise the ghosts that may appear or prevent them. One question is when
do you think you might anticipate the agreement between the dual board? Not
because it is the end, but at least it is the beginning of the optimism for
that world that you describe.
Maybe more concretely, and this probably piggybacks on what Simon said,
what does the community industry do in the interim? The interim may be even in
the next five years as we try to somehow get to ICD-10, which you describe as a
smooth transition over to ICD-11.
There are a number of folks that are trying to insert themselves into the
terminology void. Some of them have created unprincipled, perhaps even
undisciplined, terminology sets, but because they are there, people will use
them, and that may cause a disconnect with the transition and create an
impediment to the transition to ICD-10/11/SNOMED.
Any recommendations? Or how can we deal with the next five years, let’s
say, let alone the next whatever that is? You said 2015, for the next seven.
DR. CHUTE: Those are good questions as always, Paul. The status of the
negotiation is, we have exchanged formal memorandum of agreement and documents,
and they are in the hands of the lawyers. So the what we want to do is fully
negotiated across the organizations.
I would add that IHTSDO, which starts to get at your second point, is today
vastly more interested in collaboration between WHO and IHT on the ICD-10
mapping problem than they are on the ICD-11 development problem. You can
understand that. That is their urgent requirement.
In fact, Keith, you can correct me, but I believe Australia is
contractually obligated to have such a mapping tool by June across those
organizations, which is why Keith is living there, best I can tell.
DR. TANG: In ’08?
DR. CHUTE: In ’08, yes. That is ICD-10 to 10AM, as it happens, Australian
modification, to SNOMED.
Now, this is very much carts and horses and chickens and eggs are littering
the landscape here, but those mappings will clearly inform the preliminary
aggregation logic framework as we move forward. It is not as though these are
independent threads, quite the contrary. They are highly sequentially dependent
and synergistic.
IHT has made very clear that their enthusiasm for taking on the ICD-11
project as the second half of this partnership, the first half being the ICD-10
mapping, will be increased when they see evidence that their member states, the
usual suspects, United States, Canada, Australia and so on and so forth, U.K.,
have any interest in the ICD-11 project. Let’s be blunt here. That I think is
possible as we evolve into this more elegant unification of the two
environments and a mapping environment rather than dealing with the traditional
mapping architectures.
So what we do in the meantime I think is to leverage and refine
these mapping tools. Jim Campbell, the other Campbell, is chairing the
IHTSDO mapping efforts as we speak. They are aggressively moving forward with
ICD-10 mappings. I know Betsy is very involved in the whole notion of ICD-9
mapping, and HEMA of course is very involved in the ICD-10 mapping under a CMS
contract, as I understand it.
So there are a lot of forces at play here, looking at the interim, which
simplistically is a SNOMED to ICD-10 mapping that would sequentially inform the
ICD-11 development process.
DR. TANG: So you are saying that the ICD-10 mapping to SNOMED is virtually
a certainty. In fact, one instantiation of that is in June ’08, the Australia
modification will be — I don’t know whether you said completed or not, and
that NLM is working on the ICD-9 mapping?
DR. CHUTE: That’s correct.
DR. TANG: So that gives a lot of guidance, if I have stated that correctly,
to what the industry can do. I’m not sure they are well informed about this.
DR. CHUTE: And I don’t know the status of the ICD-10CM mapping, but I
believe that is the CMS contract. Marjorie, are you aware of that status?
MS. GREENBERG: Everyone is interested in what is happening today. So the
first priority as you said for the IHTSDO is to map with ICD-10. But I don’t
think we can underestimate what we will learn from that, because where it
doesn’t work is where we would want to modify SNOMED, we would want to modify
ICD. So it is just like a building block for that.
But it is the same in the U.S. The first interest is in mapping to ICD-9CM.
I think people can see how pathetic that is, to be mapping something as rich as
SNOMED CT to ICD-9CM is ancient classification, if you will excuse me. But
there is also NLM and working with AHIMA, and Linda might want to say more
about that, but is committed to a map with 9CM and then a map with 10CM.
But the thing is that of course, the international community is completely
disinterested in 9 for obvious reasons, and the map with 10CM will certainly be
informed by the map with 10, since CM is based on 10.
So as long as we are using 9CM and some of the leaders like Kaiser are
using SNOMED too, then we have got to deal with the 9CM map. But I would say
the sooner we can move to this mapping with 10CM, it makes a lot more sense.
But we are not going to get any help from the international community on the
map with 9CM, but I think we will with the map with 10CM. Also, 10CA and 10AM,
the Canadian and Australian ones, they have looked to 10CM development.
The good news is that 10CM has not been implemented, or there is not even
an official plan to implement it, but it is already part of this process, as
you can see in those slides. They went by pretty quickly, but it is a building
block. So in that sense we are glad to see some use of it because we have put a
lot of work into it.
DR. TANG: So is there a time horizon for 9CM mapping?
MS. GREENBERG: Betsy Humphreys and AHIMA would have to address that. I
don’t know what the time line is.
DR. COHN: Linda might be able to comment on it in her presentation
following this. So Paul, we will hold that question.
DR. TANG: Sure.
DR. WARREN: Simon, this is Judy. Can I get on the list?
DR. COHN: Sure. You are getting down here pretty far, but we will put your
down.
DR. WARREN: Thank you.
DR. GREEN: I have 85 questions.
DR. COHN: Take your top one.
DR. GREEN: That is hard. That was just terrific. We really have to capture
this for our medical home hearings in some way. The ideas underlying this are
crucial.
I don’t see in your slides and I didn’t hear in your comments how this
exercise would create episodes of care. I saw the groupers and I understand how
some groupers work, but episodes of care in natural language create meaningful
concepts. It strikes me as being central to one of the United States’ most
pressing issues, which is to figure out what it takes to get value for a dollar
in medical care.
What is your thinking about how this exercise can create meaningful
episodes of care?
