Hubert H. Humphrey Building
200 Independence Avenue, S.W.
Washington,
D.C.
TABLE OF CONTENTS
Page
Call to Order - Dr. Detmer 1
Review and Approval of Charges, Work Plans and Recommendations 4
Panel on Conceptual Framework for Coding and Classification 54
DON DETMER, MD, Chair, University of Virginia, Charlottesville, Virginia
JAMES SCANLON, HHS Executive Staff Director, Office of Asst Secretary for Planning and Evaluation, DHHS, Washington, DC
MARJORIE GREENBERG, Acting Executive Secretary, NCHS, CDC, Hyattsville, MD
HORTENSIA AMARO, PhD, Boston University School of Public Health, Boston, MA
CARLOS ARCE, PhD, NuStats, Inc., Austin, Texas
SIMON COHN, MD, MPH, FACP, Kaiser Permanente Medical Care Program, Oakland, California
KATHRYN COLTIN, Harvard Pilgrim Health Care, Brookline, MA
KATHERINE FRAWLEY, JD, MS, RRA, American Health Information Management Association, Washington, DC
ROBERT GELLMAN, JD, Washington, DC
RICHARD HARDING, MD, Richland Memorial Hsptl, Columbia, SC
LISA IEZZONI, MD, MS, Harvard Medical School, Boston, MA
THOMAS LA VEIST, PhD, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, Maryland
SHEILA LEATHERMAN, United HealthCare Corporation, Minnetonka, Minnesota
JOHN LUMPKIN, MD, MPH, Illinois Department of Public Health, Springfield, Illinois
CLEMENT JOSEPH MC DONALD, MD, Regenstrief Institute, Indianapolis, Indiana
VINCENT MOR, PhD, Brown University, Providence, Rhode Island
BARBARA STARFIELD, MD, School of Hygiene/Public Health, The Johns Hopkins University, Baltimore, Maryland
GEORGE VONAMBURG, Michigan Dept of Public Health, Lansing MI
ELIZABETH WARD, Washington State Department of Health, Olympia, Washington
P R O C E E D I N G S (9:00 a.m.)
Agenda Item: Call to Order.
DR. DETMER: Good morning. It is nice to see everyone. I am particularly happy to see Jeff Blair here.
Jeff, we had heard that you were going to be recommended by Ways and Means to join this group.
There were other names that I must admit we had heard, but we were just delighted that some white smoke came out of the chimney. It is great to have you here.
Obviously we will have to do all the processing of papers and such, because this is very new news. We just literally found this out from Jeff this morning. So, welcome.
I don't know if you want to make any comments. If you would like to, that would be fine.
MR. BLAIR: Just that I am proud and delighted to be a member of the committee. Thank you.
DR. DETMER: It is good to have you on board. I think I mentioned Dr. Richard Harding yesterday as being at the first formal meeting. I must admit I didn't know if we were going to make it to get the other appointee in time for this meeting.
I am Don Detmer. We usually start these meetings with introductions. I am from the University of Virginia. I think we will go around the room with introductions and get
on to our agenda for the day.
MR. SCANLON: I am Jim Scanlon from the HHS Office of Planning and Evaluation. I am executive director of the committee.
MS. COLTIN: I am Kathy Coltin. I am from Harvard Pilgrim Health Care in Boston.
MR. SCHWARTZ: Harvey Schwartz, Agency for Health Care Policy and Research.
DR. AMARO: Hortensio Amaro, Boston University School of Public Health.
DR. HARDING: Richard Harding. I am a child psychiatrist from South Carolina.
MR. VAN AMBURG: George Van Amburg, Michigan Department of Public Health.
DR. IEZZONI: Lisa Iezzoni, Division of General Medicine and Primary Care, Beth Israel Hospital, Boston.
DR. STARFIELD: Barbara Starfield, Johns Hopkins University.
MR. GELLMAN: Bob Gellman, Privacy and Information Policy Consultant.
MR. BLAIR: Jeff Blair, IBM.
MS. FRAWLEY: Kathleen Frawley, The American Health Information Management Association.
DR. SONDIK: Ed Sondik, National Center for Health Statistics.
DR. LUMPKIN: John Lumpkin, Illinois Department of Public Health.
MS. GREENBERG: Marjorie Greenberg, National Center for Health Statistics, and acting executive secretary for the committee.
MR. MEYER: Jeff Meyer, HBO and Company.
MR. ETTINGER: Stan Ettinger, AHCPR.
MS. RUDOLPH: Debbie Rudolph -- (portion off microphone).
MR. SICORNING: Frank Sicorning, Blue Cross and Blue Shield Association.
MR. KALODNER: Rob Kalodner, Department of Veterans Affairs.
MS. CLINE: Kathleen Cline, Health Insurance Association of America.
MS. VERCICI: Jean Vercici, American Medical Association.
MR. EMERY: Jack Emery, American Medical Association.
MR. STEINBERG: Tom Steinberg, American Dental Association.
MR. DEUTSCH: Larry Deutsch, Children's Health Network in Connecticut.
MS. JAMES: Gail James, CDC, Atlanta.
MS. GILLMAN: Jean Gillman, CDC, Atlanta.
MS. BRONSTEIN: Lisa Bronstein, private citizen.
MR. VELLMAN: Mike Vellman -- (off microphone).
(Other introductions made off microphone.)
DR. DETMER: Okay, thank you very much. What we started on a bit yesterday and then worked on in subcommittees were where the charges, work plans and recommendations were from our various subcommittees.
I think what we want to do is just have a brief report on where the status of each of those are.
I don't know really what you have to put on the table necessarily, but we will go down through the heads of those groups and we will hear from you.
Agenda Item: Review and Approval of Charges, Work Plans and Recommendations.
DR. STARFIELD: I have nothing additional to report from the subcommittee on health data needs, standards and security. As you will recall, the committee accepted the charge yesterday, the charge to the subcommittee.
DR. DETMER: I will call on Mr. Gellman and then come back, because we do have a fair amount of work to report on the other committee.
MR. GELLMAN: I think the order of business for the subcommittee on privacy is to develop some recommendations for everybody to look at over the next month or so.
I don't know that we have made a decision about the timing of things. Are we going to set this up to be voted on at the June meeting, or I don't know. It is clear we are going to have to do whatever we do shortly.
DR. DETMER: In fact, the executive committee this morning -- this is something I need your feedback on as a group as a whole -- felt that what we ought to do is have the subcommittees work on getting their recommendations thought through and actually pass those to the full committee through mail and e mail.
If things look like we are all totally comfortable with what is on there, to agree. If not, let me know so that we can actually look at the date currently set and see if we need to have something more timely, if it requires that.
I think that from my perspective at least that should work. We will see. We clearly don't obviously want to approve recommendations that the whole group isn't comfortable with.
At the same time we realize that if we can be more efficient, that would be in the public's interest as well. So, if that sounds reasonable to people. Okay, it looks like it is.
I am going to call on Dr. Iezzoni to talk about essentially what is a renamed committee. The new name for this will be population specific issues, the subcommittee on population specific issues.
They worked into the night to get a draft.
DR. IEZZONI: Here it is. I am seeing it for the first time.
DR. DETMER: Voila, like a rabbit popping its head out of a hole, it is still warm. We believe in timely business.
I don't care, if you want me to read this or if you care to, but basically this is the charge. I would like to have you hear this.
We have it in hard copy. Let's go through it and see if we are comfortable.
DR. IEZZONI: Tom, would it be possible, while Lynette distributes it, to maybe have a brief discussion about what we talked about with Dr. Sondik yesterday during that meeting, to give it a little time?
DR. DETMER: I think that is a wonderful thing. Let me set the stage for that conversation. Ed, correct me if I truncate this or add to it, please.
The discussion was about the National Center for Vital and Health Statistics and its agenda. The discussion really came down to the way in which this committee initially as a committee as a whole might work with Dr. Sondik to look at the national center, both where it has been and where it is headed.
Through a process of a few meetings, come back then to this group with thoughts on continued advice to the national center.
The anticipation would be that once that process is complete, we would then look at it again as a body as a whole and decide, do some of these pieces fit into our current subcommittee structure.
What of it needs to be clearly at the full committee level, if that is the way it looks. Where is additional outside expertise.
So, in fact, this group could serve, to the extent possible, as a good, solid advisor to the health and well being of the center.
Have I heard you accurately? Please weigh in on this.
DR. SONDIK: That is exactly it. We do not have a formal advisory body. Given the issues and possible corrections that we think it could take, we very much need the input of a formal advisory body.
This is the most appropriate group and I hope we can do it.
DR. DETMER: I appreciate that you came to us. Is there comment or discussion, Barbara?
DR. STARFIELD: Yes. I think it is worth reiterating that the committee is fully committed to dealing with population concerns, whole population concerns. We will not let those slip in the press of responding to the new legislation.
DR. DETMER: I think obviously the committee's history has been very dominantly in that area. So, it is certainly something that we have interest and expertise in.
I am glad, actually, that we can relate to this.
DR. IEZZONI: We have renamed our subcommittee. It is now the subcommittee on population specific issues. The basic kind of purpose of the subcommittee, however, is similar to what was in the draft charge that you have in your blue book.
Perhaps we could just go through kind of the first paragraph, which sets the stage for what we are interested in.
Basically, this subcommittee addresses the fact that a number of health issues might affect different types of people within our society differently, in terms of their ability to obtain health care and have access to health care.
These issues include mental health and substance abuse, chronic illness, impairments. We are concerned that each of these affect privacy and confidentiality and even standard setting issues that we are dealing with in some of the other subcommittees.
This subcommittee will be the place where these issues will be focused upon.
We are also concerned that gender, race and ethnicity interact with all of these kinds of issues, to raise a whole other series of concerns.
So, the subcommittee on population-specific issues will identify priorities, strategies and opportunities for developing multi-purpose and integrated and shared data, and ensure that these concerns are addressed throughout the other kinds of issues that the entire NCVHS is dealing with.
The way we have chosen to structure this particular charge is more around tasks, rather than kind of general statements of principle.
The first draft that you saw had general statements of principle. This is more oriented toward tasks, although obviously, in developing the work plan we are going to have to be even more specific than we are here.
As you will be able to see from looking at the document -- those of you who have it and those of you who don't -- I will read it for you.
The first task will be basically to revisit this issue of the core data elements project that the full committee dealt with, to revisit some of the issues that the full committee dealt with before that might pertain specifically to the populations that I have just described.
It will also to be to bring to the table the fact that we have in process core data elements projects dealing with mental health and substance abuse and, at a more preliminary stage, long-term care and disability.
So, it is just trying to reach full circle on those efforts and trying to propose to the full committee where we might want to go with those.
The second thing is to really help NCHS think about methodological issues surrounding the population based surveys that NCHS does, and to make recommendations around those, as well as kind of monitoring other population based survey issues in other departments and in private sector.
The third task is to review some of the department's new initiatives that we think might have a particular effect on these specific populations.
The one we have identified that we are particularly interested in looking at is the performance -- the public health performance measures, the performance partnership grants -- we need to have that word in here -- and the measurement piece of the performance partnership grants, and to make recommendations to assure the inclusion of issues pertaining to the populations I have just described.
The fourth task is to try to identify research and development questions pertaining to these special populations, and make recommendations around them.
Then the fifth task is to work jointly with the two subcommittees of the national committee that are focusing primarily on K2 issues at this time, and work with them to make sure that the issues for these specific populations are fully aired; let's put it that way.
In fact, we thought that one of the first things that we would propose is sponsor joint hearings with each of the two other subcommittees to make sure, for example, that privacy and confidentiality issues for these specific populations are fully addressed.
Also on the standard setting side, we would make sure that issues raised by these specific populations around the health identifier, around the coding set kind of issues that we are going to be talking about today, are also fully considered.
That is just perhaps not as brief as I could have been.
DR. DETMER: That is fine. In fact, what I would like to do is open it up for discussion. Then before we move on, I want to come back to task five and propose what I would like to do relating to it.
DR. LUMPKIN: I would just like to add perhaps one thing, a refinement. There are issues related to providers who serve special populations.
The question that perhaps your committee could get to that we haven't addressed in the K2 is, are those providers disproportionately affected by the kinds of standardization or simplification processes, which would disproportionately affect their ability to provide care.
DR. DETMER: Excellent.
MR. VAN AMBURG: I have one suggestion. That is to include under task two something with regard to looking at geographic allocation of data, which is not there.
There is a typo in the first paragraph, undeserved medical areas. It is an interesting concept.
DR. IEZZONI: There are several typos in there, and Hortensia and I will jointly share responsibility for that. Part of it is my typing, or lack of skills.
DR. DETMER: What a difference a letter makes. I am glad you read carefully.
MR. SCHWARTZ: I would just like to suggest one possible refinement. Perhaps in the ninth line in the first paragraph we might consider adding a phrase such as, persons receiving substandard health care or inadequate health care.
Quality of health care might be an issue that would be important for population specific groups. It might be a new initiative for the department that might be of interest as well.
Rather than just saying, inadequate health insurance, perhaps adding a phrase, persons receiving substandard or perhaps inadequate health care.
DR. STARFIELD: Is that an issue for specific populations or is that a general thing?
MR. SCHWARTZ: Let me try to address that by saying, within the context of the statement here, the reference, examples of populations of interest include.
I am just suggesting a refinement might be to say that a population of interest that might be included in this might be people receiving inadequate or substandard health care.
DR. IEZZONI: I think that is going to be hard to operationalize, Harvey. That is my primary concern with that. I see some heads nodding around the table.
DR. LUMPKIN: I am nodding with you.
DR. DETMER: Let's go ahead and discuss this more.
DR. SONDIK: I think it is very important, though, to have the notion of health care quality in there. I see that as a very significant issue in the future. I guess it doesn't strike me as that difficult to operationalize.
DR. DETMER: Where it could go is in that first sentence. We could say access to quality health care. That is one way of doing it. I don't know if that gets at your point.
MR. SCHWARTZ: May I suggest an alternative? I am not wedded to suggesting that it belongs in the general charge. Perhaps in task three, the quality initiative could be mentioned as one of the new initiatives, then as an example, such as with public health performance measures, survey integration, and quality of health care initiatives.
DR. LUMPKIN: I guess my difficulty on this issue is that quality is a cross cutting issue. So, in a sense, it really belongs to the committee as a whole and should be something that would be part of everything that we are doing in all the subcommittees.
I am a little bit concerned about saying that there is a -- that it is an issue that creates a population that requires special attention, while I accept the concept that populations that have been identified as needing special attention, that quality is an issue for them.
On the sixth line it says quality concerns. As far as I am concerned, that covers it.
DR. IEZZONI: I thought we had it in there. Thanks for pointing that out.
MR. SCHWARTZ: I am sorry. I didn't see that.
DR. DETMER: That is great. It would seem like we had gone through this yesterday.
DR. IEZZONI: Yes, it was just that it was hiding there.
DR. VAN AMBURG: I have one minor thing on task three and that is a rewording of the end of the sentence. I really think that should read, to ensure that population specific issues are addressed, rather than the way it is worded. I think that is the objective.
DR. DETMER: Are you comfortable with that?
DR. IEZZONI: Absolutely. This is hot off the press and any other editorial suggestions are welcome.
MS. GOLDEN: May I comment on the issue of quality? In some of the previous versions of this particular task, in trying to identify populations of concern, I think we did zero in on people that we considered those receiving less than adequate health care.
The issue became not whether that was a special population, but whether we could operationalize the term quality health care.
In a large setting, that might be one of the things that we want to consider. I know we have had a lot of difficulty to come up with turning that into variables in the data system that captured any notion of that.
From a data standpoint, I think the sentence we have handles that. If we ever decide what it is, then it might follow that those not receiving it do become a population of special concern.
DR. DETMER: Any comment?
DR. IEZZONI: Thank you, Pat, for that comment.
DR. DETMER: Other questions? I really am very grateful that we have this. I wanted to mention the reason that Dr. Iezzoni is doing this is that Dr. LaViest, for personal family meetings, had to resign from the committee.
So, I chaired the meeting yesterday and she has kindly agreed to chair this committee from this point forward.
What I would like to do, I think we word smithed this to some extent, but I think it is quite good. What I would like to entertain is a motion to accept and approve this as a charge and task statement.
We can always obviously address this further if we need to, but this gets us underway. I don't know if somebody would like to make that motion.
MR. VAN AMBURG: So move.
DR. AMARO: Second.
DR. DETMER: Any further discussion?
All in favor say aye. Opposed. Abstentions.
(Motion carried by voice vote.)
DR. DETMER: I also want to acknowledge Hortensia's work on that as well. I think this is really good. I feel really good about it.
Okay, I think that I said I wanted to come back to task five for a moment. What I want to propose on that is that Lisa and Mr. Gellman and Dr. Lumpkin get together and look at how these issues have been intersected.
I think that I would propose that we consider having a hearing on the west coast, trying to actually deal with perhaps both sets of issues.
We can actually see if we could then get this accomplished, again fairly soon. Otherwise, I think we have heard from a very good round of folks. But that way we can make sure that we have done justice to do that. What are reactions to that?
DR. IEZZONI: Can I just ask a matter of timing? When you said west coast yesterday, Don, I thought this might actually be a very good topic for the west coast.
One of the things that we are concerned about on identifier issues, for example, are hospitals that care for a large number of people who may not have Social Security numbers.
I think that being on the west coast, and border states would have more access to that kind of hearing and it would be very important.
Can I ask, though, a matter of timing? I understood from your earlier comments that you think that by the June meeting we will at least have some preliminary recommendations from Bob's efforts.
DR. DETMER: Well, we hope to actually, from both subcommittees.
DR. IEZZONI: I just feel that it is too rushed to think about getting together a hearing before the June meeting, given all the other work that everybody has to do.
DR. DETMER: I think we have got to. Unfortunately, we have got a time line. I think what we need to do is do it, and do it literally as soon as we can get that put together.
DR. IEZZONI: What I was wondering was, would it be useful to have the recommendations in hand and have the hearings be for people to react to the specific recommendations.
DR. DETMER: We clearly will need, I think, to put things out for reaction. I just see that as more hazardous. I think I would rather make sure that we at least try to cover the net before we start --
DR. IEZZONI: All right, so it needs to be before June.
DR. DETMER: Right. So, we will work on the dates. I acknowledge that it is a tight schedule but I think that is really the best way to go.
DR. IEZZONI: Okay, it is good to know what the parameters are.
DR. DETMER: Obviously, folks, I want to make a comment. For all of those Americans who live between the shining seas, we could also meet in many other places between here and the west coast.
At least I think we will make one shot at least trying to accommodate a broader input.
MS. COLTIN: I just wanted to think about the April 15 and 16 meeting of the K2 subcommittee. We are going to be talking about coding issues.
I know that historically there have been a number of coding issues relating to some of the special populations as well.
I don't know if there are particular agenda items that you might like to suggest for that, or people to testify. We might want to think about what could be accommodated in that time period.
DR. IEZZONI: I think that there might be a really full agenda for the coding meeting. Probably by the end of this morning we will feel that it could be even more full. But we should certainly consider that, Kathy.
DR. DETMER: What I am anticipating is that this will be a hearing, really, sort of led by the population specific group with input from these other two subcommittees.
I think that whatever the leaders of those two subcommittees believe would be appropriate on that and meet your satisfaction, we would go forward on them.
Okay, I think at this point we have actually -- are there other discussion items that I am missing as far as some of these work plans and recommendations?
I don't know if we have our speakers here yet for our next panel. What I think I may do is, let's move on down to what we were going to talk about at 1:30 and do just a little bit more committee meeting work.
We have talked a bit about this anyway. Then we will, at the right time, come back to our panel at the break.
What we intend to do is have, unfortunately from some respects, a fairly heavy work load, trying to meet these dates this summer.
The concept is that we are going to look at -- well, we have our meeting scheduled actually in June. Those dates are what. Remind me, please, the 24th and 25th. Those dates are set and we intend to have that meeting then.
