[This Transcript is Unedited]
Hubert H. Humphrey Building
Room 705A
200 Independence Avenue, S.W.
Washington, DC 20201
Agenda Item: Call to Order - Dr. Cohn
DR. COHN: Good morning, I want to call the meeting to order if everyone's seated. This is the first day of two days of hearings of the Subcommittee on Standards and Security of the National Committee on Vital and Health Statistics. The NCVHS as you all know is the main public advisory committee to HHS on health information policy. I'm Simon Cohn, chairman of the subcommittee, and the national director for health information policy for Kaiser Permanente.
I want to welcome fellow subcommittee members, HHS staffers, and others here in person. I want to also especially welcome those calling in on the internet, and I really want to remind everyone today at the hearings to speak clearly and into the microphone so those on the internet can hear.
We obviously have a lot to cover over the next two days. Today we're really going to be devoting the full day to discussions with the Consolidated Healthcare Informatics Initiative. They are coming forward with final standards recommendations on nine health care domain areas, and then preliminary reports on addition three domain areas. These are for use within the federal health care enterprise.
As part of the federal internal adoption process the NCVHS holds open public hearings on these recommendations. The purpose is to provide input into the federal decision making process in this very important work. For everyone's information we will be hearing additional CHI reports and recommendations at our next scheduled hearings also, which will be January 27th and 28th, 2004.
Tomorrow just to remind everyone we start early at 8:30, we'll begin with a HIPAA update and then follow with some discussions around the claims attachment standards, first an update and then a discussion. Finally for tomorrow, as many of you know the recently approved Medicare Reform legislation, which was passed I guess two weeks ago and signed by President Bush yesterday, calls on the Secretary to adopt standards for E-Prescribing, and the NCVHS has been directed to develop such standards recommendations. We will be discussing a draft project plan for the effort late Wednesday morning, and Jeff, thank you for your work putting that together. And obviously we will refine that I'm sure significantly from where we are now as we begin to mobilize around it.
Now I want to emphasize that this is an open session. Now those in attendance are willing to make brief remarks if you have information pertinent to the subject being discussed, we will also have time at the end of each session for public comments if necessary. And of course those listening in on the internet we welcome emails and other communications on any issues coming before the subcommittee.
With that, let's have introductions around the table and then around the room. For those on the national committee as always we would ask as part of your introduction if there are issues coming before us today for which you need to recuse yourselves would you please so publicly state. Jeff, would you like to start with the introductions?
MR. BLAIR: Jeff Blair, Medical Records Institute, vice chair of the subcommittee, member of ASTM, AMIA, HL7, and HIMSS, and there's nothing that I can think of that I need to recuse myself about.
DR. STEINDEL: Steve Steindel, Centers for Disease Control and Prevention, staff to the subcommittee and liaison to the full committee.
DR. ZUBELDIA: Kepa Zubeldia with Claredi Corporation, member of the committee and the subcommittee.
DR. HUFF: Stan Huff with University of Utah and Intermountain Health Care in Salt Lake City, member of the committee. I don't think there's anything that's coming before us that I need to, but I do have potential conflicts as a co-chair of the LOINC committee and as a vocabulary co-chair within the HL7 standards committee, and potentially as a conflict if there are any business about 3M that comes forward, I've been a sometimes consultant for 3M.
DR. FITZMAURICE: Michael Fitzmaurice senior science advisor for information technology to the director of the Agency for Healthcare Research and Quality. I'm staff to the subcommittee and liaison to the national committee.
MS. BEBEE: Suzie Bebee, ASPE, staff to the subcommittee.
MS. GRAHAM: Gail Graham, VA, staff to the subcommittee.
MS. BRADFORD: Alicia Bradford, Consolidated Health Informatics Initiative project management team.
MS. WARK: Cynthia Wark, Consolidated Health Informatics project management team.
MS. GREENBERG: Marjorie Greenberg, National Center for Health Statistics, CDC, and executive secretary to the committee.
MS. FRIEDMAN: Maria Friedman, CMS, staff to the subcommittee.
MS. TRUDEL: Karen Trudel, Centers for Medicare and Medicaid Services, program manager for the CHI Initiative.
MS. SQUIRE: Marietta Squire, CDC, NCHS, and staff to the subcommittee.