DR. CHUTE: That is an excellent question. Some of you may remember a paper
we coauthored with Pat Givens years ago on the notion of variable episodes of
care, use case specific episodes of care. I don’t believe that there is a one
size fits all interpretation of that concept.
With respect to the ICD-11 project and aggregation logics, clearly those
aggregations have to operate over a framework of either time, space, document,
content, to be meaningful. The question then of what is that time or space is
not addressed in the ICD-11 project. Think of it as a feature. That becomes an
exercise for the user to define what is the framework over which they want to
execute these tools and aggregations. The machinery will work equivalently over
a single document or over a collection within what you might choose to call an
episode. The machinery is indifferent to what that scope is, other than
recognizing there is clearly a scope that is required.
MS. GREENBERG: I just want to add something to that. This is all going on
while other work is also going on with other classifications as part of the
family of international classifications. That is mappings between SNOMED and
ICF, the functional classification, and ICPC.
I know that ICPC starts an episode with the patient’s reason for a visit,
or why the patient came in, from the patient’s point of view. There also is
work going on between ICD and ICPC, and there is a commitment to, as Chris
mentioned, being a primary care view as well of the ICD-11.
So I think obviously the primary care community has to be very much
involved with this as well, but I think the opportunities to incorporate all
the terminology that you need and structured the way you want it, et cetera,
exists.
DR. FRANCIS: I can’t resist being a philosopher for about two seconds
whenever I hear the word ontology, which is essentially about the reality
status of a classification scheme.
That led me to think about the use cases. The Endangered Species Act
involves a definition of species, and that is actually an ontologically
problematic definition of species. There has been all kinds of interesting
questions about public policy, there being a mismatch between public policy and
ontology in that area that the philosophers of science have gotten all
interested in.
So that may seem wildly a field, so here is the question. I think there are
enormously interesting relationship between the kinds of ontological
understandings that you have and the public policy questions.
In particular, think about whether a genomic classification or a phenotypic
classification makes more or less what the ontological commitments are of
those, and even what we understand as a disease, what is it that we are
classifying here, which can have enormous differences in how you frame the
public policy.
So I just wanted to point that out. Maybe it is just that what I really
understood I thought were the first two slides. But I think it is just
incredibly interesting to view this as a deep set of ontological problems and
what is that going to tell us over the next 20, 30 years from a policy
perspective.
DR. CHUTE: Your points are germane. I suspect we could spend the next 30
hours untangling them.
Let me simply summarize by saying the primary focus — and one of the
slides included this — is scientific phenotype as a philosophical bias, if you
will.
DR. MC CALL: First, absolutely loved the work. This is very exciting stuff
for the geeks in the crowd, of which I proudly label myself one.
Second, and this is really a two-part question, it goes back to something
that Simon said about Kaiser’s views and physicians maybe being able to enter
more detail. Then the second part goes to Larry’s question about groupers and
the exercise left to the reader.
So to the first part, I don’t see that doctors are going to maybe enter a
lot more fields so that somehow they can use a deeper level of the ontologies.
But my question for you is, is it possible if an EHR naturally captures certain
elements, maybe at different levels, that the answer and all of the other links
could be much more naturally made, so that they don’t have to go hunting for
the right SNOMED as one of 150,000 possible distinctions that is imputed?
DR. CHUTE: You are talking about the human-computer interface for the
underlying terminological grist for this mill.
DR. MC CALL: Right, and there are implications for EHR.
DR. CHUTE: Now I am getting philosophical and speculative, but I have
always regarded that problem as not something a human clinician would do,
figuring out which of 300,000 SNOMED codes I want to use today. Rather, I see
much of that being supplanted by natural language processing technologies and
term mapping technologies that would draw pertinently from what the clinicians
say and identify the salient gems.
We are doing this in our own back yard for research purposes, but the
scalability of this technology is becoming a commodity, and the ability to
create this level of terminologic detail without the clinician having to fuss
about it, is emerging.
DR. MC CALL: Right. So back to your comment about what can we do in the
next five years, when we start thinking about adjacent, maturing but necessary
technologies, this is one, in order to leverage that.
DR. COHN: I would just give you a counterpoint on that one. That is one way
to do things, but if you think about it you have preference lists. You have
dropdown lists. SNOMED is very powerful, you can say anything, but you also
want to help people say what they need to say in a way that is helpful to them.
So you aren’t always dipping into a complete vocabulary. If you are an
orthopedist you deal with a lot of fractures. You may not need to have a deep
discussion of skin diseases or whatever. So this is part of how you access data
in terms within an electronic health record. So this something where the actual
tool can help.
DR. MC CALL: I think we may be saying the same thing, but using different
technologies and methodologies.
The second is, I agree that it is left to the user, so the exercise is left
to the reader to put together episode groupers. If I am a user and I want to
use this, how does that group tap what was on those slide? How can you begin to
try to use it?
DR. CHUTE: Now you are talking about what is going to be the physical
manifestation on the associated software, associated with the aggregation logic
itself. We intend as part of the ICD-11 process to publish the algorithm. We
will recommend software, but I think it is beyond our scope to write the
implementation that would actually do it. We will have reference
implementations naturally, but I would anticipate that this would be a level of
functionality that EMR vendors will begin to incorporate, given the algorithms.
DR. MC CALL: Sure, absolutely. So the question is not so much that you guys
write them, I agree it should be left to the user, to the reader as an
exercise. The question is, what is the right way to tap in. You are saying you
will provide — it is all open source, it is all there, you can tap in with
whatever you like.
DR. CHUTE: Right.
DR. COHN: We have three last questions. We are running a little late, so I
would ask if they can be succinct and pick their top question as opposed to
multiple questions.
MR. BLAIR: Applause, applause, applause. I am trying to put some things
together here. I love what you showed us. It is a wonderful strategic map into
the future.