We are going to have, as we say, additional hearings, and then this April 15 and 16 hearing as well. We hope to then try to draft recommendations.
The executive committee hopes to have a meeting the first week of June to try to lay out, essentially, our work plan for the rest of the year as well.
DR. MOR: I had the 10th and 11th in mind.
DR. DETMER: Okay, I am glad we are discussing this. That is good. As far as agenda items, we have heard from a couple of groups, of at least some presentations they would like to have at that meeting.
John, I think you had one that you might tell the group a little bit about that.
DR. LUMPKIN: The K2 work group has become aware that there are a couple of states that have attempted, in some measure, to do some data setting standards. I think Minnesota was one, for instance.
We thought that we would have a presentation from two to three states on their experience in attempting to do that.
DR. DETMER: Then if I heard it right, Lisa, it seemed to me like I was hearing some need for some discussion on also some race and ethnicity identifiers and such.
DR. IEZZONI: Yes. OMB will, around that time, be coming out with their proposal, which is very hot politically, about how we might want to capture race and ethnicity for government reporting in the future.
In the past the national committee has always responded to OMB. OMB will basically, I guess, be having a written comment period or something.
We at least would like to have the opportunity to discuss this and come up with some language that would be part of the written response to OMB from the national committee.
DR. DETMER: Okay, questions and comments on either one of these two items. I think it is a very tight window, as I understand the issue. We will just have to again move as efficiently as we can to provide our understanding and out input.
DR. IEZZONI: Don, what we are trying to do is organize a meeting of the subcommittee that will hear the issues before the June meeting.
When we come to the June meeting, we might not know at the time of the subcommittee meeting what the OMB recommendations are.
We will be at least up to speed on the background. So, we will be able to move fairly efficiently on making the recommendations.
DR. AMARO: I just wanted to say that it is a very important issue, has major implications for the validity of race and ethnicity data and the completeness of it, which then has implications for what cities get left out, because of incomplete data and things like that.
So, I think it is worthy of getting everybody's input on this.
DR. DETMER: Are there other topics that we know we will want to have, or certainly make every effort to have a part of that June meeting.
I anticipate that we will also, by that point, I hope, have some draft recommendations from both of the K2 groups. Are there other items that you know of?
MR. SCHWARTZ: I think Lisa has mentioned some of them already. Activity continues in the area of public health performance measures and program indicators and that sort of thing.
I think that is a subcommittee issue, if I am understanding.
DR. COHN: I was just going to sort of generally comment that considering the time frame that we are dealing with regarding some of the recommendations, I would urge the committee not to set a tight agenda at the June meeting, so that we really have some time to discuss and approve whatever recommendations are forthcoming.
Dr. DETMER: I think that is a good idea, particularly if we are going to try, to the extent that we can, to pass out recommendations and see if people are comfortable with them short of the meeting, I think, just to have a chance to fully air those. I think that is a very wise idea.
Okay, are there other items that we know we want on it? I guess any further conversation, actually, relative to the April 15-16 hearing.
DR. MOR: Do we have an agenda for that meeting yet? Has it been sent to us?
MS. GREENBERG: An agenda for the meeting? No, we were going to do some planning after.
DR. DETMER: That is part of what I am asking for.
MS. GREENBERG: I have prepared something on the questions that we were going to pose. I thought there was some desire to discuss this after the presentation.
DR. DETMER: I see. Let's get into it, because I think of our schedule. Our colleagues aren't here yet. We could maybe get into this.
DR. MOR: Since we really haven't heard from -- at least I haven't -- an innovative delivery system or a system, a health care system that has a multiplicity of different classes of providers in it, it would be interesting to hear their strategies on dealing with information sharing, coding, or the complexities that they are encountering in integrating alternative ID systems, as well as information transfer.
DR. LUMPKIN: Let me see if I understand. You would be talking, let's say, like a large managed care entity that has within its scope, let's say, some staff model components, group model components and so forth, so that they would have a fairly wide integrated one.
DR. MOR: Actually, I wasn't just thinking about payers and physicians and hospitals. I was thinking about other kinds of providers as well.
DR. LUMPKIN: I think that the kind of thing you are describing, most large entities include not only physicians, physicians who work with nurse practitioners, other care providers.
Many of the large groups, for example, Blue Cross in Illinois has a product where they cover non-traditional healers and so forth.
I think those kinds of things, where you are talking about an entity that really covers a broad spectrum of packages, that get into all the different kinds of providers and the arrangements that they are in.
DR. MOR: I am not thinking exclusively from an insurance perspective, but rather a delivery system perspective; that is, where there is a need to share clinical information, not just payer information.
DR. COHN: I need a bit of clarification. Are you referring specifically to classification issues or unique health identifier issues or a combination?
DR. MOR: Both. They are both, I believe, important issues when transmitting information across an integrated delivery system that owns home health agencies or is just acquiring home health agencies or nursing homes or rehab hospitals.
Issues of ID are relevant. Clearly issues of the nature of information transmitted and the mechanisms used to transmit that information, assuring confidentiality, et cetera, I just don't believe we have heard that.
I think that is, to some extent, at the heart of some of these issues.
DR. STARFIELD: I think I know what Vince means. Maybe you could give an example, to make sure that everybody knows what you mean.
DR. MOR: Okay. When an individual is hospitalized and then sent to another post-acute provider, if that post-acute provider is part of the hospital ownership structure, however that might be defined, there are a variety of different issues associated with the transmission of information, particularly as that delivery system has to begin reporting information aggregated for all the lives they serve and that they provide services to.
So, it is their rehab hospital or their nursing home or their home health agency. The information -- they didn't start from scratch with those. They bought them or they merged with them, and they are struggling with issues of ID and information transfer.
How they are establishing those, I think, is germane to understanding what the implications of our recommendations will be in the real world.
MS. COLTIN: Let me tell you an example. I think I can relate to this. This isn't even an issue of having gone through four mergers and trying to link information. That is a whole other story.
Take for example, in just one of our division where we operate 14 health centers, each of them had their own bank of medical record numbers to assign to members who elected that health center.
Now, they were continuous and non-overlapping. However, each time a member changed their membership from one health center to another, they got a new medical record number.
We had to develop mechanisms to try to link those, so that one could track care across a period of time for that person.
The other thing that you end up with is, on the insurance side, a contract number, which is generally the subscriber's social security number or some other unique number for the subscriber, with a dependency code added on.
Now if, in fact, you have got a family where the husband changes jobs and now it is more attractive to have the coverage through the wife, now the wife becomes the subscriber.
You have got a new contract number and everybody is a dependent under her. Now you have got to figure out how to link them and recognize that it is the same person, because you have got multiple lines of registration and they have all got different numbers.
There is an identifier issue for sure when you are trying to deal with these multiple systems.
The other area that I think Vince was talking about is things like when we want to look at our rate of cardiac catheterizations or cholecystectomies, and some of them are performed in the outpatient arena and they are coded in CPT and some of them are performed in the inpatient setting and they are coded in ICD-9, we are trying to bring that information together.
Any time you have a service that can be delivered in alternative settings and can be coded differently in each of those settings, and you are trying to get an aggregate picture of what is going on, you run up against these kinds of limitations.
Is that what you are getting at?
DR. MOR: As you get to other layers which don't have their established coding systems, it becomes even more real.
DR. COHN: I guess my comment is -- like Heather, I understand the issues. As we become a less integrated and more of almost a disintegrated health care system, I guess like the rest of the world, we are having these problems.
What I am hearing is an issue that probably needs to be addressed during some hearings. But my sense is that it was probably not the April 15 and 16 set of codes and classifications issue, but as a more advanced layer as we begin to look at, for example, all the transaction standards, looking at them along with health identifiers against the various types of information that is really needed in the real world beyond just billing.
I think that is really what you are beginning to talk about. We are all painfully aware that we are talking about using these transactions, but we keep talking about, gee, we aren't just billing any more.
I think that there is a real -- you have to sort of begin to look there.
I don't think we are far enough through the analysis part of the transaction pieces right now, to have I think, to really begin to look at that. But I think it will be soon. I guess that was my response to that.
DR. LUMPKIN: First of all, I am still not sure I am clear on the issue that is being raised. The first thing that came to mind is my colleague to my left.
There is an entity that Simon is involved with that owns hospitals, owns some things, and then has some other arrangements for providing care.
So, they are both an owner and a provider of a very diverse network, and there are other examples of that.
It is difficult. I just can't conceive of a model that would not be a bearer of risk, that is that complex. Once an entity gets into owning things -- you know, in multiple delivery settings they generally are involved in accepting some component of risk or some involvement of risk in that service delivery.
I do agree that it is an issue. I am just wondering if that is within the scope of what we are directed to do.
In a way, we are talking about the transactions that occur between entities. Now, in the things that they don't own, where they are contracting with a group, I see that as being under the purview of K2.
If they are having transactions between a hospital that they own and an outpatient center that they own, those transactions are internal transactions.
Maybe if someone knows K2 better than me -- I am not sure that is a transaction that falls within the scope, because they can have their own internal mechanisms.
DR. STARFIELD: I think if you are going to deal with clinical care, you can't code procedures two different ways, if you are going to follow clinical care.
I think what you are saying is that we need to have somebody who can address those issues.
MR. MOORE: The standards were not intended to change the legacy systems that exist within the organizations. We are not looking to say that those standards that are set for communications between two parties are the same as those communications that are within the organization itself.
As long as they can convert what they have to what the standard is to transmit it to another entity, then that meets the standard requirement.
DR. MOR: May I just respond to that? If a company decides that it is going to make a decision as to whether to acquire or whether to contract an associate based upon the impediments associated with the regulatory requirements of transaction standards, you are going to be influencing the way in which the system expands and contracts amongst providers, because of the set of transaction standards that are being constructed.
If one company decides to buy another company, to acquire two more hospitals as opposed to contract with them, and that decision is influenced by the perceived impediments of having to change their transaction standards, then that will begin to affect the way in which these companies make decisions about moving together.
MR. MOORE: Doesn't that already exist? Kathy says she has the problem already with four different mergers where the internal systems were all different. That didn't stop them from getting another.
MS. COLTIN: We see standards as a way to save us.
MR. MOR: But the purpose is for internal purposes. They are all going to be under one umbrella. So, there is no regulation on how they are going to communicate internally.
They are looking to this as a way of streamlining the internal processes, not just the external transaction processes.
DR. LUMPKIN: I agree with you. I think we have to be very careful in the message that we are sending. There is a tremendous impact upon administrative simplification, which was raised yesterday at our meeting.
One of the things we have to consider is cost. If everybody is on a different system and we move to one system, there will be a cost.
If we also tell people that we think our purview extends into internal transactions, they may choose to follow these standards internally, but the mandate does not extend into that arena.
I think we just have to be careful that we at least, while we may discuss the issue, where the regulatory boundary is that is being addressed here.
DR. DETMER: I guess as a committee, if you choose, you could get considered opinion on how big an issue would that be, if this committee were to say, well, let's do recommend, not as a matter of law, but we see it as desirable. I don't know.
MR. MOORE: We have asked our office of general counsel to give us a legal opinion as to how far these standards go into the organization, because the issue is already there.
If a provider is dealing with a clearinghouse, does the transaction between the provider and the clearinghouse have to be in a standard format.
Our unofficial position is no, because that is an agent of the provider and they have contracted with that agent to really turn what data they have into some standard type of format, to make it uniform.
The same thing could be true of a payer or a health plan who chose not to deal with the external world themselves, but to hire a clearinghouse or some other entity to do that for them, and to give them the data in the format that they wish.
DR. DETMER: Clearly, obviously, how the department chooses to look at this doesn't necessarily constrain what our recommendations might be. It is very germane.
MR. MOORE: I was just trying to add some more information.
DR. DETMER: I think that is very useful.
DR. MC DONALD: As someone who has struggled with standards for about 12 years now, I just warn a little bit of caution.
One can say, for example, that one would standardize on one coding system, and that would only have to be from the door out.
You have to go to look how deeply bound certain coding systems are into the sort of interstices of the institution.
In our institution, the surgeons know the CPT-4 codes. They write them down as numbers, not as names. Switching on that would be a pretty big upheaval.
I just think we ought to be sensitive to the fact that sometimes, although we can define a border, the interface between the border and the inside might be really disruptive.
If we can make big steps, I think we should be gloriously happy. If we get everything perfect, I think we will be shocked -- we may be gloriously happy, too.
DR. COHN: I am just going to make one other comment. I consider this to be sort of brainstorming right now.
I think there is a very important piece that Dr. Mor is bringing up that I think needs to be pursued, and that has to do with as these things become more concrete, we probably need to be back revisiting and looking at integrated health care entities and capitation and other environments to see how well they fit in with those environments.
I think that right now my own feeling is that we need to go off and think about how we might pursue this issue. It may be that there is a small group of people that ought to essentially work with Vince to see if we can flesh out something that we can bring back to the committee.
DR. DETMER: Clearly, I think in fact one of the things we have handed out here -- I think we should also spend some time going down through this.
Clearly, that is part of what we are doing, is a preliminary discussion relating to planning that hearing. This may be one of the items as well as, obviously, the others.
We have just handed this out. Why don't we give people are few minutes to look down through this and then we will talk about it.
DR. STARFIELD: What is this, the questions? What are we looking at.
DR. DETMER: These are some questions that have already been proposed by subcommittee members and the staff.
The question is, if you try to put all of these in the Federal Register notice, it ends up being fairly lengthy.
I think the question is, can we get some cut at this. Now obviously we want to hear from our panel as part of what was going to help inform our deliberation of this. We will also revisit this after that discussion.
DR. STARFIELD: This is the same as we got by e mail a few days ago; right?
DR. DETMER: Yes, this is hard copy of something that you have seen.
DR. MC DONALD: Mr. Chairman, would it be possible to get from the department a list of the code systems that are currently used in the transactions?
I mean, there is a focus on a pretty well defined set of transactions.
MR. MOORE: You mean, give you the ICD-9 codes or just -- you said a list.
DR. MC DONALD: Not a list of the codes themselves, but a list of code systems that are used currently.
MR. MOORE: That is being developed now, as they are coming about, and we could make that available.
PARTICIPANT: The beginnings of a draft of the inventory that we were developing as part of the coding classification implementation team, was given to the subcommittee at the February meeting.
Again, it is still being developed. We still have not collected all of the information. So, we hope to have that available in time, possibly, for the April meeting or just before the April meeting.
MR. MOORE: Could we give them what it is as of some point in time, be it April the 1st or whatever, even if we don't feel it is complete?
PARTICIPANT: We could do that.
MR. MC DONALD: Particularly what I was asking out, I think there is really only a handful of sort of code systems, and then there are the local ones, and just kind of orient us.
MS. GREENBERG: Obviously we want to know who to invite to the hearings specifically, although I mean everyone will be welcome to attend. We need to invite the developers and users of those systems.
DR. MC DONALD: The other question was, I noticed in the documents that there is an ICD-10-CM. I had been under the impression that there wouldn't be one. Is there one? I guess there isn't one. Everybody looks shocked.
MS. GREENBERG: There is.
DR. MC DONALD: Is there a way to get a look at what that looks like?
MS. GREENBERG: It is not going to be available by the April meeting. Donna, do you want to just give us a little update on the status of the clinical modification of ICD-10?
MS. PICKETT: In terms of the status of the clinical modification of ICD-10, a contract was awarded in 1994, and was completed first quarter of 1996.
Based on the recommendations that were contained in the final report and additional input that we have had through the ICD-9-CM coordination and maintenance committee, staff at NCHS has done additional work in refining the clinical modification of ICD-10.
As part of the contract that was awarded, a prototype of ICD-10-CM was developed. We have continued as staff at NCHS to further enhance that prototype.
We currently have certain chapters out for review by specialty societies. We have at least 10 or so who have reviewed the prototype. We are working their comments in.
We still have several chapters or subsections of chapters out for review. We expect those to come in. That is an ongoing process, but we hope to have that finalized within the next month.
DR. MC DONALD: For some further clarification, the ICD-9-CM included a number of developments, including procedure codes, if I understand correctly. ICD-10-PCS is the procedure codes side.
MS. PICKETT: That is correct.
DR. MC DONALD: So, could you describe sort of the content elaborations that are in the CM part. Is it like toxins and accidents? What sort of things?
MS. PICKETT: The 10-CM is building on what was developed by WHO as part of ICD-10. What we have done is gone back to proposals that have been made at previous ICD-9-CM coordination and maintenance committee meetings, to look at those specific issues where the recommendation was a good recommendation. However, because of the limitations of ICD-9-CM, we could not incorporate those proposals.
So, we are now looking at ICD-10-CM as a way of addressing some of those concepts that could not be included in ICD-0-CM.
DR. MC DONALD: I am still not getting a feel for whether this looks like an elephant or a mouse.
MS. PICKETT: ICD-10 basically has the pretty much the same look and feel of ICD-9 and ICD-9-CM. Within ICD-10, WHO has added a number of new categories or further expanded certain categories.
Part of that we were going to cover as part of probably the April hearings. It will probably be a little bit more detail than we could probably cover adequately here.
Certainly we could make some information available to the committee that gives some of the background work on what we have been doing and kind of where we have been heading.
DR. MC DONALD: So, ICD-10-CM won't be copyright?
MS. PICKETT: It will be in the public domain in the United States.
DR. DETMER: Betsy, did you and Patricia want to weigh in on this as well?
MS. HUMPHRIES: Just to say, as Donna has said, the ICD-10-CM covers the disease part of the diagnosis side of 9-CM, and then the PCS is covering the procedure part, which was previously in 9-CM.
MS. BROOKS: I think, though, that once we decide what we are going forward with -- let's say we are going forward with the ICD-10 -- we wouldn't have the initials NPCS hanging off the end.
If we made a decision to go with ICD diagnosis and a procedure classification, we would probably call the whole book ICD-10-CM, just like we called the whole book the ICD-9-CM.
The United States developed the procedure classification under contract, too, with ICD-9.
DR. DETMER: Is there any further discussion on this issue? If not, we will go on.
DR. LUMPKIN: We were reviewing the questions. I would just like to throw out, that of this list of 12, I would cull it down to, my list had five.
That would be number four. If we asked what codes and classifications do you recommend as initial standards for administrative transactions, given the time frames in HPPA(?), what specific questions would you like to see implemented regarding coding classifications.
I think if we asked that question, then the more specific questions such as one and two, which are what codes are you using and can one system work, would be answered by their answer to four. I think it is a broader question.
MS. STARFIELD: Can I just clarify that one? The word administrative transactions means to me billing. That doesn't encompass what is in one. So, I just think we ought to hold that for discussion.
DR. DETMER: He said two and three, as I heard him.
DR. STARFIELD: We can discuss that later.
DR. DETMER: We are trying to shape this toward that hearing.
DR. LUMPKIN: Let me think about one. We can come back to that. The point is, one is comparing the administrative and the clinical. I think that may be a very valid question to add to the mix.
DR. IEZZONI: Just as a general principle, I think the hearings we have had thus far, John, have been very informative and great.
I think that they could have been improved if we had had more specific questions. I think we should have general questions, but I am not at all averse to having some specific content to the questions.
I don't think we should have as a principle right now that we want to cull out the specific questions.
DR. LUMPKIN: I think that the task before us is that this list is a little bit too large.
DR. IEZZONI: Twelve is too many; I agree.
MS. GREENBERG: We have 14 here.
DR. IEZZONI: Or 14 is too many.
DR. LUMPKIN: Then I have one that is not on there. But I think my thought would be that we would give them -- ask them to address the specific general questions in their opening comments, ask them to submit written comments about the whole list in our letter.
Then allow us to make our questions, and even encourage them to send it to us ahead of time, if they wanted to.
Then our comments can be in response to looking at what their written, specific --
DR. DETMER: We can put this whole list actually up on our web site, too, as far as that goes. That is no problem.