MS. GUILFOY(?): Elaine Guilfoy, president, Guilfoy Consulting.
MR. RENDIANACK(?): Mark Rendianack, Consolidated Health Informatics staff.
MS. LESH: Kathy Lesh, the Kever Company.
DR. COHN: And I need to publicly disclose I'm a member of the CPT editorial panel, therefore I will be recusing myself from any of the discussions related to that issue.
Now before we start I think that we were expecting Dr. McDonald, another subcommittee member, to be calling in, I guess that is, where is Dr. McDonald going to call in this morning? Okay, fine, I guess we'll welcome him when he does call in.
Anyway with that, Cynthia and Alicia, thank you very much for being here and we'll hand the floor over to you.
Agenda Item: CHI Final Reports - Ms. Wark
MS. WARK: Thank you. This is a rather full agenda today, the Consolidated Health Informatics Initiative has been busy, working hard all year and we've come to the subcommittee on several previous occasions and today we come in with nine final reports and three preliminary reports. There are several handouts on the table and I'd like to draw everyone's attention to the fact that the handouts are grouped. So there are only five handouts but there are several reports in each of the handouts, so there a total of 12 reports in the five clipped together packages.
Today we'll review the CHI schedule and the progress that we're making and give presentations on each of the nine final recommendations and the three preliminary recommendations as well. Here's the schedule that the subcommittee has seen on many prior occasions and it has remained the same throughout the year. It has been a very aggressive schedule and our teams have worked hard, our team leads are here today to deliver the recommendations from their workgroups and as you can see we are nearing the finish line, have all of our 24 domains covered by the end of December.
Again, this is the slide on the standards adoption process and the arrows on the bottom of the slide indicate when we come to the subcommittee with preliminary reports following initial analysis and feedback of the workgroup and the final reports come following technical presentations to the CHI Council, department agency review and feedback, and CHI Council establishes consensus, and it is at that point that we come back with our final recommendations to the subcommittee to obtain any additional feedback and input that we should consider before we move ahead with final adoption.
This is a somewhat modified slide from what the subcommittee has seen on our previous presentations and the reason that this is modified is that as we have moved forward with our work and come forward with recommendations it has been clear to us that it hasn't been as simple as saying there is an acceptable terminology that needs only ever greening to maintain and there may be an imperfect terminology with gaps, or no solution.
What we have now found is that there are many intermediate types of recommendations coming forward based on work that is going on in the standard development organizations and so we are trying to be much clearer with our definitions and our terminology so that when we say there is a gap to be addressed that means that the standard can be used and that we're recommending improvements to the standard through the gaps that we identify being addressed. And when we have a condition placed on the recommendation that means that the standard cannot be used until the conditions are addressed. There are also temporal issues we identify when there is work in progress within an SDO and we are aware of actions that are pending with a defined timeframe, such as ballots, or work that is ongoing to make a production version available and so we identify specifically those issues along with the recommendations so that it is clear when something is adopted what it means that there is a condition or a temporal issue.
And then we also continue to have domains where we may have no solution at this time and we may also make a recommendation that appropriate actions be taken with groups to develop a vocabulary to fill that need.
DR. COHN: Cynthia, could we just spend a minute on that slide because you're obviously changing things a little bit, I just want to make sure that everyone understands this. I guess in your reports you still have conditions, by that you mean these are things that relate to imperfect terminologies as opposed to conditional approvals, or if they're conditional approvals are they one of the three addressing gaps, conditional approval, or temporal issues? Maybe you could you just clarify that a little bit.
MS. WARK: The use of the word condition is associated with the need to have an action take place so that the standard can be ready for use. And when we identify gaps that does not mean that the standard is not ready for implementation in the federal enterprise, it simply means we have identified improvements that we believe will make the standard better.
DR. COHN: I see, so when you put a condition in a report it relates to a condition or a temporal issue?
MS. WARK: No, the use of the word condition is now being used to indicate that the standard is not ready for use, so while we may be making a recommendation that a standard be adopted there are certain conditions that would need to be met in order for the standard to be available and mature enough to use and to implement in the federal enterprise.
DR. ZUBELDIA: So it's more conditional disapproval, right? Instead of a conditional approval it's a conditional disapproval of the standard. It's disapproval until the conditions are met.