The pieces that I am trying to reconcile it with, and I am having
difficulty, is that we have all this effort that is going into from AHIC use
cases down to HITSPE standards with the ISO 2 and 3 and 4 and all of those, and
the IHE standards and all of those things being implemented. It seems as if
that approach and mind set, do you see somehow where they will begin to connect
or integrate or harmonize with the approach you just gave us?
DR. CHUTE: Wow. Clearly the consolidation of standards activities in this
country has been advanced by HITSPE and ONC. There is no question about that.
If I were to speculate on an answer, Jeff, it would be continued
coalescence. The old joke about standards is, there are so many to choose from;
you have told that one yourself. As we coalesce on a consensus component or
interoperability specifications for standards on how you actually use them, I
think much of this competing activity will begin to clarify itself.
What role this would play in that process gets at questions such as how do
we evolve the U.S. realm, or to say this collection of value sets and codified
terminologies, what is the future of KAI? Is it going to continue as a process
within the standards activities to define the overarching coding systems from
which one will derive bite-sized or application-sized elements of terminologies
and value sets.
I see that as the kinds of trajectory that ONC and its collaborators are
on, and you know the answer as well as I do, Jeff.
MR. BLAIR: Thank you.
DR. WARREN: Chris, I agree with Paul. It is hard sitting here on the phone
and not seeing the slides.
One of the questions that came up to me as you were describing how you saw
the classification of ICD-11 interfacing with SNOMED. So I was wondering, as
you are putting that process together, what you are doing to insure that there
are voices included in this process from nursing, primary care and allied
health, and also how you see yourself articulating with the other
classification systems that will have to interact with SNOMED such as ICF and
some of the others.
DR. CHUTE: Well, there are a number of allied health and nursing personnel
on the topic advisory groups that have been selected so far. As you may know,
Marcy Harris, a nurse researcher, chairs the terminology committee within the
— or is co-chair of the terminology committee within the HUFIC. I think we are
not ashamed of our representation and incorporation of a diverse community of
health care experts in this process.
With respect to the other terminologies and integration, as you know, the
ICF is one of the members of the family of international classification, along
with ICD. So I think that partnership, particularly at the level of WHO FIC,
Family of International Classification, is actually quite tight.
The more interesting question is, what is our relationship at the
terminological level with other issues beyond gene ontology and SNOMED? To what
extent would we take detailed granular functional characteristics
specifications, patient status specifications and other terminologies that are
evolving beyond the scope of say diseases and functional status or diseases and
genes. How will we manage the drug problem? Will we use RX Norm, will we use
some international classification, because clearly drug allergies and drug
effects remain if you will diagnostic concepts, although this gets at the
philosophical notion of what is a diagnosis, but let’s not go there.
DR. WARREN: Thank you.
DR. COHN: Mike, you have the final single question.
DR. FITZMAURICE: Excellent technical presentation, Chris. Forgive me for
being a little practical. I am wondering how dependent will this effort be on
increased U.S. payments to WHO? SNOMED is based on large payments to IHTSDO.
Who supports WHO in this endeavor? That is the only time you can say that and
be semantically correct. Is it more important for U.S. to wait for ICD-10 or
more important to move right into ICD-11?
DR. CHUTE: The Director General of WHO issued a nota bene, I believe they
call it, which is this formal designation that the ICD-11 process has started.
So we are dealing with a political reality.
On the funding issue, to WHO’s embarrassment, most of the money to support,
and it is a fairly limited amount, comes from the Japanese Hospital
Association, of all places. The amount of money that WHO is investing directly
into this process beyond in-kind support and staff is negligible.
We are having a meeting, as you may know, on March 6 that Marjorie is
organizing here in Washington to address the very question that you asked.
MS. GREENBERG: This is just a briefing with stakeholders in the Department,
so it is a small meeting. That is why maybe you haven’t heard about it. But
Simon will be there representing the National Committee.
But as Chris said, this is not heavily funded. The briefing is not to pass
the hat, but to help people understand what this vision is, what the plan is,
how it relates to our role in IHTSDO, as well as our plans or not for moving
ahead with ICD-10CM. I think it is very under understood by most people, and I
think it is important that they — and the folks from WHO are coming over for
it, so they give it a high priority.
DR. CHUTE: I might add as a final note on that thread, Mike, this is why we
are invoking Wiki technology, quite frankly, to leverage the good will,
knowledge and intellect of the global community to decrease the cost of
development.
DR. FITZMAURICE: I think that is absolutely great.
DR. COHN: Chris, thanks very much. Why don’t we ask you to stay up there if
you would, and let me invite up Linda Kloss and Keith Campbell.
By way of introduction, I think you all are familiar with Linda Kloss. She
is Executive Director for AHIMA. I am also pleased to welcome Dr. Keith
Campbell who is joining us. I should say by way of public disclosure, as he was
getting his PhD Kaiser Permanente was one of his first employers. We were
pleased and delighted to have him for a number of years, helping us figure out
how to put SNOMED into Kaiser Permanente. So this many years ago. We worked
closely together. It is a real pleasure to see him back, and have a chance to
reconnect.
I know you have been doing a lot of very good work talking through
nationally what we ought to be doing with health care terminologies and
classifications. So please, let me let you take it away.
Agenda Item: AHIMA/AMIA Briefing
MS. KLOSS: Thank you. I will begin. Thank you, Dr. Cohn and members of the
committee for providing us the time to present to you what is a white paper
that was done under sponsorship of the sponsorship of the American Health
Information and Management Association and the American Medical Informatics
Association.
I suppose we could subtitle this white paper what we think we should do to
get our house in order to be ready for the kind of world that we just were
discussing is possible. We came together as two organizations and created an
expert panel to wrestle with just that.
What we are going to do is move through both what the charge was to this
panel, what some of the findings were from their work, what our recommendations
were, and we will end with some recommendations for this body. Again, thank
you.