DR. LUMPKIN: It would show them what we are interested in, although the specific questions that we would ask them to answer would be the more selective list.
Seven, which is recognizing the goal of HPPA as administrative simplification. How, from your perspective, would you treat/deal with current coding environment to improve simplification, reduce the burden, and also obtain meaningful information.
Twelve, because I have -- we can perhaps think about 12. I would put it on as a candidate pending the presentation of the panel.
I have a lot of concern about us recommending a proprietary coding scheme versus one that would be in the public domain.
DR. IEZZONI: Can you clarify what you mean by proprietary, give us an example?
DR. LUMPKIN: I think we heard some testimony in earlier sessions about individual companies who had to purchase coding schemes. I can't remember which one it was.
DR. IEZZONI: It was CPT. That is why I wanted to clarify.
DR. COHN: Let me make a comment, though, that there are many coding systems -- I mean, I don't think I know of any coding system that is really in the public domain.
MS. PICKETT: Well, ICD.
DR. COHN: Is it truly in the public domain? I thought it had a copyright.
MS. IEZZONI: And the alpha numeric KIKPIX(?) is in the public domain, too.
DR. LUMPKIN: See, I think there is a difference between a copyright and a system that costs money. I am not saying that we may not come down to that, but I do think that we need to explore the differences.
The difference is that -- you see this on a lot of systems with share ware, which is that people have the right to copy it but not to change it. Copyright protects the system from change. That is why I wondered a public or a purchased system.
DR. MC DONALD: For clarification, that is not public domain. I think we have to be real careful about these lines.
I think the goal would be to have something that would be very modest or pro bono or cheap, because there may be distribution costs, and not able to grow wildly, and that can be defined.
You have to be very careful. Proprietary is to me blurry. PCRI is not going to be used for ISO. It is not public domain. It is free.
DR. DETMER: There is a lever where they are either one or the other. I think part of this is probably definitional.
DR. MC DONALD: HCFA used to sell theirs for $600.
DR. LUMPKIN: Maybe in that regard we are talking about some cost benefit analysis. That is what we are trying to get to.
That is not to say there should be a cost, but evaluate the benefits that may accrue from a particular coding system.
MR. MOORE: I think along with that cost, though, you have to consider how it is going to be maintained and expanded and developed over the coming years as changes take place.
I know within the government with ICD-9 and all we have difficulties at times maintaining the right resources to keep things current for the public to use. You have that to wrestle with.
DR. DETMER: Very few things are free.
MR. MOORE: But if you are going to recommend public domain, make sure that --
DR. DETMER: This is a very germane public policy discussion. I would certainly think this ought to be on this list. We will come back and again and revisit these after the folks are here.
DR. LUMPKIN: Don't interpret my comment as saying it has to be public domain. I am only saying it is an issue that has been raised in hearings and one that we ought to discuss and have discussed before we make recommendations.
DR. IEZZONI: I agree. I just wanted to have the language of this question be clear to readers. It is not entirely clear to me right now that it would be clear to readers.
DR. LUMPKIN: So, you need to word smith it based on this discussion.
DR. MOR: I would actually like to make sure that the issue of maintenance, of who owns the system even if it is in the public domain, who is responsible for maintaining it and updating it thereafter is on the agenda.
I mean, who should do it? If we are going to make recommendations, then who is going to be around to keep them up.
DR. DETMER: Any more on that, understanding that we are going to continue to get the discussion a little bit, and then we will re-come at this again.
DR. LUMPKIN: My next question would be then 14, which has been a standard question that we have been asking that says, essentially, how do you feel about the process that is ongoing.
Then the question that is not on the list is -- and it needs some working, so I just throw out the concept -- which is, if the code says adopt it, should all the codes be available to all the users.
That was an issue that came up a number of times in our prior -- let me explain what that means.
There are certain code sets for which payers will say, you can use IDX coding system, but you can't use the following codes because your specialty is surgery on the left foot and these are the codes for surgery on the right foot.
DR. MC DONALD: I don't know that that is our domain. What you are saying is that we are going to set policy for the usage of all kinds of these codes. I don't know that we could figure that out.
I mean, that would mean that they would be published, but I am not allowed to do certain procedures in my hospital, even though I can read the codes. I couldn't report out certain procedures. I am not supposed to.
DR. IEZZONI: John, wasn't the issue that some insurers wouldn't pay for certain codes if they were submitted for payment?
MS. FRAWLEY: The problem is that there are payers who don't recognize the official coding guidelines. That is the major issue, that the cooperating parties for years have been promulgating official coding guidelines.
We have payers all over the country that do not recognize those coding guidelines. Therefore, people are manipulating data in order to get reimbursed. So, I think that is a critical issues.
DR. LUMPKIN: What it leads to is people doing one procedure and coding another.
DR. DETMER: That is not that rare, unfortunately, because of the limitations of the coding. I am not talking about a willful interest in deceit.
MS. HUMPHRIES: If I hear that the interest here is that, because of, I don't know, internal system preference or they don't want to recode or whatever they want, some groups are saying, yes, we understand these are the standard codes, but you can only use 70 percent of them or something.
It sort of seems to me that in a way the law deals with here is the coding system, so it would have to be in the implementation guidelines in general, that once these coding systems are established, that it isn't going to be local preference among major segments of the health care system, that they are going to implement this 20 percent and I am going to implement a different 50.
MS. FRAWLEY: The other thing that I just wanted to mention when John was running down his list of questions -- I know we will talk about this more after the panel presentation -- I think number five is still a very critical question.
Simon and I had raised this at one of our hearings. The industry has been gearing up, you know, for a conversion to new classification systems, in terms of ICD-9-CM and ICD-10-PCS.
Now we have all the requirements now of HPPA facing us. I think that is something that is certainly at the hearings on the 15th and 16th.
I would hate to see that question not one of the top ones up there. I think that is really critical. I am sure we will hear more from Pat and Donna on that.
There has been a tremendous amount of development work ongoing with the possibility of getting ready for some field testing and evaluation. I think we need to have some better sense --
DR. LUMPKIN: Could we perhaps finesse that, because my concern is that it kind of really says -- number four says, what classifications do you recommend.
If we don't recommend ICD-10, then we have a problem. We could say -- we say in number four, what specific suggestions would you like to see implemented regarding code reclassifications. For instance, if we use ICD-10, should we use --
For those people who do comment that they want to see ICD-10 -- and maybe that will be everybody. That will be off base, but if they use ICD-10, then do you think it should be the whole boat.
DR. STARFIELD: How are they going to know? They haven't seen it. That is the problem.
MS. HUMPHRIES: I will confess to having drafted this question from the point of view of the implementation team.
We are sitting here and saying, I think the department definitely needs guidance on this issue. A lot of time and effort and resources and people out there -- there are numbers of people in the field that are waiting for these things to come out and do their tests and evaluations of them.
This whole train started out of the station before the law was passed. Given the timing of this hearing, even though maybe within a couple months time of the hearing a number of people who might testify might have had quite a lot of chance to look at these, as it happens, they won't have.
So, I don't know whether we either decide that we have to have another hearing specifically on this.
DR. DETMER: No, it sounds to me like that is why we are talking about it. I think you have made a good argument.
DR. MC DONALD: A slightly different domain. Firstly, I don't know how to say it as a question or how to get it brought out.
There is the question when we switch from an ICD-9 to an ICD-10, if the messages and the coding systems aren't set up right, there is always that confusion period where both codes are coming in.
There are ways, and I think the EXPO specification has it, where you require that the coding system was included along with the codes and the messages.
The other part of it is, there is a history -- I don't know if there are any code systems that still use it -- of reusing codes, which I think is the highest sin in all this business.
If there is some way that we could stop thinking that way in the next era, it would make it a lot easier for people who are actually using the data to manage to build to whatever.
That business about the codes being reused, I am not sure if any systems do that any more, but there is a long history of it.
DR. COHN: I think what you are asking, Clement, really what I would almost ask as a more generic question. Now, obviously as we look at these 15 questions what I am aware of is that certain questions probably are appropriate for certain groups.
The ICD-10 question is probably not a question that you should ask of other code developers. It is really more a question of users of code systems.
Similarly, the question that I am sort of thinking you are saying, which is how sort of how would your code system work as part of an overall strategy for codes and classifications within the United States, which is really sort of the question you are really beginning to ask. It is really a question that should be asked of developers.
DR. MC DONALD: I really need to ask them, do you reuse your codes.
DR. COHN: I think that is part of the overall strategy, of all the things that need to be handled, how does yours work in.
DR. MC DONALD: Maybe no one does, so that is moot.
DR. STARFIELD: I would hate to guarantee that no one does that.
MS. GREENBERG: I just want to point out that there was a question here on maintenance and somehow it fell off this compilation.
DR. STARFIELD: So, it should be on here?
MS. GREENBERG: Yes. I don't know. It fell off the typewriter.
DR. DETMER: I think why don't we do this. Chris Chute, apparently through a bad snow storm, was delayed. The report is that he is on his way in from the airport, so he should be here momentarily.
We want to take a break anyway and this seems like a little lull We will be coming back to this. So, let's reassemble in 20 minutes.
(Brief recess.)
DR. DETMER: We wanted to convene a panel to give us a conceptual framework for coding and classification. As I said, Dr. Chute is actually on his way. I think he will be coming very shortly.
Chris is at the Mayo Clinic and would have been our lead speaker.
Jim Cimino is also here, from Columbia. Dr. Cimino also has a long history. We have one of his papers in our book and another I guess on its way around.
Carol Weigel, I understand, was actually asked to speak first. Welcome back, by the way, to the committee. She was in charge of this piece of our work when she was on the committee, and it is good to you have you with us as well.
What we will do is hear from Carol and then Jim and hopefully Chris will be here by then. We have already seen our discussants earlier and we will get their reactions and thoughts as well. So, Carol, nice to have you here.
Agenda Item: Panel on Conceptual Framework for Coding and Classification.
MS. WEIGEL: I wonder if one of the reasons I am here is to prove to everybody that the way people sounded in Fargo is really not how we sound in Minnesota.
Well, as Don suggested, a number of years ago I served on the committee as well. I chaired the medical classification subcommittee.
Many of the issues that I heard discussed yesterday were a whole lot of what we talked about back then, too.
So, let me just start with a series of overheads here and get everybody sort of cranked up.
Okay, change is inevitable. There is no question that where we are today in trying to deal with all the things that we have to do with, we can't do with it with all of the old stuff.
Now, I am going to give you a quotation from another person who is an alumni of the national committee, and that is Dr. Carter White. Many of you, I suspect, know him.
The issue that he is talking about here says, the unwritten social contract between health care professionals and societies they serve, accord clinicians the power to label the diverse manifestations of ill health encountered.
This awesome power has enormous implications. The farther we go on in history, the more we realize this.
Clearly the issue of how the physician labels a patient or labels a diagnosis is directly written in the medical record.
What we have to deal with is the piece of that that becomes coded information at some point. At the risk of being overly gratuitous, let's just go through a real quick version of what it is codes do for everybody.
Clearly, it is the whole purview of what the committee has talked about all day yesterday, in fact, what they have talked about throughout all its history.
We use codes, obviously, to bill for care, to compile morbidity and mortality statistics. We compare the health of our nation to that of others, Lisa's population considerations and so forth, studying effects of drugs and treatment modalities, improve the quality of care, epidemics, other important sorts of epidemiology kinds of things, we build facilities, we develop health care budgets and health care policies.
Why do we bother to use codes at all? There are several good reasons for that. Number one, they are obviously more compact than a narrative. They are a single kind of thing we can look at and deal with.
Sometimes they are even more specific than the information that they are trying to code themselves. That is the sort of an issue where, as you have a piece of terminology that is not very specific, but because in a grouped classification it tends to fit with something that is more specific, it in fact becomes so.
Obviously the last piece, the codes are cheaper to handle oftentimes than what are the other narrative.
The major question that the committee obviously has to grapple with here is, does one size fit all. I think the answer probably very clearly is no.
The second issue is, can one size fit all, and very likely the answer to that is an equivalent negative.
What we need to deal with, very clearly again, is what you have talked about in the last day or so here, and what has been the focus of the conversation in terms of the committee's agenda.
As you look at the K2 issues, the biggest issues clearly are one that talks about the cost and second, the quality.
What is the generalizability of one classification over another. Again, this is a very, very succinct kind of issue.
Part of the problem that we deal with is that we, as users, in whatever genre that falls, manage to mix up a whole series of different concepts.
There are things that are called nomenclatures. There are vocabularies. There are coding schemes and there are classifications.
These things are not all the same. Very often, though, we tend to use those terms as completely interchangeable concepts. That is a big piece of the problem in its own right, because we aren't talking off the same song sheet. I tend to talk off song sheets because I can't sing. I will spare you that.
Nomenclatures, again -- this is a short little primer here -- are the active process of naming something. They tend to be very precise.
I am old enough to remember having been taught how to use standard nomenclature of diseases and operations, in which one of the primary tenets is that you do not create a code.
If god didn't strike you dead, the nun that I worked for would have. So, you learn those kinds of things.
Then all of a sudden, as time marched on, you got into the international classification. There it was some easier way of doing that.
Clearly if we did those kinds of things, if we each created our own stuff, we had no way to compare our stuff with somebody else's stuff. Very obviously, again, that is where a large piece of the problem is that we are trying to represent here.
As we look down with a number of those things, again, it is the coding issues, the classification kinds of things, vocabularies.
All of those are different. Yet, we tend to assign some of similar kinds of properties to each of those things.
It is impossible these days to do a presentation without a Dilbert cartoon. One of the little functionaries there is talking to Catbird, who happens to be the company's HR person, I think.
Mr. Catbird, the company is trying to force me to use a different kind of computer. You are the human resources director. What are you doing to stop this religious persecution. Whatever happened to diversity.
To which his response is, the longer you verk here, da verse it gets. That is true a lot with classification as well.
We keep adding diversity to this. In fact, what we have done over the years as we have played with things like classifications, each one has had a set of different attributes, if you will.
What we have managed to do is say, gee, this one works really well for one purpose. As you look at, oftentimes, the distinction between what is the ICD-9-CM and what is CPT, a lot of those same kinds of things came out and came into the open.
It said, if you look at ICD-9-CM, it tends to work fairly well for defining what happens in a population for quality kinds of issues and so forth. If you looked at CPT, it didn't necessarily.
By the same token, if you tried to make ICD-9-CM fit all the reimbursement implications, it fell apart. So, instead of trying to use one versus the other or having the usual number of missionary zealots suggest that this one is far better than that one for whatever purpose, we tended to kind of hybridize them if you are a purist, or bastardize them if you are not, I think, to say, let's just borrow some of the kind of good stuff from this one and we will borrow some good stuff from this one over here. Between all of those things, we will come up with what we want.
Clearly, some of the kinds of issues that have happened with that, the discussions and the debate of which one was better, one versus the other, I am sure many of you can recall a publication -- it has been many years ago now -- but after the DRGs came out, there was a rather landmark publication that said, can the Medicare prospective payment system survive the ICD-9-CM classification system.
Indeed, we are looking at something that said, now we are using a classification for a purpose that it was never designed to fit.
That happens a whole lot and it is not something that happened just recently.
If we think back again in time to the prior classifications ago, where we had in this country two classifications that were being used at the same time -- one of those was the ICDA-8 and the other the HICDA -- that managed to coexist.
Part of the problem there, very clearly, was these were sort of designed to do kind of the same kinds of things but they didn't. They weren't the same.
All of a sudden we had data in two separate formats, which is what caused a whole series of difficult issues.
As we started looking at the collection of data, not just for DRGs, but prior to the time that DRGs arrived -- and Barbara you remember that, because you were part of that Institute of Medicine study that said, as we looked at the data, there was one in three of the diagnosis codes, or two out of the three that were better.
Sixty-five percent of diagnosis codes were considered accurate and 74 percent of the procedure codes were considered accurate.
Clearly, you can't design policy on data that is that flawed. So, what we said was, wow, there is a neat way of being able to deal with this. We are going to deal with this in a really big way. The answer, of course very obviously, is did it do that.
Well, maybe but maybe not. Here again we had a classification that was designed for one purpose, and that is as a statistical classification that tended to group like things together.
Then along came something else, and this was the Medicare prospective payment system, where suddenly a lot of other things got to be much more of a concern for folks.
We looked at DRGs, which one translation could be, damn, the revenue is gone. Suddenly people got excited about not just what happens with one little pigeonhole and what do I do with that, but how am I going to be paid.
If I do one thing it is going to say I get paid a whole lot more than if I do something else. So, the whole community of interest suddenly got huge from where it was a long time ago.
When I was on the committee I had the privilege of serving with Dr. Carmalt Jackson. He used to say that when you talked about codes, there were 100 people that showed up.
They were all the same 100 people. They might sit in a different place in the room, but basically they were the same 100 people.
Now that number has grown far beyond 100 to a much larger group, again, because the community of interest is far different.
When we looked at ICD-9 or we look at DRGs, suddenly we said, here we are again with some sort of hybridization problem, because now the classification doesn't really tell us what we want it to do.
We went through processes that began to add pieces of information, either in terms of severity kinds of indexes where people could legitimately argue, well, but my patients are older and sicker and a whole different population group than your patients are, so I need to be able to provide for that.
Again, how well did some of that work. One of the things was looking at severity. A second of the issues was looking at the things that we call data enhancement products.
Those are edits that are put into the system that say, if you use these two things together, this one has to be first, or if you use these two things you are going to generate more of a financial reimbursement.
Again, the world got far more interesting. This is a local bias cartoon. You can read the bottom there. It says, man, I miss the old days. There are two Vikings on a ship. They used to have slaves on it. Now they have little terminals. The issue that the guy is beating out of the end is input, input, input, a far cry from where we used to be.
Actually, the way the Superbowl turned out in the NFL is that we really should have had at least one she said on these guys, but just leave it as Minnesota Vikings and forget about it.
Now, into all of that arena the medical classification subcommittee, which is what the committee was called back then, sort of staked a claim, if you will, to say this is the committee which represents a public/private kind of opportunity, to begin to discuss the classification issues in an open forum where everybody knows kinds of what is going on.
What the medical classification subcommittee, with a whole lot of help from the federal sector in terms of NCHS and HCFA, in terms of the private sector, of the folks who came to give us testimony and presentation and the kinds of recommendations that we sat and worked out together was that the critical classification issues as we saw them were that the user or the coder is the key to the data quality equation.
I suspect that this is somewhat of a gratuitous comment as well. I will say that as we look at what the average coder is or what the average coder looks like, bear in mind that they tend to be fairly garden variety folks. Most of those garden varieties are home grown.
We are not necessarily talking about people that have a huge knowledge base. They are accustomed to looking for a word, finding the word, and being able to assign a code number to it. That is the person that we need to start with.
The second issue is here. It says, the classification must be flexible to respond to changes in technology and taxonomy.
In the old days, the old versions of ICD, they were in existence for 10 years. There was, again, an unwritten law or a written law that you did not tamper with the structure of that, between any of those 10 years during that revision cycle.
That is clearly not where things are at today, nor can they be. As we begin to change things, if we had left the structure of ICD-9-CM or ICD-9 inviolate, we would not have been able to track the major issue of the AIDS epidemic, for example, because there would have been no place to put that where someone could have actually got at the information.
Bear in mind, however, that as we create that flexibility, as all of you were talking about earlier this morning, every time we make those changes, that generates a cost within the whole structure.
The vendor community for the majority of hospitals and providers is really the one that provides the software to the majority of users.
These people need to be aware of what is going on. As I heard you talk before coffee this morning, where are we with ICD-10, where are we with ICD-10-CM, among the major and very legitimate conversations and arguments before, as we looked at the ICD-9, for example, is that there wasn't really a sufficient lead time in order to implement that in this country in the best way possible.
What we need to be doing is working with that now, because that is how we get that information out to folks. It is how people learn to use what is there.