MR. BLAIR: And Cynthia, this is a change in definition for how you use conditional approval when we were doing message format standards, because in that case you used the word conditional approval, is that correct?
MS. WARK: That is correct, and I think that there has been a lot of questions about what does it mean for something to be conditionally approved. For example, if we go back to the DICOM standard, at the time that that was approved for us it was approved for use within certain parameters within an institution, for example, and so yes, the term condition has taken on a new meaning as we begin to understand some of the work that needs to take place on the standards in order for them to be successfully implemented in the federal enterprise.
DR. COHN: I think Karen has a comment.
MS. TRUDEL: I think another way to look at it is that an acceptable solution is one that is ready for use now and we couldn't think of any way to make it better at the moment. A gap is there's a solution available, it can be used now, but we can find ways to make it better and we've made those suggestions but you can put it into effect even if those improvements were not made. The temporal issues are ones where we know that there is a specific action with a specific timeline that is happening and that therefore we can approve, for instance saying we expect that an HL7 ballot will take place in January or something like that. Those are the kind of temporal approvals that we're making. The conditional really means that we found something that could work but it's not quite there and until it's there it's not usable. I don't view it as a disapproval because a disapproval means that we couldn't find anything out there at all, not even anything that was close to work with. So basically those are the distinctions we're trying to make.
MS. BEBEE: Does this change the previous conditional approvals of the previous recommendations? Or how does that fit now that it's flipped?
MS. WARK: We'll go back and revisit when we get everything, the first round finished, there may need to be some notes added to some of the previous reports to clarify how it fits but I think most of them were issues where the standard couldn't be used without the changes being made.
MS. BEBEE: Well, Steve's here, Steve, you had led a group, was it LOINC? And there was a condition? So does that, from your practical application does that work?
DR. STEINDEL: I've been involved with the discussions concerning the, what I view as somewhat subtle changes in conditional and I think one of the problems that we've been having in the past is not so much with conditional approval, where people have put in conditional approvals based on something changing in the terminology, like for instance the conditional approval on LOINC concerned them developing hierarchies, which they're doing and released, so in a sense that condition has already been met. And in other conditional approvals we are looking at something that needs to occur within the terminology and in some cases that has occurred and in some cases it has, or it's starting to occur directly because of our recommendations.
I think the biggest problem we've had in the past with workgroups coming in using conditional approval was the problem of temporal where workgroups would be coming in and saying yes, we can approve this terminology, however, it's not going to be released until, or the version that we want is not going to be released until, or this terminology is planning to make the change we're requesting but it won't be done until, and I think that is where the major change has occurred with regard to temporal. And what in a sense we're doing here is when we put a temporal recommendation on it it's sort of a check-mark for the federal government to look at this terminology on or around that date, ask has this occurred, and if it has occurred the terminology then moves from something that really is not ready for use today but can be used today without requiring any other action.
MS. BEBEE: I guess what confused me was the part that says that the standard can't be used until the conditions are addressed, and so using LOINC as an example there would still be the use of LOINC except for the need for the hierarchy to be addressed. Is that correct? So again, would LOINC in your example be used and still be conditional?
DR. STEINDEL: Looking specifically at LOINC I would say LOINC has already addressed the conditions that we have placed on it.
MS. BEBEE: Okay.
DR. COHN: Okay, I'm sure we'll come back and revisit this one as we go forward. I apologize for focusing in on this one but this obviously is a change and we just need to make sure everybody understands it.
MS. WARK: The CHI outcomes remain the same as they have been, the domains with a clear strong candidate we will adopt or acquire. The domains with standards that need gaps and weaknesses addressed we will nudge. The domains lacking in standards we will stimulate new development through the standard development organizations. And the domains with terminologies used by other processes, for example HIPAA, we will validate those.
Further coordination with the subcommittee will take place on January 27th and at that time we will have our remaining five presentations and that will complete all of the recommendations for the CHI portfolio to the subcommittee.
Okay, with that we will move to the first final report, Anatomy & Physiology, and the team lead is Steve Steindel.