I would like to clarify one comment that was made earlier. Our AHIMA
foundation and AHIMA is doing some work currently with CMS, but it is not on
mapping. It is to look at internal systems and make recommendations about what
it will take for CMS to prepare for ICD-10. So we are not involved at present
in any mapping. We are doing some mapping under contract with National Library
of Medicine, but it is ICD-9CM at this time, unfortunately, as Marjorie said.
Our organizations came together. This was initiated in 2006. We released
the report of this white paper in the end of 2006. We have published it in each
of our journals. We have used the last year to go forth and educate those who
are making decisions and have interest in all of the stakeholders on the
current state of our house as a nation and our ability as a nation to move
forward on all that is involved in management of terminologies and
classifications, so that we are ready to be a full participant in the world on
these issues beyond the wonderful research and contributions we are making.
We have been proselytizing and educating. One of our early lessons learned
is that this very complex white paper that you have been given a copy of needed
to have the English language primer version. So we took time at the end of last
year to do a shortened version of this report, so when we talk to Congress or
others who are not in the terminology-classification world, they understand
what it is that we are trying to say.
We think AMIA and AHIMA were uniquely ready to come together around these
issues, because our two organizations work a lot on these issues, but from a
little different perspective. AHIMA’s members certainly are in operations
management roles as it relates to coding and classification and working day by
day with the realities of ICD-9 and the development of electronic health
records and their full potentiality to use some of the technology that is
possible.
Our organization in 1999 did its own white paper and said that we were in
full support of the advancement of computer based coding as a way to improve
data and improve efficiency, and have been moving in support of all the ways we
can to make that happen.
And AMIA’s membership are the scientists who are designing new solutions.
We have two of the best here today. So we think that terminology and
classifications are the language infrastructure for the Nationwide Health
Information Network. We believed firmly when we came together that the U.S.
lacks a robust nationwide strategy for coordinated development, maintenance and
use of terminologies and classifications, and we lacked some of the
infrastructure needed.
Standards harmonization as we are viewing that, good progress is being
made. Alone it won’t solve the whole problem as it relates to terminologies and
classifications.
Our expert task force members was chaired by Keith Campbell, and Chris was
a member of that, as was Sue Bowman from our staff and several other alum of
the National Committee on Vital and Health Statistics, so I think you will
agree when you look at these names that we did bring thought leaders together
who understood the full range from the theoretical down to what we are dealing
with in terms of inconsistent policies or execution of coding principles in the
field.
The charge that we gave them, and it was a daunting charge, and we gave
them only six months to accomplish it, we said, describe the need for a
coordinated policy on terminologies and classifications, set that stage for us,
map a high level near term plan to accelerate. Our goal wasn’t to come out with
a definitive plan, but to start the dialogue, and then to outline thinking and
requirements for ongoing adoption and maintenance of terminologies and
classifications. So that was the charge we gave them. We said, you have six
months, work hard and develop a white paper, and they did, and we are here to
describe what the thinking is behind the recommendations.
With that, we will ask — I don’t think we need to stop to define these
concepts. We have just had them well defined for us. We did acknowledge — and
when you read the white paper, and we do encourage you to read it, we
acknowledged the substantial progress that has been made in recent years. The
National Library of Medicine’s leadership, NCHS, of course, NCVHS, the
internationalization of SNOMED and work by HITSPE. This work has even
accelerated in the last year. But we still believe that it is not enough to
solve the U.S. specific infrastructure issues that are going to need to be
solved.
So we put forth a vision of the desired state. We studied lessons from
other countries, and we developed critical short term actions.
I’ll turn this over to Keith to present our vision and discuss that.
DR. CAMPBELL: Thank you. I think that the first thing that was very
apparent among the task force members was that there was a frustration of lack
of a clear coordinated vision within the U.S. as to where we needed to go with
regard to terminologies and classifications.
There was I think also a consensus — actually, I’m sure there was a
consensus, this is in the report, that one of the things we needed to do to
overcome that lack is to work to develop a funded organization that is
responsible for setting policy and for shepherding work products that are meant
to support or deliver terminologies and classifications products that are meant
for us in the U.S. market.
As part of that, Linda mentioned that we looked at some of the things that
other countries were doing. One pre-eminent example is the British National
Health Service. Because the health care systems in some of the other countries
that we have looked at are different, the ability for them to create a central
organization that coordinates activities on behalf of the nation is enhanced.
In the U.S. we often have certain proxies for organizations like that. For
example, Kaiser Permanente is a strong example. Kaiser had a need for
terminologies many years ago, recognized it, and wanted to make strategic
activities in that regard. I think the U.S. Veterans Health Administration is
another example. You get large organizations that are capitated with regard to
their payment, and they are the ones that are needing to turn to health care
electronic data to be able to improve their outcomes and other things.
In the absence of a nationalized health system, how do we create a national
strategy? It is I think a little harder than in countries that have national
health care organizations, but I don’t think it is unattainable. How do we do
that? We really need to have a central organization that does that.
As part of that, I think there was a strong consensus in the committees
that volunteerism was not sufficient. It was not sufficient to put up a
Wikipedia, for example, for the U.S. and hope the national strategy would
somehow be resolved out of doing that kind of work.
That said, I do think something such as a national Wikipedia is an
important enhancement that could insure that we have more pluralistic
representation and be a valuable contributor to that process, but it is not
sufficient. There is a need for coordinated efforts.