Lastly, what we said is, as we look at medical classification subcommittee recommendations, they were that the coordination and maintenance committee, which is an ongoing process, had to be a functional component of both the public and the private sector. Indeed, it does represent that function for us.
The coding guidelines must be developed and approved in a public forum and they have to be transferable to all sites.
When you start talking about a rose is a rose is a rose, or another quotation that I learned from somebody who taught me plant taxonomy, which was a frightfully boring proposition, is that if the cat has kittens in the oven, you don't call them muffins.
So, no matter it is you are dealing with, no matter what site it is, it ought to have a common reference in terms of the code.
The integrity of the classification has to be maintained through administrative procedures. Somebody again has to control what happens to the code in relation to the consistency with which it is to be used.
We need to develop and interpret national coding guidelines so that we can get away from -- how can we force people, as Betsy was saying -- to use what is there. I don't know the answer to that either. Clearly, that is beyond my purview, fortunately.
Looking at interpreting national coding guidelines so that we don't have something that says -- in the great state of Minnesota we have more than snow storms. We do these kinds of things, too.
We disseminate information to all users. That, again, is a very critical piece and it means that it has to have that sort of public format and framework.
We conduct evaluation programs to maintain the accuracy, just as the Institute of Medicine study was done a long time ago.
Some of the things that we have done in relation to the use of ICD-9-CM and the other pieces that have come after that, we need to continue with those processes.
We need to receive and disseminate coding information for changes, modifications, and adjudication of disagreements.
If three or four parties are looking at something in a very different context or for a very different purpose, where do we finally assign the end result of that.
We need to continue to educate all of the users. Clearly that is coders, but it is also clinicians, but it is also researchers.
How do you know you are making the right kind of a decision based on the information that you are using. Within the classification framework, there are often some things that say, it is not necessarily an enormous increase in the disease process itself as much as it may be a classification effect.
Lisa has published a number of studies on the kinds of things that happened within the cardiac codes, as you changed the basic structure of what goes into the code content.
Suddenly you are looking at what could be a rampant epidemic. It really isn't. It is the classification that changed and that, then, changes your data.
Those are great big pieces of things that we need to be concerned about as well.
Lastly, going back to my one size fits all and a Peanuts cartoon, Linus is sitting there on his blanket and Snoopy is on top of it.
Linus said, I don't recall inviting you to share this blanket, but I must admit you are kind of warm and fuzzy. Snoopy's comment is, everybody brings something to the party.
I think that is entirely accurate. Basically what I am saying there, very simply, is that we all need to be talking together about what we are doing here because it is not a one size fits all thing. At least, we are not there yet.
That is a different argument, I think, too, but those are the kinds of issues that we dealt with as part of what was a medical classification subcommittee.
Those issues have not gone away. They are still there. I suspect they will always be there. That is the end of my comments.
DR. DETMER: Thank you very much. Actually, I enjoyed the content as well as the style. I had sort of been out of that part of the country for a few years, the Wisconsin Norwegian farmer.
That wasn't my line, the Wisconsin Norwegian farmer, who loved his wife of 50 years so much that he almost told her.
MS. WEIGEL: The way I heard the story was that he had been married for 50 years and one of his friends asked him what he was going to give his wife for the anniversary and he said, I am going to Norway.
They said, oh, wonderful, what are you going to do for your 75th anniversary. He said, I will send for her.
DR. DETMER: Questions or comment for Carol before we move on.
DR. MC DONALD: The process by which the guidelines for coding come up. I am always perplexed by the fact that when we have undefined hypertension, I would call it I think a code 9 or a 0 -- .9 because I don't know what it is.
It always gets coded benign essential hypertension because that is more specific, and that is by the guidelines.
I guess the real question is, is there a way to get sort of clinical input into the guidelines processes? That is not intuitive that is how you should code it, but that is how it is always coded.
MS. WEIGEL: That isn't so much a guideline as it is what the classification does to it. It says if you don't have any further specificity -- there are in many cases a default that says the most likely possibility is this and it just forces it into that.
DR. MC DONALD: There is a specific code in that area. It is hypertension not specified or something like that.
MS. WEIGEL: I guess I would throw that back to Pat or to Donna, the guideline developers, and let them respond to that.
MS. PICKETT: The guideline is not specific as to what is coded. The guideline really addresses more what is stated in the medical record.
So, the benign hypertension would not be the default. It would be the .9, which is the unspecified.
DR. DETMER: Other questions or comments? We will be obviously opening this up broadly later. So, Dr. Cimino.
DR. CIMINO: I didn't have a lot of chance to prepare for this. That is okay, because I look around the crowd and I realize this is a very diverse group.
Most of you probably don't know who I am. I am a medical informatician, primarily, and I work on controlled vocabularies for use in clinical systems at Columbia Presbyterian Medical Center.
I am kind of way down here in the basement working on the details of our own coding system, which we use for multiple purposes.
I think I was invited here to kind of share some of my experience or insights into the problem, especially in clinical coding, but maybe some general solutions that may be applicable to the kinds of tasks that you are going to be facing.
So, let me just give a quick overview of what I am going to be talking about in this 10 minutes or so.
I am first talk about multiple needs, which Carol has actually done. I am going to go over that briefly.
I am going to talk about capturing clinical data for multiple purposes. I am going to then talk about desired characteristics and then I am going to come back to the inadequacy of current systems, just to give a quick overview.
Then I am going to push my own agenda of a model for the future, the way I could see we could maybe try to tackle some of these problems.
Carol, your list was good, but it was missing one thing. It didn't have anything to do with taking care of patients.
It said improve the quality of care. By that I assume you mean, figure out what we should be doing so that the next time somebody comes in we can figure out what to do.
Where I work, we capture clinical data and code it -- we capture Ms. Jones' data so we can take care of Ms. Jones, not Ms. Jones' grandson.
So, that is a very important aspect from where I come from. That, I think, needs to be addressed by this committee. I hope it needs to be addressed. This is sort of our one chance to tackle this until Hale Bopp comes around again.
We are not going to get another chance. This is going to be a quantum leap forward, I suspect, in the next few years.
I would like electronic record keeping at the level of clinical care to be included in that, and I am sure a lot of people agree with that. Maybe some disagree. I would like to hear about that.
So, record keeping. Primarily what we do with paper, we write down what is wrong with a patient and the next day we go and read what is wrong with the patient, or communicating among clinicians who are taking care of the patient.
Then we want to do some other things like decision support. I define that broadly to be doing things like reviewing patient data in a different way than simply sitting down and looking at the record. So, creating a graph of patient data.
That is not simply reviewing what is in the chart. That is reusing the data in some form that is going to help clinicians make decisions.
Alerts and reminders are typically what people think of when they see decision support. That is alerting systems, such as Clem has and his system and our system and others.
Then access to on-line information resources, such as MedLine, electronic textbooks, that sort of thing. That is a decision support thing, too.
That helps clinicians decide what to do. We can link those things to electronic medical records using the coding systems.
Billing obviously and other financial considerations management, things that Carol has talked about, clinical research, epidemiology -- I define that kind of broadly; I am glad Chris isn't here -- I define that kind of broadly to include things like reporting to the World Health Organization and that sort of thing.
So, those are the multiple needs. I suspect there are more. Everybody has their own. Hopefully it is included in there.
The way we do things now is we collect patient data at the point of care and we put it in some kind of clinical repository. It may be paper. Right now it is getting to be more and more electronic.
Now when we want to create our financial data base or do our statistical reporting, we have to recode the patient data into a different data base.
If we want to do clinical research, we typically have a completely different data reporting process, because whatever is being done here is not appropriate for the particular research study. So, they have got to replicate a lot of work.
At least from where I sit the goal should be something like this. You collect the patient data once, it goes into a clinical repository, and then you can generate your research data base and your financial data base from that, or whatever other views you need, whatever your other purpose is.
I strongly believe this can be done. I am not suggesting that there is one coding system that fits all. I am suggesting that there are ways to translate among coding systems if you take the proper approach.
So, desired characteristics, I guess one or two of my papers were sent around and they reference other papers by me where I list characteristics.
The one that I guess is the late one is one that I presented at a meeting that Chris Chute chaired in Jacksonville, Florida. It is not a publication copy yet, but I thought it might be useful for this committee to see.
What I did was, I synthesized or attempted to synthesize the desired characteristics that medical informaticians or others have published in the medical informatics and medical literature, of general characteristics of nomenclatures, classifications, coding systems.
It is not simply this thing needs this term in it, but more, this thing needs a different way to say things or a completely different approach or it is missing general characteristics.
I tried to synthesize that. It came out to 12 Desiderata for the Twenty-First Century. I tried to have a catchy name for it so people would reference the article.
In there, some of these are from there, some of these are from an earlier paper. Some of the ones in that paper are really very technical and get down to a level of detail that I thought probably wouldn't appeal to this group. If it does, you can ask me in the open session.
Synonymy -- I am going to go into each of these in detail -- concept orientation, clear semantics, no limits on the number, depth, breadth of the terminology, multiple hierarchies, multiple granularity, sensible compositionality, which I will define, and interchangeable constructs.
So, synonymy, many names for the same construct is a good thing, as long as you know that they are synonyms. So, you don't want to have redundancy in your vocabulary, your terminology, your coding system. You don't want to have two codes that mean the same thing.
You would like to have lots of different names for those so different people can recognize the coded term from their own language.
In general, that is easy. It is easy to say this nomenclature has synonymy. We will just make lists of synonyms, approved synonyms. There is a lot of work in getting it done but it is easy technically to do.
Concept orientation. In concept orientation you have one term equals one meaning. So, a meaning and concept are the same thing.
There is no ambiguity. That is, the concept doesn't have two meanings. It also doesn't have less than one meaning. It has exactly one meaning.
It also implies that there is concept permanence. I think when I walked in you were talking about reusing codes.
Reusing codes is a real bad thing. If you are saying that the code is the term and that code corresponds to the concept, if you reuse the code, then you have changed the meaning of that code.
Now that code has two meanings or it has an old meaning or a new meaning. That is generally a bad thing, especially for computers to try to cope with.
Concept permanence is an easy thing to say. It is not so easy to do, because of other constraints. If you run out of codes and you have to reuse one, what do you do.
If you are going to update your vocabulary, how are you going to do that in a way that maintains this concept oriented flavor.
For instance, if you want to change the name of a concept, how do you do that without changing the meaning, or do you change the meaning.
I am not going to try to answer all those things in 10 minutes, but we can do it in discussion.
Clear semantics. Not only does the term have to have one meaning, but we have to know what it is. It has to be clear to the users of the terminology. Those are the people who have coded it and the people who use the information that has been coded.
Everybody has to understand what it really means and how it is supposed to be used. So, penicillin, is penicillin a chemical or is it a pill. In some terminologies they are considered the same thing.
Not to me. To me, a penicillin pill is something that has a size and a shape and a dose and made by somebody and I buy it someplace.
Penicillin the chemical is something that is the composition of that pill. I can do some very interesting things if I model it that way.
For instance, I can add new pills all the time, but I don't have to change my chemical classification scheme.
What about gunshot? Well, you can have an injury and if it is not clear, if the semantics of that term are not clear, then somebody might use the same code for both an act and an injury.
Gunshot in the atrium, is that an injury or an event that caused an injury?
DR. DETMER: Colonel Mustard.
DR. CIMINO: That is right. I also had, my uncle shot an elephant in his pajamas. How he got in his pajamas, I will never know.
The notion there is, the semantics of the first statement are ambiguous. Everybody assumes the one thing, but then when you see it the other way you realize that there are multiple interpretations and the computer can't tell the difference.
No limit on the depth, breadth of terms, the number. So, don't pick a coding scheme like ICD-9 has that says you can only have 10 at this level, and if you run out you have got to move it somewhere else. Don't limit your depth of your level.
Be careful, as we look at how we start to do this, don't let the computer how deep you can go in your hierarchy or how broad.
As soon as you set that, somebody will come along and say, I need another one at this level, or I need another one at this depth.
Multiple hierarchies, this is maybe a fairly controversial one to a lot of people. Medicine, itself, though, is multi-hierarchic. We look at it by organ system. We look at it by etiology.
So, you know, pneumonia, does that go into the lung disease or the infectious disease. Somebody has got to decide that if you have a strict hierarchy. If you have multiple hierarchies, you can have both.
The macro level versus the micro level, we classify diseases. What is the site of disease of diabetes. Is it the pancreas? Is it the small vessels in the nerves and other organs? How do we define it, or is it the whole body? We can look at diseases at lots of different levels.
Each user of the data will want their own view. So, the person who collects the data, the clinician who initiates this data -- this includes people who generate data in ancillary systems, the laboratory and pharmacy and so on -- they have one view of that that they need for their purposes.
Of course, the abstractors need a different view and each abstractor may need a different view. In research they need it one way and the World Health Organization needs it another way and so on.
Multiple granularity. I think this is something that Clem might have been getting at a little bit, that users are going to want different levels of granularity than the coders want.
In general, the clinical users will want very fine grained terms. In the laboratory they say, this is a serum fasting glucose, or this is a post-prandial plasma glucose.
The epidemiologists and the payers might want to say, how many lab tests were done. They want a completely different level of granularity.
Then decision support and researchers want it somewhere in between. So, if I am writing a rule that says, you know, send an alert if the blood sugar is too low, I don't want to say if the serum blood sugar or the plasma blood sugar or the fasting, I just say, if any blood sugar is measured and it is less than 40, send an alert to somebody. I don't care how that blood test was done.
So, there are all these different levels. Again, this is where the one size fits all problem comes in. You can't have one size that is going to satisfy all of those.
I think rather than try to force one or ignore the others, we have to find ways to accommodate them all.
Okay, sensible compositionality. This is the one I said I would define a little bit. The first thing is, it doesn't matter how finely detailed you make your terminology, the users will always find another level that they want to go below that.
You can have every possible fracture you know -- you know, a right compound fracture of some bone with spiral fracture, green stick or whatever. Then suddenly somebody says, yes, but I want to say there is a secondary infection with that. To me, that is an important level of detail.
We need a way to define codes, put them together, rather than simply ignore that those data are there.
ICD-9, for instance, has not elsewhere classified which just means, they wanted to add something. We don't have a way to code it, but we want to indicate that they said something else.
It is not not otherwise specified. There is a modifier. We just don't have a way to code it. We need a way to code that information. It is important.
Equally important we need rules for combining them. We can't let people just slam them together in a haphazard way. We have to have a sensible way to do that.
One of the reasons we need that is so that we can recognize when synonymy occurs. If two people construct a code or code something with these compositional constructs, we can recognize that they are synonymous or different as appropriate. We can recognize how to classify them and so on.
Just a quick example, we might have one nomenclature where we have pneumonia in left lower lobe. So, some person is coding pneumonia and then says, well, I want to say it occurred in the left lower lobe.
So, we have a control terminology for the disease and we have a control terminology for the anatomy, so let's just combine them together.
We want to make sure that we have a sensible way to do that so that we can recognize that the concept left lower lobe pneumonia, if that appears in a scheme, there is a way to understand that this is a pneumonia that has a site in left lower lobe and is participating in a finding which has to do with context, something I haven't really touched on yet.
This I feel very strongly about. No more not elsewhere classified. We can't have codes for everything else.
Those are only useful for the one person that decided how they wanted to view the world. If they say, I really care about this pneumonia, this pneumonia, this pneumonia, and then if there are other ones, just put them under other because I don't really care about that.
The next person that comes along says, yes, I want other too. I want some that are lumped under your other classified, but the ones you care about, I want to lump them under other.
Look at ICD-9 and SNOMED. They both have -- SNOMED-II has whatever, pneumonia not elsewhere classified. But if you look at the other pneumonias that are classified, you can see that those terms are not synonymous.
There are ways -- I presented this in Jacksonville and the classifier types jumped down my throat. Hell, you can't do it, it is heretical. Especially the Europeans, they want to have a pigeonhole for everything.
I kept saying, this is the Twenty-First Century. We have computers and we have technical ways that we can code patients data without having to use an other code.
The example I love is LOINK vocabulary where you have rules for constructing your terminology for laboratory tests.
If there is a code for your test, great. If there isn't, there is a way to construct a code for that, so that somebody can come along later and figure out where they want to group that code.
We can talk about that, if people in the United States feel strongly about that. Certainly the Europeans were all over me about that.
So, inadequacy of current systems, I don't think I have time to go over all of these. The ones I have gone over, in general, ICD-9 doesn't fit any of those.
From what I have seen of OCD-10, I don't believe ICD-10 keeps up with those either. The one in particular, concept oriented, if you reuse a code, you are not concept oriented.
If you change a name to change the meaning, you are not concept oriented. That is why there is a minus sign there.
SNOMED has got a lot of plus signs in there. It is working to add more. I am on the SNOMED editorial board as a consultant. I know the direction they are taking. It seems that SNOMED is moving in the right direction.
CPT, whatever else you can say, there are no limits on the code. It does have a coding system that doesn't impose limits on hierarchy if it has one. Otherwise, it is pretty miserable in terms of some of these other formal constructs that we need in control vocabulary.
READ vocabulary, other than that they spell everything wrong, they do pretty well in terms of sticking to concept orientation, they have multiple hierarchies, they have different levels of granularity and they allow compositionality.
They don't have formal grammar for composing these things as far as I know. They don't have clear semantics as far as I know. It has been hard to tell from the samples of the vocabulary and reading their papers just how far we can go with that.
The Gabrielli nomenclature I have listed. There are others that I haven't attempted to evaluate. This one comes up a lot because it is big and there are claims that it covers everything in control vocabulary.
When you look at the structure, the synonymy is sort of there, except that if you look at the synonyms you might not agree that all the things that are identified as synonyms of a concept are really synonyms.
It is not clear to me if it is concept oriented or has clear semantics, because it is just this big hierarchy. You get it in a book or an electronic form and there it is. I have no understanding how it is maintained, so I can't answer those.
It does not have a multiple hierarchy. It has a strict coding system, and so on.
By the way, synonymy and ICD-9, if you look at the book ICD-9 has synonyms. But they are not in electronic form, or at least not available to me in electronic form, and they are not a formal nomenclature for synonyms.
So, I gave it a minus sign, but I could be persuaded to put a plus sign there, if somebody really felt bad that ICD-9 couldn't even do as well as CPT.
So, here is my view of how we get around a lot of this stuff, and other people have made these statements, too. These are not just my ideas.
The first one is meaningless codes. We have got to find another way to say that, Betsy. Meaningless just has such a bad connotation. Well, if it is meaningless, who needs it.
By meaningless, we mean there are no semantics in the code itself. So, you don't look at the code and go, oh, it is a disease, or oh, it is a third level in a hierarchy or oh, it belongs under some other code because that code has one digit less.
You get into trouble. You can't do multiple hierarchies if you have a hierarchical coding system. There are lots of problems that you run into.
A semantic network, as opposed to a simple hierarchy or collection of terms, there have to be relationships among the terms and they have to be semantic.
That is, they have to be labeled in ways that give meaning, that tell you how they are related. So, you don't just say that this pneumonia is related to this organism, because your computer won't know whether the organism causes pneumonia or the organism gets pneumonia.
So, you have to have ways to label that. The computer can help you with some of the tasks, including maintenance, which is why I have been mostly involved in semantic networks. It helps me with the maintenance problem.
A formal grammar, which ties into formal semantics for representing terms, it helps you with compositionality.
What a lot of it boils down to is the inclusion of definitional information. When you put pneumococcal pneumonia in there, say that pneumococcus causes it and link that to another code for pneumococcus and link it there.
Say that it occurs in the lung. If you don't want to say that it occurs in the alveoli of the lung, okay, but have some kind of definition that helps people understand what that is.
It helps the computers understand, so that they can help with the coding and the translation. If you want to add a term to the vocabulary and you define it, the vocabulary system may be able to tell you, hey, this thing is already in here. It has got this other name. They are synonymous, or this is where you should put it.
If you have got multiple hierarchies and you put it in one place, the computer can say, hey, you know, based on your description it really belongs in a bunch of other places.