DR. COHN: And Steve before we start let's just talk about process for a second. I am presuming as in previous occasions you'll take notes on comments and modifications and I think what we'll try to do is we'll have the committee vote to accept each of them as we go from one to another rather then holding them to some later dates. Is that okay with the subcommittee? I should actually ask at this point has Dr. McDonald called in?
DR. MCDONALD: Yep. Simon I can hardly hear anything.
DR. COHN: Can you hear me now?
DR. MCDONALD: Not too well but better then the previous discussion.
DR. COHN: Steve, do you want to try speaking into the microphone and seeing if Dr. McDonald can hear you?
DR. STEINDEL: Dr. McDonald, is this clear? Do you hear me clearly?
DR. MCDONALD: I can hear you, I wouldn't say it blows my ears off.
DR. COHN: Well, we're thankful for that since you have to travel to Europe later today.
DR. MCDONALD: It's much less loud then it has been on past meetings.
DR. COHN: Dr. McDonald, do you want to try calling back in again and see if the connection is better? We apologize for the technical delay.
Steve, I don't think I expected it was going to take five minutes to get him back on the line.
Dr. McDonald, are you back on?
DR. MCDONALD: Yeah, it's not substantially different but I can hear you.
DR. COHN: Well, Dr. McDonald, since you're on the line and actually going to be a voting member for this discussion --
Dr. McDonald are you still there?
DR. MCDONALD: I didn't hear much in between there.
DR. COHN: Well, that's probably just as well. Anyway, we just wanted you to take a minute and have you introduce yourself as well as if there's any things that you need to recuse yourself for for this conversation.
DR. MCDONALD: This is Clem McDonald from Regenstrief Institute and Indiana University, and I will recuse myself from any votes or discussion about LOINC, I don't know what's really on, if that's on the agenda today.
DR. COHN: Okay, thank you. Steve, would you like to start off?
Agenda Item: CHI Final Reports - Anatomy & Physiology - Dr. Steindel
DR. STEINDEL: Yes, thank you, I'd like to present the report from the Anatomy & Physiology CHI domain to you. This team was a mixed group, we had a member from NASA, a member from VA, a member from the Department of Defense, a member from NHI, in particular at NCI, and I asked a colleague of mine at CDC, Dan Pollack, to join the workgroup except I was very specific when I asked him, I asked him to join as a clinician and not as an epidemiologist because I thought we needed some direct clinical input on the Anatomy & Physiology team. And Dan was very gracious and actually wore that hat throughout the whole discussions and it was very helpful.
The first thing we did was to define the scope of the domains and we developed an extensive scope for the anatomy domain, used to describe anatomical locations for the following purposes. One was clinical, the site a procedure, and the location of an observation. Another was surgical, then it was pathology, and finally research. And that we felt that the anatomy terminology to be useful had to have certain characteristics from an informatics point of view. It had contained and is a hierarchy, a partonomy(?), or a part of hierarchy, had to be able to indicate laterality, anatomy uses a lot of synonyms so synonyms had to be accommodated, and we also felt that lacking a standard term we used the term virtual locators to describe features of anatomy.
Anatomy terminologies tend, anatomy themselves tend to describe lots of subparts and a lot of times they use a concept that may not actually physically exist to describe the collection of those subparts and SNOMED actually is one terminology that does use virtual locators where they have a liver structure and they define parts of the liver with respect to that liver structure, and they also have another concept, which is the entire liver, but it does not have parts.
We also felt that the anatomy terminology must accommodate modifiers because anatomists tend to use those a lot, and it must have compatibility with animal models, not necessarily as a decision factor but because we do reach out into the non-human world. So that was our definition, our scope, for the anatomy terminologies.
For physiology we looked at that first from the narrow concept and we found that we couldn't really accommodate that with standard terminologies that are used in the medical care profession. So we defined physiology as being used to describe or infer human physiology, at least at the organ system, cellular and biochemical levels. Physiology terminology includes tests that are used to infer physiological states at any level noted, so this is an extension of the standard way people think about physiology in that we decided to include in our definition tests to infer that and we did that primarily because when clinicians talk about physiology they generally don't talk about physiology from a scientific viewpoint, they talk about it from the tests that they have run. If they want to talk about lung function they don't talk about how the cells in the lung are behaving, but they talk about the test that was used to infer that the cells and the lungs may not be behaving properly.