Another country that we looked at in some detail within the committee was
Australia. They also have some similarities with the British National Health
Service, but they have some differences as well. Whereas in the NHS, their
terminology strategy is basically within governmental organizations, in
Australia they chose to use public money to create a private company that was
called the National Health Transition Authority, that was meant to deliver on
the terminology needs of the country for them
So in the U.S. we need something here. I think that is the fundamental
message that the task force wants to come across. The rest is supporting
information, but how are we going to get this national governance, how are we
going to migrate away from unfunded mandates or unfunded activities that are
dependent upon — some organizations are stepping up to the plate to try and
make an impact here, I’ll mention Kaiser, I’ll mention the Veterans Health
Administration, but that is not a majority of the health care dollars in the
U.S. How do we get everybody else to coordinate? How do we get the electronic
health care vendors confident that the directions that are being promoted are
the ones that they should be investing and getting their systems to
interoperate with.
Another thing that was recognized was that the U.S. shouldn’t be trying to
create policy in a vacuum. I think there was a recognition that we are behind
in some regards where other countries are because wouldn’t have central
coordinating policy bodies. So it is incumbent upon us to mesh our policy with
those of others and to leverage activities through the new International Health
Terminology Standards Development Organization, leverage activities through the
World Health Organization and others to try to make sure that what we are doing
is in synchrony with what is happening internationally, not something that is
independent and an increasing burden on implementers to implement something
that is unique for the United States, whereas for everywhere else it is going
to be using different sets of standards.
So with that discussion of vision, let’s go ahead and move on to
collaboration.
I think there is also a recognition that we can’t do it alone, nor should
we want to. The health record vendors want to sell health record systems to
multiple countries. They don’t want to have to create a custom product for each
country because the regulatory requirements around terminologies are different.
So we need to collaborate. We need to have infrastructure for that. Chris
talked about the need with WHO and the ICD-10 and 11 efforts to have that kind
of collaborative infrastructure. We need it, but there still has to be a
central policy body and a central organization that insures that the lights are
on on a daily basis for that kind of activity.
Let’s go ahead and move on to the context. The goal here is that we want to
create the adoption of models and of standard terminologies. One goal would be,
we talked a little bit today about the needs for mapping between SNOMED and
ICD-10 and ICD-11 and how might we automate that coming down the track.
I think that those are absolutely essential. But one of the first steps I
think is to standardize things at a lower level, to make sure that we have a
consistent and comprehensive way of managing laboratory information and
pharmacy information as well as the problem with some diagnostic information,
because these machine readable sources are going to be the reliable means with
which we can compute the aggregation classification that Chris is talking
about.
If we are only relying on a natural language processing promise to be able
to realize this vision, it will not be realized. We need standardization at a
lower level. We need to know better about how these terminologies evolve over
time, so that we don’t have the ICD-10, ICD-11, ICD-12 upgrade problem. It
starts to become the ICD that evolves gracefully with continual changes, and
you don’t have to go through this big process of re-updating your systems. That
applies equally well if not more so to the laboratory information, to the
pharmacy information and to the lower level clinical encounter information that
may be recorded within your records.
That takes us on to the next slide. The goal here is a longitudinal cradle
to grave electronic health record. In order to attain that, we have got to get
away from this notion of episodic releases of terminology that create great
disruption to the health care organizations and to the software vendors trying
to implement these systems.
So as much as we talk about efforts at harmonizing ICD and harmonizing
LOINC and harmonizing SNOMED and harmonizing with RX Norm, we need to talk
about what is a unified model of change so that when terminologies are
implemented in an environment, we can evolve those environments gracefully over
time so that we don’t end up with a situation where we can’t realize the
benefits that we are intending to accrue as a result of electronic health
records, because we can’t upgrade the darn system, because what we need is
significantly different than what we have implemented.
With that kind of background, we will be able to deliver on the next slide,
which is by having that foundation, we can have real time patient specific
decision support that provides things like alerts, reminders, protocols,
advising and other things, and from that strong foundation of granular clinical
data in a machinable form, I believe that is where we can use techniques to
analyze that together with others to help make the classification problem more
automatable, more reliable and more granular.
Finally, with regard to the data, we can then share that data not only
within a single organization which today there is still a problem. When you get
an organization that has more than one pharmacy and more than one laboratory
and more than one physician, it is frequently the case that you cannot
aggregate those data on an enterprise scale, let alone aggregate that data on a
national or international scale, which is the purposes that ICD-10 and other
classifications maintain.
If we can make it over this hurdle of how do we get terminologies to evolve
gracefully over time, we then start to have a foundation of creating a new kind
of health care marketplace. This is where we can separate application
development from data persistence. So if your health records are not tied
tightly to a particular vendor’s implementation of something so that you cannot
— one of the things is, if you have to upgrade your system you have to throw
out the last 15 years of data because it is incompatible with the new
environment, that has a dramatic impact on switching costs. So people aren’t
able to incorporate latest technologies and latest innovations that may benefit
their patients because of the other costs of upgrading incompatible data within
their environment.
So we spent a little bit of time saying what is needed to create this new
kind of marketplace. We want to be able to have a standard set of detailed
clinical data models, independent of whether it is populated with SNOMED,
LOINC, Rx Norm or something else, that would need this universal representation
of terminologies and change.
Within that, we can then say, you want whatever terminologies, we have got
to make sure they fit within these formats. We have to have standard APIs for
the software that uses them, and have open sharing of these models, coded terms
and APIs.
This gave a set of challenges. We are going to come back to what was the
primary message, fragmented governance and lack of funding. We need to have a
unified governance model for the U.S. that sets a coherent strategy that people
can work towards. This used to be done in coordination with other countries. It
needs to be funded. Through that funding will come incentives to collaborate.
There is really very little incentive to collaborate when everybody is working
on independent funding streams, and within those funding streams there is no
requirement to collaborate. In fact, the primary requirement is to get your
deliverables out independent of anybody else. There is good reasons for being
able to do that, but we need to have some carrot that can help people to
collaborate in a more meaningful way.
So with that, the next thing that came back were specific recommendations
that the task force made to the AMIA and AHIMA boards. So I am going to ask
Linda to wrap up this set of slides.
MS. KLOSS: I will offer a couple of comments on some of what we learned,
and maybe ask Keith to comment also.