So, either put it there or fix the definition, because your definition implies something that maybe you don't want to imply.
Finally, reject these other terms. Computationally there are ways to handle not elsewhere classified that don't require that we pigeonhole things into these other garbage cans.
That is all I have. I will take questions now, if I can.
DR. DETMER: Yes, we will take questions now and then we will move on and come back.
DR. MC DONALD: There is another way to define thing that is at least conceivable. IUPAC -- the International Union of Applied and something Chemistry -- has taken the position that you name the code produced by somebody who has it attached to something.
Their best example is the ATTC is the organization that actually keep cultures. They keep organisms and tissues. They have shelves of them. If you want to write to them, you can actually get those bugs and infect everyone with them if you like.
They use those numbers, so that that is defined by that thing that is on a tray. Are there other areas that could be done or that could be used?
DR. CIMINO: There are probably some areas where we could probably punt to some group that is charge of some area that we don't want to get into controversy in.
For instance, in my own vocabulary we have lab tests and we have drugs and we also have chemicals.
Now, the laboratory doesn't give me a list of chemicals. I assign a chemical or a body part or whatever it is that the test measures, I assign that to that test. LOINK does the same thing.
I don't get into the business of deciding, are these chemicals the same or different. I don't try to come up with vocabularies. I choose them from UMLS, because it is convenient.
Certainly you could say there is a group that is in charge of chemicals and let them decide how they want to do it and we will just use those for representing whatever we want, but we will let them in charge of that.
If they are reasonable and not changing the names of concepts and behaving themselves, I think we can divvy up some of those domains so we don't have to try to do everything.
DR. MC DONALD: The chemical abstract society -- CAS -- they actually do that.
DR. CIMINO: The Chemical Registry.
DR. MOR: From the point of view of administrative simplifications, then, if I understood your recommendations at the end, basically as long as they are reasonably common sense-ical vocabularies, the computers can do all of the assignments as far as optimizing a whole series of particular applications.
DR. CIMINO: I don't want to make that claim lightly. You don't get something for nothing. In our vocabulary the thought is -- we haven't done this except on little models because it is just such a huge task that I can't do it myself. I have no resources to do this.
If I take my cardiac disease terms and I represent those and fully specify those with these kinds of semantic definitions and so on, if I can do that with the ICD-9 codes, then these terms in the hierarchy, they will fall under the ICD-9 codes automatically.
So, if a clinician identifies a patient as having one of these things and somebody says, okay, now I need to know what the ICD-9 equivalent is because we are going to do financial reporting or whatever, just pop up the hierarchy until you hit one.
Now if it is a multiple hierarchy and you hit two, well, then we have got to have ways to do that. That other thing sticks out like a sore thumb. The definition of other is just by definition undefinable, because it is defined by everything else in the vocabulary.
So, if I have pneumonia not elsewhere classified, and somebody adds another pneumonia, the meaning of pneumonia not elsewhere classified has changed.
I have to have some way to represent that. The way I do it is put all my pneumonias under that other one. But it is pretty ugly.
As I say, I haven't had the resources to build these things. But in theory and at least my experience is that if you do that, build these definitions, things fall into place.
When you want to do the administrative simplification, yes, that is very easy to do. It just falls right out of the system.
DR. DETMER: I think at this point -- by the way, I ought to mention that Dr. Cimino came through sleet and ice and Chris came through snow. It is sort of like amazing, like the Post Office or something. Chris, we appreciate your particularly valent effort to get here and we look forward to your comments.
DR. CHUTE: I want to begin by apologizing for stumbling in late. It obviously doesn't reflect in any way on the importance of the committee.
It had a lot more to do with those two closed airports and six canceled flights which I had to get past which was sort of the challenge.
As some of you know, I have spoken here before. In general I tend to rail somewhat comically about things. Today I am going to do something different and actually present the collective and inclusive works of others who have contributed substantially to what we are presenting.
Specifically, this is the ANSI health information standards board working group on codes and vocabulary, and the CPRI working group in terminology.
What we are presenting today really reflects the contributions of many of the people at the table and many of the people in the room and obviously folks elsewhere. I want that to be at least understood.
One of the premises that we have made -- and I don't think there is a lot of controversy about this any more -- is that terminology or the way we think about it has become quite central to several processes in health care.
If you think about it from a knowledge representation point of view or a practical data coding point of view, really information begins with the patient, from which we can build clinical data bases, observational repositories, and yield, through outcomes research or epidemiologic analysis, new knowledge.
If we are clever, we can reengineer that new knowledge into guidelines which can build expert systems. That closes the circle to continuously improve patient care.
I think that is the dominating premise that causes us to think about terminology in very broad context.
If any of these processes are to be undertaken efficiently, if linkages are to occur practically, the common core terminology needs at least to have a consistency or comparability that I think many would agree is not yet present.
There are many ways that one can achieve this. In our work we assert that really there is a continuum of thinking that goes from the notion of a nomenclature, detailed, nitty gritty specifics of what actually happens to a patient and what should actually record in a patient record, all the way to broad, statistical classifications or aggregating classifications that really serve different purposes.
For example, patient data is clearly very detailed. It can include, and does include in clinical records, notions of anatomy, stage, severity, extent and so on, and qualifiers which are distinguished from modifiers, in that they are operational or administrative, probability, history of a disease, status post of procedure and so on and so forth.
On the other hand, there are real world practical needs to have consistent aggregated data, be it for reimbursement, be it for public health reporting, be it for community health interpretation.
Those really can focus on groups of patients or strata of patients for analysis or broad statistical or fiscal categories which are distinct from these very detailed patient data.
It is important to recognize that in an ideal world there would be a consistency along this continuum, that there is not a mutual exclusivity here, and that one winner or one loser need not come out of a match.
There is room for coordinating and integrating terminologies at these various levels of granularity and at these various levels of application.
At the end of the day, one can have an interlocking suite of terminologies that can be readily mappable, that can be applied in a patient record but fall out to yield aggregate data in a very efficient and cost effective way.
They can fulfill the requirements and services that the national committee is charged to serve. At the same time, they leave fertile ground for the evolution and standardization of electronic patient record content which is clearly, I think, where the future is in the next decade.
Toward that end, the ANSI has created a framework which we hope can characterize terminology development. I think I can say briefly that it is not at all inconsistent with the Desiderata that Jim has prepared.
I had an opportunity to hear Jim's paper in Jacksonville and read it subsequently. While Jim has focused on the descriptive characteristics of the terminology, I think the purposes of the framework document are to say, fine, since we agree on the principles, how the heck would one go about operationalizing that kind of thing.
How would one go ahead and develop an integrated, interlocking suite of terminologies.
Some of the characteristics of terminology that derive from that document include sort of the obvious apple pie and motherhood types of issues, completeness, in-depth coverage, it has to be comprehensive, all segments of health care and so on.
More important, it should have non-redundancy. It doesn't do to have many ways of representing the same concept and then you kind of pick your choice.
Then when you try to use that data, you have to anticipate, well, gosh, how did people choose to code that information. How did people choose to represent that information.
That is distinct from the notion of synonymy. There should be a lot of ways to get the information in or, for that matter, to get it out again.
It gets at this notion of the concept level which Jim had characterized, in which there has to be a single concept to which various strings or phrases or synonyms can map.
That has to be non-redundantly expressed within and among and between terminologies.
The structure of the terminology is actually quite critical to its practical utilization, particularly at the level of granularity of patient data.
Jim described this as the compositionality of okay, we invent some big long term to define fractures, was his example. Then somebody else decides they want to modify it in a new and innovative way.
A way around that is to support under the hood, perhaps, notions of atomic structure, which simply means that you can get down to the unit pieces of how you compose these parts together.
You have specific rules of compositionality. One of the problems with several of the existing terminologies is that, sure, you can have all these atomic pieces and different ways of expressing things, but you can put them together in the darnedest way.
It is hard to predict how you are going to do that and, worse, it is hard to predict how you are going to get them back again.
So, detailed rules of how to compose them are really quite crucial.
Similarly, an explicit representation of synonyms is really quite important. Otherwise, it becomes confused with the redundancy problem.
If it is not present, you cannot support practical entry in real world applications that are going to have to use that.
Terminology should, by its nature, have an attribute that would allow the notion of inheritance, the notion that tumors as a class are going to have notions of stage, are going to have notions of grade and so on and so forth, in a predictable kind of way.
In the context of tumors, if we talk about multiple hierarchies, it is obvious that one might want to think about stomach cancer as a tumor. But it is also a GI disease.
As one thinks about that concept from a different hierarchy, the term, no matter how you navigate to it, if you want to think of it as a navigation problem, should be the same underlying concept, very clearly.
Notions of uncertainty, lexical rules, these really have more to do with representation and navigation.
This one is quite key, in terms of the representation of the terminology itself. Many of the existing terminologies, I think, would be ready to admit that they are stymied by arbitrary limits of the way you represent this.
Heavens, we have got X number of digits, period, X number of digits. Gee, we ran out of digits, we have to re-invent a new category, or no more conditions need apply. There is no room at the inn.
That kind of arbitrary structural restriction is clearly incompatible with the goals of accurate and precise representation of information. I mean, it just is.
Furthermore, the heritage of that kind of representation really grew out of a Seventeenth Century data model, if you want to go back to the London Bills of Mortality.
Folks, we are in the late Twentieth-Century now. There are ways to accommodate this from an information science point of view. The notion of arbitrary restrictions is really not one that we should accept any longer.
We did make up a new word, Jim. Instead of meaningless identifiers, this week we decided it might be, let's try on context-free identifiers for size.
It is the same basic notion, that if you have an identifier it can't be linked to meaning.
As we learn more about concepts, as we learn more about diseases -- I mean, AIDS is a classic example.
When AIDS was first characterized as a syndrome, some people actually thought about it as a type of a tumor. The reality of it being a retroviral infectious disease was not clear.
As a consequence, any classification at that time that had pegged to the wrong hierarchy and had attached an identifier to it that way would have made a terrible blunder.
One would have had to have changed that, which is in violation of the next principle, that once you identify a concept, you can't change it, because historical retrieval, longitudinal retrieval is going to be deucedly difficult if what the number means or what the concept means keeps changing from year to year or month to month or day to day.
That introduces, of course, explicit notions of version control, of obsolescence marking.
Sure, terminologies are going to move forward. Terms are going to be irrelevant or obsolete or inappropriate.
That doesn't mean that you eliminate them. You simply mark them that way, explicitly so.
In the maintenance context, terminologies clearly should have appropriate definitions. This is sort of an ideal if you want to leap to the international levels or even in multi-ethnic communities in the United States; of having language independence as an intrinsic capability, and with a concept oriented terminology that is really quite possible.
Finally, the notion of responsiveness to change.
I put neat little doo-dads on the top of this slide. It doesn't project well and I notice it didn't print well either, but trust me on this.
This is the National Summit on Health Information Solutions, which was the CPRI summit. This was their logo, which I cribbed off their power point slides and put on these slides.
The reason I did that is these recommendations which I am presenting, which derive from the framework from CPRI and ANSI, really were the formal recommendations made at the national summit last November, was it?
I am simply repeating them here, because we haven't changed our mind.
Some of the recommendations that we think are important is that everybody has to sit down in the same room and talk this out.
One way to do that is to have a conference. We could have a party, too, but conferences tend to be more readily accepted.
The folks that have to be there are really the terminology developers, for sure, health care providers, system vendors, health care payers and representatives of government.
The hope is that, given the dominance of the information intensive nature of medicine today, that a forum could therefore commit to the critical infrastructural need to coordinate terminology development for all its multi-faceted purposes.
The next recommendation follows somewhat. At this party it is hoped that folks might select, or at least engage in a process to choose and entity -- this was terribly carefully phrased -- serving the public interest to coordinate health care terminology development and maintenance.
As best we could tell, that doesn't leave any candidate options out for being that entity. That was very deliberate.
There are no preconceptions of who or what such an entity might be that would fulfill this role. But it is clear that somebody has to be a coordinating role, because this is a very involved and complicated problem.
The third recommendation was to promote the evolution of existing terminology. Obviously, we don't want to throw out a lot of excellent work, nor can we ignore the practical requirement that impinges upon us every day, to adhere to existing mechanisms for reimbursement, existing mechanisms for public health reporting, and existing mechanisms for internal quality improvement and the like.
The evolution of those terminologies toward an integrated set of interlocking terminologies, which would require the endorsement of the framework requirements, and obviously establishing partnerships among the people who had the responsibility for maintaining, using or otherwise authorizing those terminologies.
Finally, a recommendation which seemed very clear, a lot of folks in health care and obviously outside of health care don't yet grasp the strategic and, indeed, tactical importance of having common, cohesive, comparable terminology resources that can be integrated into patient records, that can derive aggregate data for multiple purposes.
Hence, an education process needs to be undertaken to ensure that that understanding is present as widely as is practical.
So, terminology is the second-most urgent issue in health care information today. Even I would admit that. I usually take bets as to what the most urgent -- I usually take bets about the most urgent thing, but I will spare you that. This group is too big.
I think confidentiality and security is the most urgent issue. Quite frankly, a lot of the goals that we have articulated -- of learning from patient data and so on and so forth -- are completely moot if we cannot have fair and equitable and responsible access to that data, while protecting the rights of that patient and their confidentiality.
That remains the most critical agenda. This comes in second.
The problem of terminology probably underlies virtually all secondary uses of patient data in the nation today. I think that is obvious.
This is sort of an opinion. Uncoordinated initiatives prevail, but they are getting better. It is quite striking, over the past 18 months in particular, the number of developers and terminology groups that have begun to coalesce to converge to form partnerships and to dialogue.
Hence, I think it is feasible and, indeed, would be acceptable to propose a convergence to a common set of systems.
Furthermore, it is critical that we do so to have efficiency in the health care system.
As a closing observation, this is sometimes perceived as a battle, the nomenclature folks versus the classification folks.
There is literature going back with those kinds of titles -- the nomenclature classification controversy -- into the mid-1950s. This is not a new argument.
I think the assertion can be made very clearly that these are not mutually exclusive. On the contrary, they are highly complementary and can and must exist on a continuum that is coordinated and interlocking.
Terminology development should ensure the smooth transition and detail from patient records.
What we are saying is if you have got patient data in the record, gosh, it should be easy to make that patient data into other forms of aggregation that are needed. That is all we are trying to say there.
As an implication, intermediate levels of granularity -- because the level of aggregation that one might want to do might change depending on the application -- but an explicit representation of intermediate levels of granularity, essentially a hierarchy of detail, can and should be incorporated as interlocking systems of terminology are coordinated and evolve.
DR. DETMER: It is always terrific -- I always tend to think presentations are terrific when I agree with them. I have to say, I really found these terrific.
Okay, a question or two to Chris and then I think we will move on to our three colleagues.
We have Patricia Brooks from HCFA and Betsy Humphries from the Library, and Donna, I guess you are from the National Center for Health Statistics. I don't know which of you wants to go first. Have you sorted that out?
MS. GREENBERG: We asked them to serve as discussants. I think any of them can -- they will be preparing specifically comments or presentations for the April meeting.
We felt that since they were co-chairing the classification groups, we would get their reactions.
DR. DETMER: If you want to put ideas on the floor, and then we will open this up.
MS. HUMPHRIES: I just want to say that I think you all received a copy of the charge for our implementation team this morning.
I think the speakers received it. In view of the implementation team, in looking at the task before us in terms of implementation of Kennedy-Kassabaum, I think it is really quite in line with the notion of all speakers, that one sizes doesn't fit all.
We need a continuum, and we need to worry about the higher level aggregation of data or granularity that may be appropriate for administrative transactions and statistical reporting.
We also have to look forward to the deployment of computer based patient record systems over the next five to ten years, where a much lower level of granularity will be captured, for the reasons that Chris and Jim elucidated.
If the K2 process is to serve all these needs well, then essentially by the time the year 2000 comes along, and we implement the standards at the higher level of granularity, that we will have to have an implemented strategy such that at that time there will be an effective map between the lower level vocabularies recommended for the patient record, and the higher level of granularity which we, at this point, are assuming will be required for the transaction and classification purposes, particularly on the time frame that we have before us.
One of the things we are interested in getting out of the hearing is whether the field at large agrees with that point of view; that if we are going to do something that is administratively possible, feasible, practical, that we need to probably define the administrative level classifications for the transactions that are the main focus of the immediate agenda in the law, at a level of granularity that is not all that different from the one that is being used.
We need to also move on as quickly as possible --
DR. DETMER: Set the stage for the future.
MS. HUMPHRIES: Say these are the lower level vocabularies and get working on an effective mapping from the lower level granularity up to the higher one.
MS. PICKETT: I think Betsy pretty much has covered all the different areas that we have been trying to address as part of the team.
We have been working with a very large team, so we are trying to obtain as much input as possible. We have started looking beyond just the team to see if we can incorporate additional information.
So, I want to thank the presenters for providing us additional information that we could then incorporate into the work that we are doing, as we try to look at all the different nomenclatures, classifications and vocabularies that we are currently addressing, and as we are finding new ones.
They will keep us focused on what we are doing and to help us along in the deliberations of the team.
MS. BROOKS: I think they have covered it real well. I don't have anything to add except that we are very early in the process.
We have only had two meetings of the team and we have 17 people working on it right now. But that changes.
DR. DETMER: We are open now for discussion on this.
DR. COHN: I personally want to thank all the speakers. I personally want to thank Dr. Chute, who I know went through extraordinary efforts to get here.
I think that he did a very good job of representing, I think, the views of NCHISB and CPRI.
I actually want to reiterate that those documents actually were approved as white papers by both organizations, just not by the work groups associated with them.
I think they actually stand up very well to the sort of work we are doing.
I guess I also wanted to comment that we talk as a committee about the issues of managed care and issues of other uses of administrative data.
I think this may be an approach as we are all looking at this data, about how we are going to try to move into the issues of outcomes as practices, decision support and all of that.
That is certainly something that needs to be considered by the committee as we begin to talk about the extensibility of code sets and all this for administrative transactions.
Having said all that, I actually had a question for the discussants related specifically to the charge of the coding classification implementation team, which I think -- the last time I looked at this it was actually the medical coding and classification implementation team.
Obviously they have expanded their charge, which I think is great. I am actually trying to figure out in my own mind how this is going to cross cut with other implementation activities, and evaluations of the administration code set, within each administrative team activity. How do you see all that working together?
MS. HUMPHRIES: Do you mean, how are we going to be coordinating our activities with the other implementation teams that are identifying what data elements there are in various transactions and other standards?
DR. COHN: Right, evaluating the code sets with the administrative transactions.
MS. HUMPHRIES: One of the reasons why the original proposal for how the implementation would be handled was extended to include the infrastructure team -- they were represented here yesterday by Bob Mays.
In essence, they currently are putting all the pieces that they see of the different transaction standards that are being considered into a data dictionary.
Of course, we need to be sure that we understand what are all the elements of the various transaction standards as they go forward, that require a coding scheme.
I know that we even had some comments on our charge, when people said, gee, why do you have so much in this charge. They were only thinking of procedures and diseases.
We said, well, wait a minute here, we have got transactions where the names of the drugs come in and the equipment codes are used. So, we need to worry about those as well.
They said, oh, I guess you are right. But they were only sort of thinking of part of this.
We believe, from our perspective, that what we have here covers most of what you would call the medical or health care coding needs.
You could then argue whether the demographic codes belong here. They have been assigned to another group. Then we brought up the issue of locations yesterday.
So, as the infrastructure team works through their analysis of where we have multiple places where some sort of code will be required, some sort of table of valid values will be required, it is conceivable that this might even broaden. I think we have it all.
Some people have said, how can the one hand go ahead. How can we go ahead with the claim when we don't know the coding system.
On the other hand, how can the coding group go ahead if they don't know exactly. While they might be, in an ideal world, doing these things serially, we sort of believe that we understand that we do have to have a coding system for procedures and we do have to have a coding system for a range of conditions. We have to cover these things.