So that was our definition and one thing, because we do feel that is an extension of the definition and as we'll note later we do not make a recommendation in this area, this may be something for the committee to make a comment on. We then defined the domains and the sub-domains and that's pretty much an abstraction of the discussion that I just made.
We looked at as alternative terminology the Medical Subject Headings, or MESH, the National Cancer Institute Anatomical Terminology, which at the point the study was done which was the summer was in early stages of development, SNOMED CT, Clinical LOINC, the Fundamental Model of Anatomy from the University of Washington, the HL7 Site Table, and the Veterans Administration NDF-RT Physiological Effects Hierarchy.
I would like to point out at this point that we actually have approved the NDF-RT Physiological Effects Hierarchy as a CHI recommendation for the Medical Domain. It doesn't have any bearing on what we're going to be talking about here, but it is an approved CHI terminology.
For anatomy our prime recommendation was a set of hierarchies that are found in SNOMED CT and those hierarchies are noted. They are primarily found in the body structure area of SNOMED as one might expect. We also noted that SNOMED does have a good set of modifier terms and we noted where those are found and which hierarchies. So our specific recommendations for SNOMED CT and our primary recommendation is a series of hierarchies found in SNOMED.
One important contribution that came from Dan Pollack from a clinical point of view was that as a clinician where he has to do things in the emergency department and note what's going on he felt that he would be greatly inhibited if he had to select terms from an anatomy terminology that contained somewhere between 3,000 and 5,000 terms, which is what these terminologies usually have. And he felt that it would be very useful to have an abbreviated form of the anatomy terminology for clinical purposes.
Our conditional recommendation for that was the HL7 Site Table. The HL7 Site Table could be a subset of SNOMED, it could be independently developed. It exists at this point in time but it is too weak to rise up to the level of a full recommendation, but what it is in essence can be viewed as as a subset of maybe 100 to 200 anatomical terms that would be useful for clinicians to use on a day to day basis.
Our third recommendation delved into the area of research. As you noted that was one of the areas we felt was in scope for the anatomy domain. Research SNOMED is a clinical terminology and usually does not have terms at the subcellular level where the research people actually work and we felt it was important to recommend a terminology to use there. If you recall a similar type of recommendation was made for laboratory result reports, where we said that for research purposes we would recommend the NCBI terminology as an extension to the SNOMED terminology. And in this particular case we recommend the NCI Thesaurus for research purposes. We also note that this has been developed into a fairly good and well ordered complete terminology.
The NCI Thesaurus at the time that this was written was not complete, it was still in development stage, I have checked with NCI recently, the terminology is anticipated to be certified as complete somewhere around the end of this month, first of January, it's going into internal test at NCI the January/February timeframe and is expected to be released as a production product in the March version of the NCI Thesaurus. So we feel that temporally this terminology is well on the pathway to be put into use.
We also note that NCI has developed and will make available a map between their anatomy terminology and the others. We also noted that mapping is very important in anatomy terminology --
DR. FITZMAURICE: Steve, by the others do you mean both SNOMED and the HL7 Site Table?
DR. STEINDEL: HL7 and in general for any anatomy terminology, because anatomy terms tend to be synonymous and it's fairly easy to do one to one mapping.
One terminology we did not recommend was the Fundamental Model of Anatomy from the University of Washington, one reason we did not recommend it was we felt it was not a clinically related terminology, it was more developed for views of anatomy and primarily for imaging. But it is still an extremely good model of anatomy and mapping could make it very useful.
DR. FITZMAURICE: Just a little bit aside but if you're looking at an x-ray and you want to describe say where a tumor is located in the lung, is that an example of where an anatomy could be helpful in showing the dimensions and where it's physically located on that picture? Or is that out of the scope of the anatomy?
DR. STEINDEL: I'm hesitating in answering that for several reasons, one is because DICOM has released specifications on how to locate spots in an x-ray. Also I believe the Multimedia group maybe making some recommendations in that report. But having said that what DICOM does is DICOM provides a structure to give the coordinate location of the spot in the x-ray that they're talking about and then tends to use SNOMED as the anatomical term to describe the features that they are seeing. So the answer to your question is most likely yes. But it wasn't not considered in scope.
Are there any other questions on anatomy?
DR. MCDONALD: Steve?