One of the things that we did was to create tables which even if you just
look at the pages, it comes home as to why this is an important thing to talk
about, on pages 23 through 30 of this report. We did one table listing the
current terminologies that are being used somewhere in the U.S., who owns them,
I don’t mean literal ownership, but their sponsor and their owner, and what is
the governance process, then another separate table on what is the maintenance
all about for that.
I think it brings home how diverse this is and how uncoordinated it is, and
the release process, the maintenance process, the development process,
ownership process. We talked about that and tried to map this out, not to offer
specific solutions, but to lay out the enormity of the problem and how we need
to work at solutions for moving to a more coordinated state.
At that point, the task force realized the enormity of this challenge, and
said that recommendation number one is to do more formal R&D, looking at
preparing specifications for coordinated solutions, plans for consolidated
terminologies, dealing with critical elements such as licensing technology,
data integrity requirements, maintenance, release schedules, education, prepare
a road map for change and make recommendations about a governance model for a
centralized authority.
So understand that in this white paper, there was no way to push it to the
level where we could say, here is the answer. But we hope we raised the issues
and put enough meat behind them to create some sense of urgency and some
commitment to needing to do this. Otherwise we are going to have a hard time
knowing what the solution should be.
Our recommendation number two was then to fund a centralized terminology
authority. We did say that is both public and private. We rely on the national
center, we rely on CMS, we rely on NLM, we relay on government agencies to do
important pieces of this, to represent us at World Health Organization. But we
have a lot of important private sector work going on and we need to come
together.
To oversee funding and development of U.S. terminologies and
classifications. We could envision this authority would have a source of
funding, that it in turn would become a funding mechanism to do important
research and development work, to divine the governance structure. Here, we
didn’t get a chance to update that from SNOMED SDO to ITSDO, but to look at
that governance structure to support non-proprietary development of standards,
identify certification standards and processes.
We are already seeing in the market that we know it will come to an issue
of compliance, to understand how algorithmic coding is working and whether it
is in compliance with reimbursement requirements, and how will those
application software be confirmed to be in compliance with maps, and oversee
development of implementation guides and linkages to other HIT standards and
policies.
I think our belief was, having looked at this, that it is a huge work
effort in and of itself. Obviously it needs to be coordinated with AHIC’s
successor or whatever larger more umbrella public-private mechanism it is, but
it in itself is enormously complex, and it is not a matter of determining what
harmonized standards are.
Then our third major recommendation was, then once we get there, we have
got to figure out how to get everybody committed or incented, more likely, to
adopt the sound principles of terminology development, so that going forward we
do have consistent use of these tools, which is not also today’s state.
So I think we will stop there and call for some discussion. We saw the
future. I hope we have given you a dose of reality. I guess having listened to
your strategic planning discussion this morning, I would just say that we would
urge that this be an element, because I don’t think we will get to the
information framework for health care improvement without tackling this very
specific thing, and tackling it broadly enough to deal with this from a
development to a release and education standpoint.
Thank you.
DR. COHN: I want to thank you both very much. I think this has been a very
helpful both review as well as insights into the current state of what is going
on. I think people sometimes think that certain efforts are solving all the
problems, and obviously it is a layering issue. You get to a certain point, and
underneath it is all this data.
I am going to ask perhaps a very basic question. I was listening and
reflecting on previous NCVHS letters and recommendations. I was remembering
back to a letter that we sent to the Secretary a couple of years ago that was
recommending that the National Library of Medicine do maybe not everything you
are describing, but many of the things you are describing. We didn’t think to
invite Betsy Humphreys to provide her input on this one, probably we need to
have her in May to talk about it.
MS. KLOSS: She did review this white paper.
DR. COHN: I’m sure she did. I guess I am curious what your perceptions are
about — that is not public-private, that is just public, but that role versus
the role of a separate public-private terminology authority. Can you give us
some of your thoughts on that?
DR. CAMPBELL: Yes, I will try my best to faithfully represent the
deliberations of the group. Chris and Linda were both present, so they can try
to reorient me.
I think the role of the NLM has been very helpful over the last 15 years in
this area. The efforts that they have put into the UMLS, the efforts that they
have put into educating the community and other groups are critical.
One of the things that is needed for — there were some discussions within
the group as to, is the NLM it or is something else needed. I think the group
discussed that there really wasn’t the connectivity between some of the
activities of the NLM with regard to this we are doing and the health care
vendor community. It wasn’t quite that tight. For example, the UMLS was meeting
a particular set of needs, but some of those needs it was meeting weren’t
necessarily the ones that the vendor communities felt were of high priority.
So there needed to become a governance process by which there was industry
representation and industry input directly into the activities to be able to
help prioritize and steer things in ways that meet a different set of
interests.
So within NLM, you could discuss or debate to what extent implementation of
electronic health records fits within their mission. I think there was some
recognition that some aspects of what they do or some aspects of what is needed
in terminology falls outside of that mission; could it be reconciled within a
group within NLM by improving governance and accountability and having industry
boards that sit and direct activities perhaps?
I don’t think we meant to preclude exactly what that government and
public-private would look like, but the thing that was quite clear is that it
needed to become more relevant to industry, and industry needed to come and sit
and participate in that in some way.
DR. FERRER: I had a question. With the AHIC 2.0 clearly there is a lot of
alignment, but do you feel there is a point where you would undermine the
effort of what you have outlined? This is really the architectural foundation
of a good bit of the world of medical informatics, and AHIC although it has
done a lot of very good work, the products have heavily been laden on the
informatics standards. There is a certain slowness in the machinery that is
coupled with that.
Do you feel that — you talked about, is truly well served in an
organization that has such a political underpinning, if you will? This is not
something that is going to be done in a very short period of time.
MS. KLOSS: I think our experience over the past year talking about this
work and even over the past ten or 20 years talking about these issues, is that
the deep issues of terminology and classifications are not very well
understood. You all know that because this is this committee’s work.