Even though we may have to make some adjustments later based on what comes out, we figure we can at least work on the big pieces that are identifiable.
DR. STARFIELD: These wonderful presentations actually addressed an issue that we were talking about just before the panel convened. That was question one on our proposed list of questions for our April hearing.
That is, can one system serve most, if not all, purposes, and then we listed the various purposes. Thee answers I heard were different. I think some of you implied that, yes, if we do it right from the beginning. Could we talk about that?
DR. DETMER: Let's discuss that.
DR. CHUTE: The HANSI HISB group went back and forth on that quite a bit. The phrases that we used were the monolithic system versus the interlocking system; same idea.
The truth of it is, either solution from a technical perspective is possible. The conclusion or the opinion of the ANSI/HISB group is that an interlocking suite is more practical and is achievable in the context of existing systems evolving toward that, rather than a dramatic revolutionary introduction of a massive monolithic system that would serve all purposes for all people.
We thought it would be kind of hard to build such a system also.
DR. STARFIELD: An interlocking suite is a system.
DR. CIMINO: Another way to look at that would be to try to figure out, when you draw your circle around what you call your terminology, what is in the circle.
Let's say for the sake of argument that ICD-10 is not going to go away tomorrow. So, whatever we do, ICD-10 is going to exist in some form.
We can draw a circle around that and call that a terminology. If we can define the concepts that are underlying ICD-10 and represent them in some other framework, what Chris calls interlocking, and draw a circle around that whole framework, is that the terminology?
If that is the case, then yes, one size fits all, if you want to think of it that way.
Different people will want different views, and if you represent this the way a lot of us have been talking about representing it, you can get that view.
So, you can actually have multiple views. Depending on if you are the blind man examining the elephant, you may feel you are getting the piece the way you want to perceive it. In fact, there may be a much bigger thing underlying that.
MS. WEIGEL: I guess I am not sure if I agree or disagree here. I don't honestly think that right now, what we have, that there is a one size fits all, if you look at the things that currently exist.
Again, they are designed for different purposes and we put them to different kinds of uses.
I don't disagree with what either Chris or Jim are saying at all. On the other hand, I think back to the level of the average user in the average either hospital or, if you will, an almost lesser trained individual many time exists in the individual physician office.
This is where the data actually starts. This is what gets put together. This is what gets sent someplace to do something with.
No matter how fanciful the nosology is we design, if the user can't use it, we haven't gained anything. I am still stuck in that place, I think.
No question that sort of behind the scenes mapping of thing A to thing B to make a giant sweep is going to be the answer of where we need to go. How we get there I think is going to be a difficult issue.
MR. MOORE: This is back to Simon's question that Betsy was trying to answer. In the original legislation it said that we would be responsible for coding and classification of those transaction sets.
When we looked at that, it would take two different kinds of people to work in that area. We tried to separate those people who were working in the transactions as a business reason for the transaction, and those who were working from a medical or clinical, to identify what care or things of that nature.
We tried to separate those, because it took different people with different knowledges. We considered this was going to be, in itself, a more difficult process to wrestle with and we broke it out.
The code sets, the elements that are within a transaction, whether it be a place of service or type of service or a revenue center or whatever, or sex or whatever, will have a different code set with it.
The people working on that transaction will be responsible for identifying those code sections.
DR. COHN: I guess I will ask another question here. Now, I adjudge that both the ANSI/HISB framework and Dr. Cimino's Desiderata are not so different.
I think both of you are shaking your heads that you are in agreement with that.
I am curious from the other participants here, both the discussants and Ms. Weigel, whether the type of criteria we are talking about might be reasonable criteria for the committee to begin to move forward with in terms of our look at codes and classifications, as well as -- as we proceed on.
Are these basically the area that we should be going and the way we should be proceeding.
MS. WEIGEL: I will start by saying that I think that we need to hold the hearings. One thing that I feel is that we have to give the most play in this initial hearings on this subject to the people who are using -- who are creating transactions now.
I think it is just an artifact, the way the list was drafted, that we had this page full of those of us who have been involved in vocabulary generation or, in the case of the MLS, mapping between them.
It seems to me that we need to reverse the order and we need to hear from people who actually do this.
I think that one of the things that we are looking at is a very short time frame for the implementation of administrative standards for transactions.
One of the things that I think is a very good strength of the law is that it is built in there that you establish the standards and they are not supposed to last for the next 25 years after you establish them. In fact, there is a way to move forward.
The fact that there may not be any standard, for example, that may be readily available -- or shall we say there may not be classifications readily available for use at a comparable granularity to what people are used to in administrative transactions, that have many of the characteristics that I desperately agree with.
For example, the context-free identifier, which I am a total bigot for, if, in fact, there is not an available system that we could announce a standard within 18 months and implement by the year 2000 -- I think what we are thinking about is that we have to keep these in mind and we have to look at an evolutionary strategy.
Whether we make the first selection with all of these characteristics, it may not be possible, given what we have in front of us to use within a very short time frame.
Obviously we have to hear from people in terms of their burden of implementing new systems and so forth.
DR. DETMER: Part of what I see that we are in, I think the situation, as you said, is improving but it is still disjointed.
There is no question, I think, that there is enormous urgency with the development of these systems globally, that the timing, in a sense, for trying to put this together really is now.
I guess from this committee's perspective, and my role as chair of it, we are in what, I think, our 48th year of this committee's history.
The K2 is an important issue right now, but obviously we are part of a trajectory. We are part of a trajectory that is hoping to represent the future beyond obviously these important times.
These are important times as well, because this is the very central agenda, it seems to me, just curiously at this time.
The question I have, I guess, is how would you advise us as a committee to really -- what role might we play besides just what I have done by trying to give some voice to this.
What role might we play to try to move this process along?
MS. WEIGEL: Let me respond to that a little bit, by going back to the question that Simon asked originally. That is, with what both Chris and Jim presented -- there is no question that is the direction I think we need to head in -- the problem is, in part, that we can get there far quicker if we can convince every one of the how many hundred thousand physicians there are to use exactly the proper terminology.
With all due apology, even though you don't like NEC, the poor coder who is sitting there and has no place to put what someone says, has to go somewhere.
Until we can get over that hurdle, that will be a huge achievement. That means, in part, that we need to head in that evolutionary process.
We all know how long evolution took, and that is a problem for us as well.
What the committee, I think, can continue to do is be the forum where everybody has the opportunity to come and feed out of the same trough and let their views be heard, but with some kind of organized effect to get to where we need to be.
Now how do we do that? I don't know. I am retired. It is tougher.
DR. CIMINO: I agree that it is tough. When I think about it, I think about retiring, too.
I think that first of all -- I think one thing the committee can do is educate everybody -- we are doing a lot of coding now and we are trying to fix that coding process.
In fact, that process is an artifact of other things that we have been doing. Really, Chris and I both had pointed out that we need to be coding the origin of the data, doing that right.
That is a whole new ball game. It is a whole new problem. People will want to avoid that problem because it is going to be very expensive to try to solve that problem. But when you solve that, a lot of the other things go away.
The NEC problem does go away because if I can code the main piece and then the NEC piece, the part that is not classified, I can code that with something else or that is a part of free text, I can put that in the computer can tell you that it is NEC.
So, we don't have to do the coding that way any more. There are a lot of technical solutions to a lot of these problems.
They all boil down to, you have got to code the data at its origin. That origin is not a coder in the medical record room. The coder is the provider, the ancillary system, is the patient themselves.
That is where the data are coming from and that is where we have to code it. If we can do that right, everything else will fall out of that. I really very strongly believe that.
What can this committee do? The ANSI\HISB recommendations are a good start. I think that there was a plan that was kicking around. I think Ed Hammnod had drawn up a plan for actually how to do this.
It had to do with hiring people away for a couple of years and forming a committee of people who were really going to put their shoulder to the wheel and really just do this full time and make that happen.
Maybe that is a kind of a plan that should be looked at, but it is not going to come for free. It is not going to come from volunteer committees.
I have been putting volunteer groups, the Canon group for instance, the electronic medical records collaboration. I have been involved with a lot of things that are more or less voluntary.
That is not the way it is going to happen. It has got to have funding. It has got to have a mandate to make it happen.
DR. CHUTE: Yes, yes, and yes. Carol has raised twice now the notion, gosh, we have to be a little careful what we wish for or we might get it and then what would the coders do.
I think we have to separate out the notion of navigation of this elegant way of representing information from the underlying terminology itself.
Part of the properties that are articulated in the framework really allow ordinary human beings to get at these kinds of meanings very easily. That is a point I want to make.
To answer your question, what can the committee do, I agree with Jim. Perhaps the first one is to endorse the notion that there would be a conference, a summit, a meeting, name your noun here, that would have the endorsement of this committee and other groups, that would seek to resolve at a technical, practical and strategic level, some of these issues.
I certainly agree, Betsy, the goals that you confront as an HHS committee, as a data council subcommittee are urgent, are practical, and have to be met. I don't think that is in conflict.
I think we are simply saying that in parallel we can begin to layer on a strategic approach to these same questions and ensure that in the process of forging a strategic vision and plan, that the tactical decisions that we make in the meantime don't trip us up.
MS. HUMPHRIES: I don't disagree with what you are saying. We think that the reason why our charge is organized the way it is, which does not only address the issue of what we need to do in the short term to meet the requirements of the law and improve things a bit for the administrative transaction standards under administrative simplifications, but in our committee discussions we really feel that we need to do that and, following straight after it, we need to say, okay, but here is the interlocking set, or the best basis for the interlocking set that we can come up with.
We actually believe that the NCVHS should help us by holding the hearings where we can -- I don't know whether we are holding the meeting at a summit or whether it is in this room, where people are coming to say what are these interlocking sets that they want.
Then we say, all right, the strategy is that we have the NCVHS assisting us in coming up with what we believe is a practical recommendation about how we get from that base set to where we want to go in terms of a maintenance situation.
As we have said, we really believe that the reason why we need to define both layers -- that is, the administrative transaction layer and what is going to be the basis for the lower level as quickly as possible -- is because we are assuming that the development of these patient record systems is proceeding apace.
We are presuming that by the year 2000 there will at least be some lucky souls in exactly the state you describe.
They will be capturing some of their data at a lower level of granularity. They will not want to have to read those records and recode them to the statistical classification level.
We need to have a map that is at least reasonably functional in place when that occurs.
As I said, I was actually regarding the committee, given their charge, to advise Congress and the Secretary on the overall requirements for patient records.
This is a big requirement. I was assuming that they were going to convene some of these people to discuss these issues.
DR. CHUTE: It occurred to me that I should amend one of the things I said about voluntary committees. Some of the successes that we have had in standards in medical informatics, HL-7 is a big one, and that one was done with a small committee that just did it.
They didn't worry about pleasing everybody at once. They got something out there that was expandable and extensible and flexible.
Another one is LOINK. I think LOINK hasn't made much penetrance yet in terms of a household word. The model and the way they did it is incredible.
I took LOINK's standard and I have adapted my own vocabulary to it with relatively little pain. There is a lot of grunt work that is involved because I have got 4,000 lab tests and I have got to find ways to model those, but it was relatively painless.
I can put a master's student as a project and get them to do it and get them to understand it. It is one of the right ways.
There has to be an over-arching framework that is going to get the job done. When you can get down to one of whatever an interlocking piece is, one of the models is to do it the way HL-7 and LOINK have.
It is sort of voluntary, but there is one person with the whip and they have some funding to support it and they move it forward with small groups, flexible, intelligent, motivated. That has gotten the job done in some relatively small domains compared to what we are talking about here.
DR. DETMER: I hear two sets of tasks, two sets in this first piece, administration and then this infrastructure.
Then I hear a substantial piece of work in the implementation of this and kind of its roll-out as well. So, probably the committee needs to kind of relate, I hope, to both of those.
DR. LUMPKIN: I would like to thank the presenters. It has been very thought provoking. I am not sure I have reached all the conclusions that your presentations will lead us to.
I have a concern, because I think that we are sort of in a nexus in information processing in relationship to the provision of medical care.
When you design an information system, it is my belief that you design it to carry out the business processes.
If you fully automate the business processes -- I don't mean billing and paying, I mean actually taking care of patients -- when you automate that, that you generate enough data in that automation process to adequately allow you to deal with the financial processes of any system.
That certainly has been our experience in Illinois when we developed our integrated information system for maternal and child health, that all the statistical things that we wanted to know about were available in ways that we had never been able to imagine.
When I was in my first part of the career, I spent it teaching medical students and physicians in training how to make diagnoses, which is in fact naming diseases.
We taught them how to do that in a structured fashion. We taught them how to chart in a structured fashion. So, most clinicians in this country learn that in one way.
The difficulty is, of course, that once you leave medical school there is no reason to maintain that structure that you were taught.
We are now getting to the point of the nexus, when we look at computerized patient records, where there is, in fact, a reason to maintain that structure that you learn, both in how you identify a particular injury or illness, as well as a whole host of other reasons, what we do in that process of diagnosis and treatment.
So, my concern is that if we, in fact, got to this point of having to have short names for things because of the cost of keeping that information, whether it be in pencil and paper or electronically.
That cost now is being dramatically reduced by increases in technology. So, a computerized patient record and using natural medical language in some sort of structured fashion which can easily be taught to clinical staff is on the horizon, and I believe very close.
My concern is that we may not take steps to give future computerized patient records the ability of the physician to write that this patient has an interstice section with obstruction and so forth, which is information that we want to now.
Yet, if we get a whole new group of physicians who instead write down an ICD-9 code, then it no longer serves its purpose, the business process of communicating information about the individual care of Mrs. Jones. It really becomes a business system.
I don't know the answer to that question. We always keep that in the back of our minds, that we are adopting things that will not preclude the very important future that we are just on the bring of.
DR. DETMER: I think your comment ties to Chris' point. Maybe you are suggesting that we do find a way to try to sponsor a meeting that would pull the right people together to, indeed, look at this and other key questions.
DR. LUMPKIN: I say that regardless of whether or not we evaluate our charge under K2. I think under our charge as our committee as a whole, that that is something that is important for us to do.
DR. DETMER: Dr. Sondik, I would be interested in your thoughts.
DR. SONDIK: I completely agree. I had the privilege of attending a meeting that Chris put together in January in Jacksonville.
I came back actually concerned that the ideas and the talent attending that meeting were not as plugged into this process, and the implications from this process, as they should be.
I couldn't be happier that you braved sleet, snow, ice and all of the rest, to be here.
John really made the point that I would make. I think Betsy really states very clearly, in light of the constraints that we have, that we need to move forward with this.
I am concerned that we don't want to make decisions here that are going to hurt patient care and hurt biomedical research.
If anything, I think we should try to make decisions that would lead on the trajectory, as you put it, toward being able to, in fact, enhance research and enhance patient care, especially patient care. I think the two obviously go hand in hand.
We have heard now a few suggestions as to what we could do. Jim said, for example, that we could free up a group of people who can work very intensively on this system, if you will, for a period of time.
Chris recommended that we have a conference. I guess my own view is that this is going to be a process. I think within two or three years, regardless of how hard we work, I am not sure we would come up with the system.
I think what we are going to have is something that is going to evolve. Whatever we want to do, we need to put in place the mechanisms that will allow it to evolve and to change.
I am concerned about the management of that, the resources of that system, the research that needs to be done and the support of that research over time.
I see it as absolutely essential to improving the health care system in this country. I would be interested in any kind of suggestions that you would have about that.
DR. DETMER: I would like to tag one second, and then I am interested in hearing from the others, because it is fascinating.
I think if you look at the GAO report on telemedicine, there are very good things going on, but I don't see a strategy. I don't see an overall strategy on this.
I think part of what we are dealing with is informatics as a field is sufficiently new, although it is going to be incredibly important, that I don't think we have quite the history of vested interests and parties who have really worried and thought these issues.
I do think that although there is an enormous amount of momentum and issues as such, we don't really have a thought-through strategy on really what our health information infrastructure, in a sense, ought to be.
We are really talking about some of that language, which is more important. So, I guess Jim, you would want to call it the Manhattan project. We could call it the Rochester project or whatever.
It seems that is just a piece of what really is what our charge might be. Without going to the whole world, what are your comments?
DR. CHUTE: I think Dr. Sondik has an excellent point. The second recommendation in the ANSI/HISB CPRI was that we pick a coordinator.
Maybe this isn't the forum or the time to discuss those candidates, but there is a very broad range of heritages and legacies, not the least of which includes the NCHS with its decades of information development and terminology work, the National Library of Medicine with its very dominant role in terminology maintenance and linkages, as well as private sector groups.
Without a mutually agreed upon coordinator, all of this infrastructure and commitment and resource can be unfocused and diffused.
That would require really some group to play a very strong and I think seeded role for coordination.
DR. STARFIELD: I would peg on what Chris is saying and ask a very fundamental and practical question. Who makes the decision as to who that group would be.
There is going to be, as Chris has suggested, either neutral or very biased posturing about which direction it goes.
DR. DETMER: In one sense it is an interface between the public and private sector. Obviously just because somebody puts up their hand, doesn't mean that everybody says, yea, verily.
By the same token, we have some imprimatur, without question, to at least try to speak to the issue, not whether we will necessarily get the nod to try to ordain and shape it.
DR. MC DONALD: The whole problem of leadership and control in America is almost impossible, because this is America.
I think there is actually -- maybe the best way to declare the coordinator is not to declare now, but to encourage the good work and perhaps increase the funding in areas that are already going on.
NLM has already -- if you look at ICD-9, there are Read codes, all the ones are sort of coordinated in the data base as it stands without causing any fuss or anybody harm, and still allowing the pluralistic systems to move on, which is important in the United States.
So, one short term, before we figure out the final answer, would be to sort of encourage further continuation and maybe expansion of that work.
DR. DETMER: I am just kind of interested. How are some of the rest of you responding to some of this discussion?
DR. COHN: I was just going to say that I was delighted with the discussion as it is moving forward. Obviously the committee has a charge. I must remind everyone our discussion, our charge is to make sure that we do no harm, along with the medical oath.
I think we have all been feeling that we want to make sure that none of the actions in any way get in the way of future evolution as we seem to envision. We need to be very careful as we move forward.
It would seem to me that regardless of what happens at a larger conference or exactly how we do this, certainly the hearings that are going to begin in April around codes and classifications need in some way to be asking questions that are very pertinent to the discussion we just had.
Part of it is probably asking developers of codes, how are you going to integrate. How are you going to fit into other code systems. I mean, beginning to have that be an issue that gets noted by them.
I think certainly coordination is maybe something that comes as a recommendation down the line, although it may be premature to try to figure out exactly how to do that today.
I guess I am beginning to see for the first time sort of a way toward the future, I am delighted to say.
DR. MOR: I am really thankful. I have thought about nosology in other areas before, but this is the first time it has all been laid out as nicely. I am very pleased.
I am also very pleased to hear a sense of concern that we have a dual track responsibility. On the one hand, we have to have something recommended that will work reasonably quickly. The existing code sets are there.
I have been obsessed for the last X number of months with, well, what if we put those code sets in and everybody begins to behave accordingly and that makes difficult any future developments that are going to be more pertinent to natural language and more granular information sets.
On the other hand, I think that the issue of, should this committee take on the role of steering this process, takes K2 from what has been sort of a time limited set of responsibilities and projects it out into the future in a way that I think requires considerable discussion amongst the committee.
It is an expansion of considered responsibility in some very serious way, particularly given the fact that we had a discussion yesterday about the committee serving in an advisory capacity to the committee in a more formal way.
DR. CIMINO: This concern about locking ourselves into something that closes some doors to future work, I think it has been reiterated now a few times.
I think the kinds of approaches that Chris and I have been talking about -- normally what happens is you say, okay, there is some level at which we are going to code things and that is the lowest common denominator. Then we are going to look at how do we aggregate them and we sort of work our way up.