DR. STEINDEL: Yes, Clem?
DR. MCDONALD: Could I ask a question? The digital anatomous one, did you look at that hard, or how did that compare? What was its issues/problems?
DR. STEINDEL: That was the Fundamental Model of Anatomy --
DR. MCDONALD: The model was what the problem?
DR. STEINDEL: That's the Fundamental Model of Anatomy that I mentioned just a moment ago. The reason why we didn't, we excluded that was we felt that the model itself, the way it was put together was more aligned towards people who were looking at anatomical structures and images and not aligned to clinicians. But we did feel the model itself was an extremely good model of anatomy, we just didn't feel that it was useful for the specific aspect that we were making the recommendation for. We did note very strongly that the model is usable from a mapping point of view because we can map terms in the fundamental model for virtually all cases, one to one, to SNOMED terms. And of the terms that we looked at the ones that we couldn't locate, couldn't map one to one, had to deal with the concept that we referred to as virtual locators, where SNOMED would create an anatomical type structure that didn't exist that would differ from what the digital anatomous, the Fundamental Model of Anatomy, would create as their same type of virtual locator. They both used that concept.
MS. GREENBERG: The Fundamental or Foundational?
DR. STEINDEL: Foundational, I always use, I did it in our workgroup discussions as well, thank you for correcting me, Marjorie.
DR. COHN: Steven, are you done with your presentation and the floor is open for --
DR. STEINDEL: Physiology just very quickly. For physiology we actually found that while there were a number of individually developed research level terminologies for specific projects in the area of cellular physiology, outside of the NDF-RT physiological axis we found no nationally proposed terminology for physiology. We also found that the NDF-RT physiology axis was highly specific for drug physiology and was not suitable, hence we made no recommendation for a physiology terminology at the cellular level, which is the traditional view of physiology. And we do not know if we want to recommend that that be created because I do not know how well that's used in clinical medicine. I suspect outside of research it's not really used at all.
Going to our extended definition where we talked about the tests that are used to infer physiology we found that there were two terminologies that do explore that area in different ways and we found problems with both terminologies and could not make a recommendation. Those two terminologies were SNOMED CT and the clinical axis of LOINC. We have approved the laboratory axis of LOINC but we haven't looked at the clinical axis. The reasons for disapproval in both cases differ, they both would require extensive work to clear the condition and we can explore that further if we wish. But my recommendation, and it's not noted in the report that we're passing on to you, is that if we do wish to explore the expansion of the physiology axis into tests it's probably more worthwhile to do it on a more detailed level that can be accommodated at this time.
Now I'm open to questions.
DR. COHN: Now you're open for questions on both anatomy and physiology. Well, first of all I want to thank Marjorie for clarifying the name of the model from the University of Washington since somebody might be actually listening in from Washington.
Actually I had a couple of questions here, and I think I want to start with anatomy. And I guess I'm getting a little confused here and I think it gets down to the issue is I think about at the end of the day of simplicity versus how much redundancy we seem to be willing to accept. And I actually do understand that mapping is the solution to all problems but I also worry to that certain point maybe people have to make a decision or there has to be some things that are sort of in and out of scope, and so I guess I'm wondering first of all, I mean it seems to me that you are recommending sort of unequivocally the SNOMED CT terminology for at least the intended purpose here. I guess I'm trying, in my own mind, trying to figure out the HL7 Site Map.
Now once again being also an emergency physician who talks about femurs and elbows and arms and all that stuff, the last time I looked that seemed to be accommodated in SNOMED though it was mixed in with very much more granular concepts at the same time. So I was trying to decide about the independent utility of the HL7 Site Map as opposed to it being some sort of a subset of SNOMED CT, which being used as an interface terminology would seem to me to be more where you were going with that. So that was question number one.
And question number two is the NCI Thesaurus, which I'm actually not as familiar with as probably should be, but I'm certainly supportive of the idea of extensions into areas where a currently existing terminology doesn't have strength, which I think was what we accepted in your previous recommendations. But I'm getting an unsettled feeling that the NCI Thesaurus as you're describing it here seems, there's not just extension but there's significant overlap.
Now once again I'm obviously talking about two different issues but maybe you can address both and why we don't change this recommendation, or recommend changing the recommendation.