I think we have an enormous amount to do to help all parts of this come
together to understand how foundational this is. Message exchange standards
aren’t going to do it. We have had many discussions, I know we all have been in
those, where people will say once we have the electronic health record, then we
will be able to report the data we need for quality measures. Having missed all
those other parts there, we need a standard vocabulary and we need middleware
that will extract.
So I just think it really still is part of the industry education. Whether
at some point there can be — I think there has to be a very tight connection
with this work and the AHIC too, however that works, dotted line, direct line.
I think that is exactly why our group stopped short of saying here is the right
way to design this. We said, here are the enormous needs.
It is going to take funded R&D, even to figure out what that connect
should be. But we wanted to raise the sense of urgency, with moving that
deliberate study.
I think it really comes home when you hear about how other countries have
looked and viewed with organizing just this set of functions. It needs to be
funded. They need staff. People need to go to work every day to do this.
DR. CAMPBELL: Just to add, the sobering look at the budgets these countries
have just in terminology. If you were to try and tease out how large the NHS’
release center budget is and the amount of money they put in to Read One, Read
Two, Read Three, SNOMED releases, other things, and look at things like the
budgets of the National Health Transition Authority in Australia, I think you
would find considerable disparity with regard to the amount that is directed
funded in the U.S.
DR. COHN: Mike Leavitt has mentioned in public that he wonders at times
whether we really have a national health care system. From what you are
describing there is something much more formalized which this of course
supports as a strategic agenda.
DR. MC CALL: Simple question, four words. How can we help?
MS. KLOSS: Put it on your critical planning agenda, the issue of helping to
come to a consensus as to what our U.S. solution needs to be. I think it is
central to the mission of this group to take that up.
The struggles we have had to get from ICD-9 to ICD-10 are just symptomatic
of this issue. That is what triggered our white paper. It is symptomatic. This
should be a routine.
DR. COHN: I think we are also asking Betsy Humphreys to come for our
meeting to brief on her part of the role in all this stuff. I think it would
help us to get another view of all of this.
MS. GREENBERG: I was just going to say, I hate to say the world is flat and
everything is related to the 21st Century for a Vision for Health Statistics.
But the central role for classifications is statistics. This is certainly a
component of what we were talking about before the break. You can’t really do
statistics on terminology, but you can on a ghost.
I think these all fit together so well, I would have to say. It does seem
to me that it should be a critical element. I think classifications are
mentioned in passing in the old vision, but maybe they need to be —
DR. MC CALL: If not number one, the top three.
MS. KLOSS: Just the issue of — I remember when the code sets got rolled
into HIPAA. We were sitting here clapping. Okay, this is good, things are going
to move. These kinds of releases probably in the long run shouldn’t be
regulatory, but how should it work. We didn’t presume to know.
I do think that there is real merit to putting some dedicated time of
experts into a formal R&D. We haven’t seen that that is a priority at this
moment, but we are going to keep talking to agencies and organizations and
seeing if someone wants to fund some formal collaborative R&D project.
DR. CAMPBELL: And time has moved on since this report was written, was it a
year and a half ago, about?
MS. KLOSS: Yes.
DR. CAMPBELL: One of the things that I am observing both in the U.K. and in
Australia is that the IHTSDO is becoming what that central terminology
authority is, and that the release centers are doing that. Both organizations
are working on how should terminology evolve gracefully over time.
The release centers in those countries are actively generating white papers
and reviewing proposals about how the structures are going for future releases
of terminology. The U.S. is not represented in a meaningful way there. We do
have delegates that attend. I am one of the people funded within the U.S. to
attend. Bob Dolan from Kaiser, there are many others. But outside of being able
to fund travel expenses to attend the meeting, you need to be able to spend two
or three weeks reviewing proposals and creating proposals and being able to
have a dialogue with them. Otherwise it is basically NHS and NIDA and New
Zealand and Canada will be the ones that have the input into this, not the U.S.
MS. GREENBERG: We have already fallen so far behind because of our lack of
moving to ICD-10.
DR. COHN: Mike, I’ll let you ask the last question, and then I’ll try to
wrap this one up.
DR. FITZMAURICE: I don’t have to be convinced that this is a public good
and that there is also private benefit, and therefore the gains should help
fund the costs of it. But what makes it a hard sell is, it is a vision at this
point. Is there a place where this vision works on a smaller scale than a
national scale? An example of a success story, a scaled-down model where you
have clinical data coded, aggregated and that serves to produce quality of care
or better patient outcomes or better population health measures or lower costs
or longitudinally meaningful data for research and quality improvement.
I’ll take just one of those and carry the example to people who can talk to
their bosses and make things happen.
DR. CAMPBELL: I’ll be so bold as to probably try and give a couple. One
that Simon mentioned, I do have some experience within the Kaiser health
system. This isn’t a SNOMED related example, but there was a point in time when
Kaiser was able to create a national pharmacy data warehouse, where they were
able to aggregate all of their prescription data across their entire enterprise
and be able to use it to identify patterns of prescribing, and be able to
intervene when they found outlying patterns that were traceable to pharmacy
representatives that were going around and making visits within the enterprise.
They were having quite an impact on the budget.
DR. FITZMAURICE: Did they tell people?
DR. CAMPBELL: Yes. So there is one example.
DR. OVERHAGE: You said there was a time. Does that mean it is not happening
now?
DR. CAMPBELL: No, it is just that I am not currently an advisor to Kaiser,
that’s all. I am also not an official spokesman for them as well, so Simon may
have to modify anything that I may say in that regard.
One of the greatest areas of demonstrated — and you can go to the journal
articles and look these up — areas of clinical benefit from coded terminology
is drug-drug and drug-disease interaction prevention, and adverse outcomes.