We act as though that lowest level, those things are primitives or they are atomic and we don't need to say anything more about them.
If we talk about having definitions, we can actually grow downward and not just aggregate upward. That has been my experience with my own vocabulary, where we would set something and then somebody comes along and says, I need more specific terms, or I need now to do natural language processing, something we weren't doing when we originally designed the system at Columbia, but which now will accommodate natural language processing and a lot of other things that nobody envisioned that the dictionary would be used for.
Because we took a disciplined approach and said, we are not going to just enumerate things, we are going to actually describe them in a formal way, or allow people to describe them where possible, we didn't close doors. It was a way to get at this other information.
One way to keep the door open is to involve people like you saw, Dr. Sondik, in Jacksonville. Those of us who are on the fringe or the cutting edge or whatever you want to call it, we are doing things now.
We are coding patient data now at a level far beyond what we think will happen at a national level in four or five years.
If you continue to keep one eye on the way we are doing things and we can continue to help provide some input, then whoever is doing this will have some way of keeping this door open.
You know, we know that five years from now we are going to want to expand in a certain area and how are people doing it now. Let's make sure we can at least accommodate that in some way.
DR. DETMER: I am clearly hearing how deeply we really want to and engage in this. There is no question I am hearing obviously a great deal of interest.
DR. COHN: I was actually just going to comment in response to Vince's comment. My understanding is that one o four charges is within the next three-and-a-half year is to advocate for standards for the electronic health record.
I think this actually is a big piece of it. So, I think that it needs to be moved forward.
I guess my own feeling is, having dealt with this a lot in my own organization, with both Dr. Cimino, who I do not consider to be part of the fringe, and Dr. Chute, is that some of this is obviously things that are done now.
Some of this is a research agenda and we need to begin to sort out sort of what is today possible, how is it extensible, what needs to be recommended as we move forward to make sure that it is workable before we start recommending.
MS. WEIGEL: The big piece -- and I will go back to that again -- we clearly don't want to design something for the future that lives with the lowest common denominator.
At the same time, we can't forget about that. If we take your taxi driver from yesterday, who has to say something about what they did that ends up in somebody's reimbursement slot or whatever, we can't forget about the fact that that is a tail for everybody here, in terms of looking at longitudinal health care and how are we going to deal with that and how are we going to describe that.
Ideally, if we look at natural language, no matter how sloppy somebody is, it is going to end up in some sort of pot that we are going to be able to deal with.
There are a heck of a lot of folks out there in the hinterlands that have never seen a computer. So, again, we can get real fanciful with what we design, but not everybody lives and breathes rarified air.
So, my only plea is that we need to pay attention to that piece, too.
MS. HUMPHRIES: I was just going to say that I know this is a committee that, in the past and now, is very interested with data.
I think that when we next address this issue, or within a reasonable amount of time, the National Library of Medicine and AHCPR is actually going to provide some data.
We have recently run a very large scale test involving people working on the Internet around the country, more than 60 people, who basically searched existing terminologies for more than 41,000 specific concepts at their particular level of granularity, whatever they were interested in.
They then provided data to us about what they did and didn't find. We are now doing a sort of secondary subject review of the data, so that is why we are not providing the analysis now.
I think it will at least provide perhaps more data than we have had in the past, showing what people want, and how that maps onto a whole range of existing systems.
There have been some very good studies done in the past and Simon and Chris have been involved with that. This will give us another piece of that.
I think it will sort of help us see what the continuum is and where people are on it.
DR. DETMER: Dr. Cimino, I know, has to leave. I think we can continue a few more minutes. But thank you very much. I know you do have to go.
DR. MC DONALD: If I could have about 20 more seconds, I have one more question. I heard maybe a concern that these systems would impede going forward.
I would like Clem maybe to comment on -- I didn't hear anything you described that would make it any harder to use ICD-10 or CPT right now today, as long as one can build these things into these other connected trees.
DR. CIMINO: If you took ICD-9 and you got rid of the coding system, the hierarchical coding system that it has, and you added some structured definitions and you got rid of the NEC terms, it would be a lot more usable than it is today.
DR. MC DONALD: You actually have ICD-9 included within your system.
DR. CIMINO: That is right. ICD-9 is included in our medical entities dictionary. As I said, I don't have the resources to expand the definitions fully, but in little areas where I have worked with that, I can then add the codes at whatever level I want to add them and use a knowledge based approach.
The system will then say, okay, this term belongs under the ICD-9 code. If it belongs under two, it belongs under two.
If I have to resolve a rule about how to automatically code that, then I have to develop something. If the ICD-9 concepts had some clear definitions that included things that said, do not code, this concept excludes the cases where tuberculosis is present, then I would know that would be formally defined in there.
I would not classify that thing, because tuberculosis would be present, let's say, and it wouldn't end up under that concept.
I could do a lot more with it if it had those formal definitions.
MS. HUMPHRIES: I think one of the issues is that in many cases the definitions exist, but they have not been traditionally distributed in an easily computer-analyzable and usable form.
DR. CHUTE: Then they don't exist.
DR. DETMER: Thanks very much. I am sorry you have to go.
DR. AMARO: My comments are kind of broad in nature. This isn't my specialization. Through my own research and my own clinical programs, I am very familiar on a much smaller scale what havoc the introduction of new data systems can bring, not only to the system but also the burden on the staff.
My thinking is sort of along those terms. I just want to highlight some things that I want to make sure that we attend to.
I am wondering how new requirements are going go impact on providers that work in the context of fewer resources and less developed infrastructures.
I think we really need to have that at the forefront. We need to look at the burden that is maybe unequally distributed with respect to new requirements.
I think we need to hear from the sector in our hearings. Maybe you know where to go for that.
I am also wondering whether if the new requirements are more complex or require different things from providers, how that might over-burden providers and also impact on the provider/client relationship, the kind of things it would focus on.
I am not sure that it would, but I would think i would want to look at that.
Sort of I think we need sort of an environmental impact, sort of on systems and providers as well. I want to make sure that we hear from the people who sort of are at the low end of the technical resources spectrum, and how it is going to impact on them.
I think they are going to have a harder time adjusting to changes.
I would like to also have some attention paid to, how can we help develop the resources and the infrastructure among that group, to make sure that they are going to be able to benefit also from all these changes and the improved system.
I don't know if that makes any sense to you, those suggestions, but those are the things that I am concerned about.
DR. DETMER: I think they are all germane. Do you want to respond to that?
DR. CHUTE: I think those are very important considerations, and please don't misunderstand. Many of us who work in this field are exquisitely sensitive to the human interface issues, to the burden issues, to the technology impact issues.
Impact statements, as you know, ideally look at both sides, if you do and if you don't. I think one recognition of late is that with the de facto installation of electronic medical records systems, not in huge urban hospital systems, but in small clinics, the computer costs the week's salary of a technician these days.
The data is demonstrating that the rate of installation in small areas vastly exceeds that in large urban centers. It is a simpler piece.
DR. AMARO: What about like the data entry piece.
DR. CHUTE: This is what I am getting at. The other side of the impact statement is that if we do nothing, if we basically ask providers to learn the very intricate and elaborate -- I have actually enormous admiration for the discipline and rigor of the ICD-9 definitions, don't misunderstand me, I don't mean to disparage them in any way -- but if we require poor providers -- I mean poor in the unfortunate sense -- to have to master those rules in the day-to-day delivery of care and preclude offering them something that is more clinically sensible that they can relate to that is sort of closer to what they learned in medical school or nursing school or whatever provider environment they were trained in, then I think we are doing a disservice and imposing a greater burden.
If we are somewhat thoughtful about the way we think, hey, if we have a consistent way with friendly synonyms and practical navigation engines that don't require expensive machinery but you can get at Best Buy for $400, then we have done a service.
We have actually brought the process closer to their view and understanding. I think an argument could be made that the change would reduce the impact of this transition.
DR. AMARO: Could I ask you, do you have any data on sort of how many, and the profile of providers that have gone to computerized records?
Do we know who it is that haven't gone to that and where they are and the kinds of clients they serve?
DR. CHUTE: Most of that data comes out of industry surveys, as you might gather, and sales figures. The sales figures are really dominant in --
MR. BLAIR: There is a survey conducted at the industry's largest conference, Health Information and Management Systems Society in February.
They will typically ask the attendees of the conferences -- about 14,000 attendees in the last conference -- what their plans and projections are for information systems.
It is not going to give you a definitive number. It will give you typically a percentage of the people that attended that chose to answer that question. It is the best data that I have seen.
DR. AMARO: I think it would behoove us to get a good sense of that.
DR. STARFIELD: Can I just add to that? There have been a few surveys. The vast majority of providers use computers for billing. It is something like three or four percent.
DR. AMARO: For billing.
DR. COHN: I was going to comment because in our organization this whole issue is a very interesting issue. We, like probably most private providers, and even though we are in a managed care environment, are in a situation where, when we see a patient, we write down a diagnosis.
When I see a patient coming in with left forearm fracture with cellulitis, I write that down.
Now today what happens is, somebody has to take that and either I code it into something that usually isn't very close -- and that is an additional activity for myself -- or the coder has to figure out how to put it into something that isn't very close either.
Now the way a more granular code could work could either be, yes, I have the ability to use computers to get that more granular topic.
Like many people did before, they hand it off to a coder who is now using a different coding system to get to that granular stage.
DR. AMARO: But we are not just talking about billing records. I am going to go home, when I go back to Boston, maybe you know, but all the community health centers that I work with, I don't think they have computerized patient care records.
DR. CHUTE: No.
DR. AMARO: We are talking about things that impact not only billing, but clinical records, right?
DR. DETMER: But we are on a trajectory. The fact is that there are places where they are. But clearly, this is on the cutting edge.
DR. AMARO: Right, and I understand we want to get there. I am just saying that we need an assessment of where we are with respect to the clinical records.
DR. MC DONALD: I appreciate your anxiety and concerns. I think we could reduce it if we made a clear distinction between enabling and requiring.
They get all blurred together and I think we get in lots of trouble if we don't keep them clear.
There are forces amongst us -- in our own minds, sometimes, we would like to have everything coded in the record, because wouldn't that be nice for X, Y and Z.
That requires time. So, we can't do it by getting resources for coders, because that is too expensive. So, we will let the provider do it and we kind of hope it won't take any extra time and it does.
I think we ought to be thinking about this extra granularity as an enabling technology that could help those places that want to use it.
Then be very careful that it doesn't become one of these coding requirements that makes you have to use it. We are just shipping the five digit ICD-9 and our doctors are going crazy.
When you have got more, longer lists it takes longer, in one way or the other. If we could really keep those real separate, I think we will be happier and we will avoid political confusion that we shouldn't have to get into to get this enabled.
DR. DETMER: Last comment.
DR. MOR: It is a question. I guess it is for Dr. Chute.
Part of this issue that is being raised is the problem of the cost of data capture. Then there is also the cost of converting that data capture to some kind of code.
Now, if I understood anything about what you were saying, it was that the data capture issue is really a matter of proper vocabulary, not numbers.
Algorithms can then be subsequently applied of any different sort, to be applied to those vocabularies, to provide sets of numbers; is that right?
DR. CHUTE: Yes.
DR. MOR: So, really what we are talking about is data capture.
DR. CHUTE: At the point of care, in an ideal sense. Recognizing that that is a strategical, and we have to acknowledge the very urgent and real world classifications and how to bridge them.
DR. MC DONALD: But could you classify, are you talking about constrained vocabulary or unconstrained. I am not sure you are thinking unconstrained vocabulary. I would just like to have it clear.
DR. CHUTE: We could have a whole hearing on the notion of user interface design and user navigation tools and mapping.
At some level, you know there is a continuum better than most, Clem, between the notion of a rigidly defined vocabulary and controlled terminology and one that has a fairly forgiving -- heavens, even Microsoft Word help screens have a natural language interface to them under Office 97 these days.
DR. MC DONALD: I still want to clarify. I know Steve has been thinking, what we should just do is take the dictation and understand it. I don't think that is quite the same thing that we are necessarily saying.
DR. CHUTE: No.
DR. MOR: I didn't mean dictation. I meant some semi-constrained vocabulary.
DR. DETMER: As you know, after lunch we intend to get back into a discussion as we move toward that hearing. I think in that regard we will pick it up.
This has been, I think, really an extraordinarily valuable discussion, where the rubber meets the sky as well as where it meets the road, and I thank you very much. We will adjourn for lunch.
(Whereupon, at 12:50 p.m,. the meeting was recessed for lunch, to reconvene that same day.)
DR. DETMER: To call us back to order, we have a couple of things at least to do. One is to re-engage our conversation on the list of questions that we were starting working our way through, and got quite a long way, I think through this morning.
It was an excellent panel and I think it really did help us. So, I would just like to come back to shaping this list of questions.
Then I think we still need to have a few housekeeping issues relating to dates and such and what the executive committee is hoping to do before the next meeting.
Unless there are other things, we will hear from the audience and then we will adjourn.
MS. GREENBERG: Just to expand a little bit on the first task, in addition to questions, which I think this is critical, because we really need to contact these people and let them know about this, we need to contact them by next week, any individuals, organizations that you feel should be heard from at the April 15-16 hearing that are not on this list.
This list compiles what I discussed with people, but it certainly wasn't meant to be a complete list. I would like us to have a little bit of time to talk about that, too, because then we can go back and invite everybody.
DR. DETMER: So, we are back open, John. Do you mind in mid-bite there, re-engaging us?
DR. LUMPKIN: I think to -- one of the things we were working through was the list of questions. The ones we had kind of singled out were question number four, as to which codes and classifications do you recommend.
Question five was identified as for ICD-10, if you are going to use ICD-10, do you want to use ICD-10-CM, 9-CM or 10-PCS.
Seven was an issue related to how do you think that you would deal with the current coding environment and improve simplification, reduce administrative burden but also obtain meaningful information.
Number 12, what are the arguments for having such systems in the public domain versus the proprietary sector. With the modifications that we had on that, it needs to be rewritten. I think it is something about the cost/benefits or something.
MS. GREENBERG: I think we agreed that maybe that wording wasn't exactly what we wanted, but I am not sure that I am clear what wording would be better.
DR. DETMER: Let's discuss that.
DR. LUMPKIN: Let's come back to it. Then question 14, what do you think about the process the way it is going so far.
Then question 15 --
MS. GREENBERG: There was a question on maintenance that somehow fell off the page. That one seemed of importance, too.
DR. LUMPKIN: Maybe we can wrap that into 12, which we haven't defined. Question 16, then, since maintenance fell off, something to the effect that, should we allow variance in how the code set is used.
Then a specific question that came up for the people who developed the code sets, I think the two specific ones are, what kind of maintenance do you have and what is your commitment to maintaining the code set.
Two, how do you see your code set integrating with other code sets.
MS. GREENBERG: Can I ask you why you didn't want question three, which seems sort of basic.
DR. IEZZONI: I was going to say, three is sort of a demographic question.
DR. LUMPKIN: I thought it was okay, but since we are trying to limit the number of questions, which we haven't done a good job of, if you know what code sets people want you to adopt, it almost doesn't matter what they are currently using.
DR. IEZZONI: I think it gives you context and it is a very quick answer. I am using ICD-9, period.
DR. LUMPKIN: Okay, that is cool.
MS. GREENBERG: We want to hear what they consider are the strengths and weaknesses of what they are currently using.
DR. LUMPKIN: Sometimes I am task oriented. I was told to try to reduce this list down. I tried. You guys still liked it. Okay, I am easy.
DR. STARFIELD: I want to make another pitch for including number one. We talked about it this morning and I think it is really important to get a feel for whether people think that they are just stuck on billing, or whether they can use these systems for other purposes.
DR. MC DONALD: It is a good ice breaker.
DR. COHN: I actually have no problem with number one, if we want to add that. I am a little concerned about the purpose, since I think maybe clinical care ought to be one of the purposes.
Maybe we just want to say -- I don't know if we want to list them because we are likely to miss some no matter what we do.
DR. IEZZONI: We are calling this the hearing on coding and classifications for administrative and financial transactions.
I guess to what extent do we want to focus people on that?
MS. GREENBERG: Is there agreement that this should be a hearing primarily focused on coding classification for administrative and financial transactions.
DR. IEZZONI: Given that we have to make decisions by June, yes.
DR. MC DONALD: The problem is that this one will be the one we use in that part, too, don't you think?
MS. GREENBERG: We will have to have a second hearing, I am sure, to get more into some of this.
DR. DETMER: We were going to have a set of questions go out. I think what we were saying is we could take input.
The question is, I guess, how do you really want to use the meeting itself.
DR. COHN: I was going to say, I would be happy with that question just as long as we don't try to identify all the uses.
If we just say, can one system serve most if not all --
DR. STARFIELD: They won't know what you are talking about if they are only focused on billing. That is the problem.
MS. HUMPHRIES: I think this question as worded, from my perspective, is a little bit ambiguous. It became clear to me in the discussion this morning that when Chris and Jim were saying we could collect more granular data and we could map it to other codes and classifications for other purposes, that Barbara was viewing that as one system.
It certainly would be, if that is the way it worked. But I think that if you send this question out cold, people will, I think, assume that you are asking, is there a single level of granularity that will suit all these purposes.
They will come in and say no, and go on to the next thing.
DR. STARFIELD: What about if we say, is it possible to serve.
MS. GREENBERG: The only thing is if there is agreement that the answer is no, maybe there is not much point in asking it, maybe.
I still feel -- I will be real candid. One of the things I had hoped to get out of this discussion, and I thought it was very stimulating and well worthwhile, was to get a little more teasing out of the differences between vocabulary terminology and classification systems, et cetera, and the different systems that one needs for these different purposes.
So, that is one issue. Then another one is John's, which is how can these be integrated. If we acknowledge, as Betsy said, differences in granularity depending upon what we are doing, then obviously -- but we still need to integrate things. That is really the issue.
I think that we didn't really get as good a discussion on that, because I think after a while we started talking as if they were all the same thing.
DR. MC DONALD: I think they maybe are, in the sense that one person's classification is another person's terminology and vice versa.
It depends on whether there are pieces below, whether the thing above is a class. I don't think that is a permanent distinction.
DR. DETMER: I think some of it is, are you a clumper or a splitter and how do you go at it.
MS. COLTIN: I think there is a perhaps more obvious problem in the way it is worded. I can't tell, from this wording, whether you are asking me, can one develop such a system, or if you are asking, can one of the existing systems do this. Those are two different questions.
MR. MOORE: I just have one more piece to throw into it. Sure, we have talked about coding for services and skills, but we also -- this group has to face another type of coding, coding for durable goods, which happens to be in part of the procedure codes.
That is a large part of what is on the reimbursement side and also I guess could affect the clinical side, if you are talking about what devices, what kinds of things do you use in order to treat patients, et cetera. I want to make sure we don't leave that part out.
We met with the people from the --
MS. BROOKS: The National Product Code issue.
DR. MOR: It seems to me that the UPC code, or whatever it is, meets all the requirements that had been discussed earlier, that they are very, very basic codes.
You just apply for them. They have no meaning in and of themselves, and they can be aggregated up into any particular package you want.
A payer can choose to pay for or not pay for something that is identifiable and uniquely there, which is a totally different thing than documenting the fact that it was there.
DR. MC DONALD: Is that true, that you can do that? I didn't know that was true, that you could aggregate them up to any level.
MS. HUMPHRIES: You certainly could aggregate them. I don't know that they have been mapped to higher level aggregations that would be useful for any particular purpose.
DR. MC DONALD: To understand it, the manufacturer is giving it a number and they can make up anything they want for it. I don't think anyone is keeping track of which ones are the same as which, unless you know otherwise.
MS. BROOKS: It is correct that they are all different and unique and you would have to map them yourself. Some people consider that a strength.
The Department of Defense, we understand, is probably going to mandate these codes, beginning with all their products they are going to buy this summer, or else they are not going to buy those products.