DR. STEINDEL: To address the first point the HL7 Site Table. In either case, whether we changed the wording and we say that a subset of SNOMED be created, or that we leave it as the HL7 Site Table, the main purpose of this recommendation was that a subset of a large anatomy terminology be made available to clinicians. Now from a practical point of view our recommendation for the HL7 Site Table clearly states that that table needs to be redone, it is inadequate at this point in time. The discussion came up at the CHI Council meeting regarding the creation of a new HL7 Site Table and Betsy Humphreys from the National Library of Medicine raised exactly the same question as you did.
And the comment from a subcommittee point of view, a workgroup point of view on this, was very simple that if we need to create a new HL7 Site Table the HL7 Site Table could be created as that subset of the SNOMED terminologies.
DR. MCDONALD: Steve? HL7 may have some problems with its international constituency if they can't have the codes that would be distributed without special licensing.
DR. STEINDEL: Well, it's my understanding that Betsey Humphreys is planning to discuss this with the HL7 vocabulary group to see how we can accommodate this. It may be perfectly reasonable that SNOMED would be willing to place those 100 odd terms free for use. But Betsy is well aware and that is the approach that she would like to take on this so in essence we are merging the needs of using SNOMED and SNOMED having the terms and isolating those terms in a separate fashion to make it readily available for everybody. And I'm encouraged that that approach would be happy in either a change in the recommendations from this subcommittee or in the letter that we transmit to the Secretary to specifically note the use of SNOMED to form that new version, a subset of SNOMED, so I agree in essence with what you're saying.
Now with regard to the second issue with regard to the NCI Thesaurus, we do actually have the same problem with respect to our approval for the NCBI terminology for organisms where SNOMED contains approximately 18,000 living organisms and the NCBI Thesaurus contains over 100,000. There is this need in the research community, which does do federal exchange of information, to have terminologies that do meet their needs, and just by the nature there's going to be overlap. What I basically feel in this particular case, we are meeting the needs of the research community and if we exclude the NCI Thesaurus we are not meeting their needs and the specific recommendation is for the use of that terminology for research purposes. So I do feel we are covered.
I think we have a little bit of a practical problem in that within NCI itself and their clinical trials which routinely use the NCI Thesaurus, starting in March the anatomy terminology they will be using is the NCI Thesaurus Anatomy Terminology. And it is mapped within SNOMED CT.
DR. COHN: Okay, so basically one question and then I'll let Susie have a question. So basically I guess what I'm hearing is that we probably ought to accept the NCI Thesaurus though it is noted here that it is anticipated that the anatomy terminology will be included in the NCI Thesaurus by Fall, 2003, this is in your recommendation and conditions, and I guess you're now stating that it will be there in March 2004.
DR. STEINDEL: It will be in production in March 2004.
DR. COHN: So that would be the revised condition. Susie?
MS. BEBEE: I support the idea of a subset from a practical standpoint, my question, I wonder how the anatomy perspective is different from a subset of anything else, supplies, diagnosis, any other type of vocabulary that the anatomy would need to have a subset versus other types of domains.
DR. STEINDEL: Well, I think you're statement is quite correct, I mean most domains are going to be developing smaller subsets of work terminologies that they would be using. I think the major difference in this particular case is we're really looking at kind of an inverse terminology use then in a lot of the other domains. There is a specific generally used clinical subset of anatomy that clinicians tend to use. As Simon pointed out they're major body parts. When people, the people who actually use the finer levels of anatomy terminologies, it's a very small subset of the clinical specialties, like pathology would use them, radiology might, surgeons might, but in those particular cases like for surgery or radiology they might create their own subset of a more extensive use of anatomy terms but in general the full anatomy terminology really is not used by most of the clinical community and therefore we felt it was very important to note the need for a specific agreed upon clinical subset since the main terminology is not used extensively.
DR. COHN: Jeff then Stan then Kepa.