A few years ago I went and looked at the data there, and it is quite
impressive, the return on investment for putting in a pharmacy information
system that can prevent drug-drug interactions. I wrote a paper on that a
number of years ago, talking a little bit about the Kaiser experience when I
was part of the Kaiser activity.
So there are examples out there.
DR. FITZMAURICE: In reading those articles, I don’t get a sense that there
is an awful lot of data work that had to be done on ontology building, on
making sure everybody used the same terminology. But that was probably part of
it, right?
DR. CAMPBELL: For the pharmacy data warehouse they had to do work to
normalize two or three different — Molten and First Databank I think were the
two pharmacy information system terminologies that Kaiser used. The pharmacy
data warehouse group had to figure out how to normalize those data. I think
they relied on the NDC codes as an intermediary in order to do that. They had
to then have a standard representation that they mapped everything to, and then
they were able to result in a useful end product.
DR. FITZMAURICE: It is those nuts and bolts, here is the beginning, here is
the end and here is what was accomplished. It is hard to write up in an article
or someplace what you had to do in the middle of it. Yet that is what this
large national organization would do, is help to do it for everyone, so that
nobody has to suffer that big expense. They glom onto something that we would
hope would be as free as possible to make things work.
MS. KLOSS: It is like other parts of our health IT agenda. We have
incalculable costs going into the administrative complexity of doing things the
way we are right now. There is a great opportunity there also to see new tools
developing and streamlining some of the business processes.
DR. STEINDEL: Mike, I do believe there are a lot of national models that
are evolving that are starting to produce evidence, like in New Zealand,
Denmark, Netherlands, et cetera, that are very similar to this.
DR. FITZMAURICE: Maybe they would fund our stuff.
DR. STEINDEL: You can request it.
DR. FITZMAURICE: But your point is also, maybe we can learn from their
models too, and I take that.
DR. COHN: Maybe to wrap this section up, I do want to thank all three of
you. I think it has been a very useful overview.
Linda, you are absolutely right, I think in the context as we think towards
21st Century Vision for Health Statistics, this is an important — and has been
somewhat neglected recently — piece of the whole thing. It is probably time to
shine a light back on it. But it is also exciting to hear the pieces that are
starting to fall in place. I think ICD-11 sounds like a really key and critical
aspect. Maybe one of these other foundations will move things forward.
I think we have all observed that the implementation of clinical
terminologies has lagged everyone’s expectations. I think probably part of it,
at least my own view has always been that if you live in a world where you need
to report ICD and CPT for all the reasons that you need to do that, and you
have a clinical terminology that isn’t connected, or requires great effort to
map, and maybe that mapping is not to the point where it has been sanctioned by
the government or otherwise, it becomes a major barrier to the introduction of
all this stuff.
So we are beginning to see all these pieces beginning to be thought of
seriously, and potentially it is time to put all this together. So we will
continue the conversation.
We are going to finish up here very quickly. Why don’t you stay sitting
where you are, since we are going to be done in just a couple of minutes.
For the full committee, this is the end of our meeting for our February
meeting. Our next full committee meeting is May 21-22. You all are aware of the
change from June to May for that one. It sounds like we are having some
hearings of the Population Subcommittee right before that, so those of you who
want to spend a full week in Washington as part of this conversation, it is
springtime in Washington.
As always, I just want to open up the floor for comments by the committee
members of things that we should be doing better or differently or whatever as
we plan the May meeting. The May meeting will be a time for celebration.
Hopefully we will have new members identified at that point. But that will be
an opportunity to celebrate and reflect on the good work we have been doing in
the last couple of years.
But any comments about this meeting, what you think we ought to be doing in
May or otherwise? The floor is open. I think you all are helping me get this
meeting done on schedule here, because it is 12:01. Oh, Larry.
DR. GREEN: Simon, I was quite taken with Jim’s comments about positioning
the May meeting and the September meeting. If I heard you right, Jim, you used
the word preparatory.
I would think in terms of the function of the committee and the next couple
of meetings that we would do well to frame our preparation for those meetings,
the agenda, the way we spend our time with that being our modus operandi for a
couple of meetings, to make transitions in membership and to prepare.
I found that to be a galvanizing statement this morning that helps me think
about the work of the committee for the next six or eight months. What do you
think?
DR. COHN: I think the executive subcommittee is charged with coming up with
an overall plan that relates to 21st century statistics. How we execute that
plan I think goes along the lines of what you are describing. There may be a
fair amount of preparatory work as we create that.
We all know that the work of the committee typically from an idea to an
actual report is anywhere from — except in extraordinary circumstances that we
probably prefer not to replicate is usually a 12- to 16-month process. So I
don’t think we want to come out with a major report in September, but hopefully
we will be moving forward with all of that, and talking about it very seriously
in the May time frame. Make sense? Melding all the ideas together.
MS. GREENBERG: And we obviously will need to have at least one
teleconference of the executive subcommittee probably, too. But we will go back
and digest all of this, but it is going to be a busy spring.
I think that all fits in, but I think what I am hearing from Larry, and we
will have to discuss it, we have various items that have been on our agenda to
have presentations, and we have deferred them or whatever. We should see, do
they fit into these high priorities or do we do our regular updates but then
focus more on what we have been discussing these two days. I think that is what
you are suggesting.
DR. GREEN: It is. For example, we have agreed, and we have got plans for
May 19 and 20 for hearing about the medical home. It seems to me it would be
very important to put the agenda for that hearing through the filter of how it
is going to help us prepare for 21st century statistics, as an example. It is
not just another hearing. It needs to fit in with this line of thinking, in my
view. It is part of preparation, it is preparing.
Does that make sense?
MS. GREENBERG: It definitely does. I think looking at the quality report
that we talked about yesterday, too, how does it fit, what aspects of it fit.
DR. COHN: With that, I want to thank you all very much. The meeting is
adjourned.
(Whereupon, the meeting adjourned at 12:05 p.m.)