Many products do not have those codes on yet. So, the issue is, do you mandate something that isn't labeled, or do you wait until the get labeled and then mandate it?
That is an important issue for our group. There is a lot of benefit to it. It is drastically different from alpha numeric codes.
DR. DETMER: It sounds like this could be a useful part of this question.
DR. LUMPKIN: That last question certainly gives me the concept of taking a jet fighter and doing a bar code scan on it. I am glad they are doing something about those.
We are having a lot of difficulty in trying to define number one. One of the things that I think we need to -- I believe that if we are talking about a coding system, we are talking about a currently existing coding system.
DR. STARFIELD: Why don't I withdraw the suggestion that we include it. I think there are just too many problems with the wording. We could tinker with the wording, but I don't think we have time to do that.
DR. LUMPKIN: I think the questions that don't make this cut will be there. We do want to let them know that there are other pieces of information that we would like to get in writing for our deliberations that we may not want to deal with particularly at the hearing.
We can certainly ask that question. When it is not written down, it is easier to put it within the context of the presentation and get the answer that we want.
DR. DETMER: Are there other comments on what is still on the list, and then it sounded like there were some possible issues that might also need to get on the list.
DR. STARFIELD: I just have a question about the title. Are we including attachments?
MS. GREENBERG: I don't think so. The first phase of the implementation team does not include attachments.
MS. HUMPHRIES: As we were thinking, as you look at the list of interested parties in this, at any level of this, it just wasn't going to be possible to deal with all aspects of this.
Once you get to the claims attachment, you are potentially at the entire medical record, which is the other issue of more granular vocabulary.
We were thinking that we probably did want to have additional hearings on that issue. One of the things, of course, we are assuming is going to happen, if the committee decides that it will have follow on hearings, even though it hasn't set the date for them, that this will be part of the announcement of the initial hearing.
People will then not assume that this will be the only time they ever get to say what they have to say about every other issue.
Maybe they will be happier to focus on whatever questions you put forward to them in the first hearing, because they know there is another opportunity coming to deal with the other questions.
MS. BROOKS: I thought attachments were another work group, though, weren't they?
MR. MOORE: Attachments is another transaction set. It is more like an envelope that you put the documentation in and move it.
I don't think X12 has been able to figure out what all that is that goes in that envelope.
DR. MC DONALD: The attachments runs from whatever -- I think the decision process has kind of not even been poked at yet.
You could send it as a fax, sort of really literally like a xerox copy. We could send this structure. Then you get into all kinds of other things right away.
MR. MOORE: That raises another issue. One of the other team, the security team, and the question of, are faxes of medical records electronic transmission and, therefore, fall into the security purview? We thought so.
DR. MOR: I would like to suggest the kind of question that I would be interested in. Clearly the administrative standard creating group, I would like them to respond to the question we sort of dealt with a little bit this morning.
That is, to what extent would their code sets be flexible for developments for other applications in the future, ala the conversation we had this morning.
I don't have an exact frame for that, but how well can CPT codes be cross walked into or amplified with other sets of vocabularies, for example.
MS. GREENBERG: There is a question about cross walking and linking. One thing you could do is put a core of basic questions in the Federal Register notice and in the body of the letter.
Then attach, these are other areas that the committee has been thinking about. If you want to address these in your written remarks, or there may be an opportunity to address them in questions, fine.
They can throw it out or they can really worry over one of them and say, this is something I really want to address.
They could write a paper on it, but try to limit certainly their oral prepared remarks to maybe these, the core of questions. That would be one approach.
DR. COHN: I was actually just going to comment for the committee's sake, that I think as Betsy had commented, I think this is a pretty big area.
The fact that we all as a committee have only some little feelings about this I think is probably reflective of the society in general.
I think we just need to be aware that this is going to be the first of probably a number of different sessions on this area.
I think the question is really not trying to argue over how much we can handle, but really just trying to focus in on very clearly the first steps.
Then sort of figure that there are a lot of questions we are going to ask as we go along. We are not going to have all the answers after a day and a half.
DR. DETMER: That would probably be a good comment to make at the start of this piece for the Federal Register. So, again, people won't think this is sort of the only word, why are they only looking at that.
MS. GREENBERG: This may be -- if we stay on the time frame that is specified in the law, we are not going to have a lot of additional hearings on this topic for the K2 objective.
DR. DETMER: Yes, but this doesn't just say that.
MS. GREENBERG: Certainly, it is the beginning of looking at coding issues, both for the K2 objectives and more broadly, and more will follow.
MS. HUMPHRIES: Following up on Marjorie's point, then, I want to go back to something that Bob said about the durable goods. We might say the same thing about the drugs.
We don't want to -- when we say to people, what are you using now, if they answer, ICD, that is not really true. They are using some other things.
I mean, they are using ICD and they are using some other things. It seems to me that we don't want to not get in the folks who may be dealing in the device area and the drug area which may, in some cases, differ.
MS. GREENBERG: Obviously your charge includes all this. Frankly, if you do not object, I really this think charge has to go probably in the Federal Register and in the letter.
We are asking a question, do you have any concerns about the way it is being undertaken. This is the only way they know.
I think that will also explain -- because it says equipment in here, rather than -- it will speak for itself, I think.
DR. MC DONALD: Now that you brought up drugs, that is scratching the surface. There are at least three companies that sell sort of the classification and extra information connected to NDC.
At least one of them, and maybe two, have put it in the public domain, or at least made it free, these codes. None of them are on here. Again, I don't know whether we want to get into this.
McKinzick is the company that at least they say they put it into the public domain.
DR. MOR: There is the American Hospital Formulary System.
DR. MC DONALD: That is a classification.
DR. MOR: That is a classification. Then there is also Medispan, which I don't think is in the public domain.
DR. MC DONALD: They have been wrestling with that. So, that is the second one. McKinzick, which I think may be medical economics, I got a letter that it was in the public domain, but I haven't seen it verified.
DR. DETMER: I don't know when we want to engage this, but I think a discussion on that matter, we were on that issue right before we broke.
I think again, when in the order of all this it ought to come in, I don't know, but I think it is a very germane public policy question.
DR. MOR: I think it speaks to whom we invite, if we bring in somebody who plays with NDC codes or UPC codes. I think that is a germane issue, who do we invite.
DR. DETMER: Do you have a suggestion or comment?
DR. LUMPKIN: : We could do one session on non-clinical codes, which would be those codes that are not directly related to -- drugs I put in there because --
MS. HUMPHRIES: I don't know that we would know. Why don't you do one on drugs and devices. Then we will understand it.
DR. LUMPKIN: I appreciate that was a divisive answer I gave.
DR. AMARO: I want to come back to the issue I had raised before. I would like to make a recommendation and I would like some help in thinking about the best way to implement it.
I want to make sure that the technological have nots are heard from. I would like some help in thinking how we are going to do that.
I am thinking about folks like from the National Medical Association, The Indian Health Services, clinics in farm worker camps, community health centers, rural providers.
What are we going to do to make sure that we hear from them to hear about that end of the spectrum. I feel like we are doing a lot to hear from the other end of the spectrum, the people who really do this all the time, are very sophisticated.
I think if we want to know what the impact is going to be on the other end, we have better be very purpose in inviting them, and also having this facilitating somehow their coming to these meetings.
These are exactly the folks who may not always have access for getting there. I am sure there are lots of ideas on how to do that, and I would like some help with that.
MS. COLTIN: I would agree with that. I am wondering whether -- these are folks I would like to hear from not just about coding. I would like to hear about the standards and everything else.
This particular meeting in April is confined to the coding and classification. I am wondering if there ought to be another hearing where we hear from groups like that, not just about coding and about some of the other standards that we are debating.
Then they could come one time, since they are have nots and it may be difficult for them to come.
DR. IEZZONI: It could be part of the California set of hearings that we are thinking about; perhaps, Hortensia, the portion that deals with the standards, since a lot of these folks are serving the specified populations that committee is addressing.
MR. COLOBET: One of the things that I have been wondering about the hearings, when we are talking about east coast, west coast, is whether we can do some video conferencing.
At least for some of them, the VAs are now getting more and more video teleconferencing equipment. For some, we may be able to make that available, for them to travel less, and if we can get a set-up, to actually people testify by video teleconferencing. It may facilitate some access for some of those people.
DR. AMARO: In sort of the spirit of the approach that we took in our subcommittee of not ghetto-izing all the issues pertaining to these populations, I would like these issues integrated throughout.
Otherwise, we are going to have one gigantic --
MS. GREENBERG: I think we could, certainly. I think that is a great idea, Rob. Unfortunately, we haven't even figured this out for our implementation team, video conferencing.
We know how to do it but it still isn't always feasible. Even now, we still have our trusty telephone, as you know.
I think, though, that we could certainly try to get a few of the organizations. We have to remember this meeting is a month from tomorrow, although it is not on a Saturday.
We have got to get this out, we have got to get letters out to people, we have got to get people here. So, we are really talking about a quick turn-around.
If you could give us a few suggestions, we could deal with their travel, et cetera, and we could deal with that on a case-by-case basis.
DR. IEZZONI: Marjorie, I still feel that this hearing is about classification. It is not -- I don't know. I think that the issues that Hortensia is raising are broader than just classification.
MS. GREENBERG: They are broader.
DR. IEZZONI: That is why I was suggesting the broader forum, even though it might be perceived as ghetto-izing. I think that the California hearings were going to have a broader reach, that they weren't just going to be on classification. That is why I suggested it.
MS. GREENBERG: Many of the groups that are going to come in for this classification have already been in for other topics. It is better sooner than later to integrate them.
DR. LUMPKIN: I don't think these issues are mutually exclusive. One of the things, as I was trying to think of the various groups that we want to hear from, we want to hear from the code developers. We want to hear --
MS. HUMPHRIES: I have a question about that. I do want to hear from them. In a two-day hearing where you are really trying to say, what are the needs of the people who have to use this, what do they see as their issue, is it more important for you to hear first from the people who are using the codes?
We have received a lot of information about existing coding systems from ANSI/HISB. We have names and contact people of the people who developed these things.
I really feel their points of view, in many cases, if the organizations that are developing these represent user groups -- which they may -- the AMA certainly represents a user group, the College of American Pathologists represents a user group.
I am concerned that in two days there are a lot of developers of classification systems. Is there going to be any time left for the users?
DR. LUMPKIN: I didn't say that we would necessarily have a lot of time, but there are pertinent issues that we need to address that weren't addressed in any other venue.
So, what is their commitment to maintenance, how do they integrate with other systems. I think we have developed a set of questions.
We may want to do as we did with the standard development organizations at our last hearing, which was to have them in last, and then listen to the users in various subgroupings and then help us develop the initial questions.
Certainly we would want to have the payers and groups of providers, which would be the primary users. Maybe there is another subclassification we would want to have.
My suggestion would be that as we look at that, particularly as we bring in providers, that we represent a more diverse group than we have in the past.
DR. DETMER: Yes, I think that is certainly good. We may also want to d it at the other end, too.
DR. AMARO: That is fine with me. I think that it is good to have sort of a focused place as well as integrated throughout.
I just want to make sure that we make an effort to invite. You know who the groups are, to just make a special effort to seek their participation, so we get their input.
DR. MC DONALD: Just to clarify, I don't know that we can assume that the populations you define are technologically deficient or underserved.
The Indian Health Service that is all computerized. So, there are some of them that are ahead of the rest. There is a distribution.
DR. DETMER: That could be useful, too.
DR. MOR: I would like to strongly suggest that amongst the folks that we hear from in terms of other provider groups, would be the long-term care industry.
I don't know exactly which groups, but there are a number of national representatives. I think we need to have home care as well as the institutional care settings.
MS. GREENBERG: If you can identify specific contacts, that would be great.
DR. DETMER: Okay, anything more on this?
DR. IEZZONI: I was just thinking, we don't have to have these hearings in California. This would be okay with me.
I had thought that we had talked about having these hearings in California that would deal with the privacy and confidentiality issues around these population subgroups, as well as coding, classification, UHID, some of these issues, like the long-term care, that we would talk about at that meeting.
Again, I am sensitive to what Hortensia said about ghetto-izing. It is just that I feel that these two days are becoming absolutely packed.
What I would like to see, though, Vince, going back to your comment this morning, is about the integrated delivery systems.
Have the people there who represent the tertiary all the way down to the home health agencies.
DR. MOR: : The problem is that the people they will send, the CEOs are not even going to think about their home care providers or their nursing home.
DR. IEZZONI: We need to get the Kathy Coltin equivalents. Kathy I think might be able to help us with that, because you would know the folks.
DR. DETMER: Okay, thank you, John. It looks like you have reached a wonderful summary of this.
I think this is what we needed to do for this. I think the other issue we do need to talk a little more about some dates, and I don't know that we will be closing in on this.
DR. MC DONALD: Vince has really stirred up a whole hornet's nest in my mind. I mean good, it is a good thing.
One is that the DOD is going to use UPC everywhere. How does that interact with NDC codes, which are also products. We have some other tangles in this, which is also used everywhere.
If the UPC code -- it is a wonderful code. It is very granular. If we can get it classified -- we ought to be sure we get that discussed.
Anything that is a product could have a UPC code on it. We already have some preferred coding systems in the world. NDC, I think, is for the prescriptions and outpatient area. We need to grapple with some of these divisional issues, which one is going to be used.
DR. DETMER: I think we will still be working this as we move toward this date.
DR. MOR: Let me just differentiate. The coding structure at the most granular level is quite different from any kind of classification system that you happen to impose on it. The NDCs are classified in at least three different ways.
DR. MC DONALD: That is not my problem. The NDC is also a very granular coding system, which exactly overlaps what UPC would do for those drug products. That is all I am pointing out. There are some overlaps.
DR. GREENBERG: We have the dates for this next meeting set up. I guess in time the next meeting will be the west coast meeting.
DR. IEZZONI: There was also a sense that the subcommittee on specified populations -- I have got to get that rolling off my tongue -- should have a one-day kind of working meeting.
The question is, should it be before or after the west coast meeting. It would be better to have it before, but I don't see how this calendar is going to work.
DR. DETMER: I think it would be better to have it before. I must say, because of the short time anyway, we will probably have to be working these issues somewhat simultaneously, I am afraid.
DR. IEZZONI: Yes, I agree.
DR. DETMER: We will just have to work through that. It seems to me that trying to have --
DR. AMARO: I think we should have our subcommittee meeting before that. When were you thinking of having that meeting?
DR. IEZZONI: We were thinking of the subcommittee meeting -- Hortensia, if you remember, Olivia said that the research undergirding the OMB wasn't going to come out until May 8th.
So, we had talked about having our working group meeting after May 8th.
MS. GREENBERG: So, this would be a meeting not where you would have a presentation.
DR. IEZZONI: We would have an afternoon where we would have two-and-a-half hours of presentations. In the morning it would be mostly our group working.
From 1:00 to 3:00 we were going to have presentations on the background for this OMB decision, the research for the racial and ethnicity classification decision.
So, when were you thinking about these west coast hearings occurring, Don? Had you a feeling about that?
DR. DETMER: End of May, beginning of June. I think I would like to see them before June.
DR. IEZZONI: Before June 1 or before June 24.
DR. DETMER: Before the end of May, in other words, before June.
MS. GREENBERG: We tentatively said first week of June.
DR. DETMER: I would like to have as much of this into the pipeline, if you will, before we start, and still have a little time to write our recommendations and so forth before we come up to our June meeting.
The executive committee hopes to meet the first week of June. I think we would also like to have the benefit of looking at some of that.
DR. MOR: Are we talking about one day's hearing or two?
DR. DETMER: My guess is it will probably be a couple of days.
MS. GREENBERG: I think you really need two.
DR. MOR: The last week in May is Memorial Day, that Monday. Two of those four days would be available, if that works for other people.
DR. DETMER: Let's check our schedule for the last week in May.
MS. GREENBERG: That is the California meeting, and you are thinking that the other meeting would be earlier.
DR. IEZZONI: So, the working group could meet maybe the week of the 12th of May. That would allow us a little slippage in case the research doesn't come out on the 8th.
MS. GREENBERG: A one-day meeting?
DR. IEZZONI: It was going to start at 9:30 in the morning and end at 3:30. So, it was going to be in and out for those of us on the east coast.
DR. AMARO: How about Friday the 16th?
DR. IEZZONI: It is possible.
MS. GREENBERG: I guess at this point the population based work group --
DR. IEZZONI: I think we need to talk about that. The population based work group is made up of two people. I am one of them. George Van Amburg is the other. That is a virtual work group.
DR. DETMER: For the moment it is virtual.
MS. GREENBERG: I know they met with you yesterday.
DR. IEZZONI: I met with me yesterday.
DR. DETMER: There was absolute consensus.
MS. GREENBERG: Okay, fine. I won't worry about them.
(Meeting dates discussed.)
DR. IEZZONI: Why don't we check with staff. Why don't we come up with one other day that week.
(Meeting dates discussed.)
DR. DETMER: What we will look at is either then or the week of the 19th.
DR. IEZZONI: I can't do it then, but I don't need to be there.
DR. DETMER: Yes, you do. All right, we have consensus on that. We will try to do it, then, the first week in June.
DR. IEZZONI: So, you will get back to us which dates that week.
DR. DETMER: Is that holding the 3rd and 4th or the 2nd and 3rd?
MS. GREENBERG: I heard an objection to the 2nd and 3rd.
DR. IEZZONI: The 3rd and 4th would be fine.
(Travel arrangements discussed.)
DR. DETMER: So, why don't you look at the 3rd and 4th.
MS. GREENBERG: Yes, we will be in touch. We will try to work the executive committee in there.
DR. DETMER: Do we have a fall date set?
MS. GREENBERG: You know that the meeting of the full committee is June 24 and 25. There is a September meeting. I know there is some confusion even about the June dates, so I want to make sure everybody knows what the dates are.
DR. DETMER: The dates are?
MS. GREENBERG: The dates are June 24 and 25. The September meeting is September 8 and 9. The November meeting is November 5 and 6.
DR. DETMER: Did everybody hear those dates?
MS. GREENBERG: You have three full committee meetings scheduled for the first of the year: June 24-25; September 8-9; and November 5-6.
DR. DETMER: I think we have proven that we have worked ourselves almost to exhaustion the last couple of days.
I want to thank everyone. Obviously in a sense we started with an essentially largely new committee. We have really picked up an awfully lot of responsibility and, as well, I think a lot of reshaping of ourselves.
This has required a lot of patience, I think, as well as work. I think we are getting there.
It is also great to have our at least now full committee appointed and here, and I am delighted on that. We will be adding some folks.
I do want to give members of the audience an opportunity to say something, if they would like to do that. We can sit for a few more minutes or whatever it takes. Is there input from the audience.
MS. BRONSTEIN: I guess I will give a little bit of input. I guess I am the only private citizen in the room, except we do vote, so I guess we all do vote.
DR. DETMER: What is your name, please?
MS. BRONSTEIN: My name is Binnie Cyril(?) Bronstein from Baltimore. I guess what I haven't heard enough of, and I did have to leave early yesterday, I haven't heard enough on concerns about confidentiality, on security.
I guess I am disturbed about the fact that everybody will be identified. I don't know why we have to be identified by race and ethnic origin and so on.
Frankly, I don't need to sound like an hysteric, but a country about 50 years ago did that and it had quite disturbing effects.
I would hate to see us going down that kind of a road. I will use a political term, a slippery slope. I guess the confidentiality and security thing does bother me.
It bothers me that Dr. X can look up your records even though he is not your primary physician, that anybody can have access to this and that it can be hacked into.
We are on such a close deadline here that things may be put into place and there are not enough security measures. So, that is my little presentation, but I am worried about that.
DR. DETMER: Thank you for your comments. Anyone else? Is there a motion that we adjourn?
(Motion made and seconded.)
DR. DETMER: Thank you very much. Have safe trips home, everyone.
(Whereupon, the meeting was adjourned.)