MR. BLAIR: Steve, as I listened to you step through the anatomy and physiology CHI recommendations one thing that I observed was that the priorities that CHI is using in this case are a little bit different then the priorities NCVHS has been looking at as we've gone through the PMRI terminology selections in that, well, actually the PMRI standards and terminologies, in that we came, we began with the perspective, NCVHS began with the perspective of identifying clinical data standards and terminologies with a vision that we would be capturing information once at the point of care, clinically specific information once at the point of care, and using derivatives of that information for all other purposes. And here as you've gone through anatomy and physiology it appeared as if you were focusing on it from a research and public health standpoint and then recognizing the need for its use in patient care and therefore recognizing that there needed to be a subset, the HL7 Site Map or a subset of SNOMED. It appeared to me as if what you did accommodated the fact that you're coming from a different perspective but I wanted to bring up this point in case you wanted to make any other observations about it or whether this difference in perspective does create a problem that we need to consider.
DR. STEINDEL: Jeff, I think the only reason that you sense a difference in perspective was primarily because of the unique way anatomy and physiology is used in the clinical community. That as I noted in answer to Susie's question it's not really widely used in the clinical community, the broad terminology, and that when we were charged with this we looked at the broad terminologies and focused the clinical need as a subset and I think it's just unique to this terminology domain.
DR. COHN: Kepa?
DR. ZUBELDIA: Steve, help me understand what's the advantage of having HL7 Site Table as a separate new creation to accommodate the subset rather then having a recommendation to SNOMED to identify the subset and keeping it within the SNOMED domain.
DR. STEINDEL: I think the answer to that was at the time this was written is basically availability. As I noted this was done over the summer timeframe, it was actually started before the federal license was put into effect and we are still not sure of the distribution mechanism for SNOMED and the ability to create subsets within SNOMED, but we are sure of the ability to create subsets within HL7 so we focused it on something that we felt comfortable in it happening.
DR. ZUBELDIA: And the HL7 recommendation is a conditional recommendation?
DR. STEINDEL: Yes, because it's not there yet. Those on the internet, Kepa and I are sharing a microphone.
DR. COHN: Stan, do you have any comments?
DR. HUFF: Well, I had a new question. Could you say more Steve about what needs to be done in physiology? In a sense I guess I would be looking for input that would allow LOINC or SNOMED or somebody else to correct, it's not clear what needs to be done there based on what you said, so could you say more?
DR. STEINDEL: I actually, the basic answer to your question is yes, the report that was transmitted to NCVHS did not go into specific details of what we found were the problems with Clinical LOINC and SNOMED, but I could go back to the report and look at the exact recommendations and wording, but as a first approximation the problem that we found in Clinical LOINC was basically that it was not complete. The areas that it did define the tests it was very good but it was omitting a lot of clinical areas like neurology, etc., so our feeling about Clinical LOINC was one of lack of completeness.
Our feeling about SNOMED was one of confusion. We found that when we looked a the EKG area of SNOMED what they had done was brought in what they had before, merged it with the read terms, and it appeared that several of the same, what we would consider to be physiological tests appeared in several locations. Some of it was not cleaned up, for instance EKG and ECG were used interchangeably. We looked for items that are very important like an electrocardiogram, like results from various leads, and while they're clearly defined in Clinical LOINC they are not clearly defined in SNOMED. So what we found was there was a tremendous amount of lack of clarity in SNOMED. What we were saying were both terminologies before they could be looked at from a recommendation point of view needed extensive work.
DR. COHN: Karen and then Carol Bickford.
MS. TRUDEL: I'd like to clarify that the first phase of CHI, which is what we are in, is to roll through the first set of 24 domains and do what we can in terms of making recommendations and adopting standards. Our second phase will involve developing work plans for each of the areas where we've identified gaps or conditions or where there are, there was no recommendation because there wasn't anything that was ready, set up teams to go back to the SDOs to work on those areas and address them as we can. So phase two will actually have teams going back to each of the organizations saying this is what we would like to work with you on.
DR. COHN: Carol Bickford?
DR. BICKFORD: Carol Bickford, American Nurses Association. I have two questions. The first one is when you were looking at anatomy and physiology has there been consideration of the complementary and alternative approach to health care? Or is it strictly the Western medical model?
DR. STEINDEL: It was strictly the Western medical model but I can say that I think that might have a bearing for consideration for physiology, where we did not make a recommendation, but I'm not sure it has a bearing with anatomy.
DR. BICKFORD: There may be an approach from the anatomical perspective that there is a different structuring, for example energy zones or, and I'm just picking this off the top of my head, from the standpoint of how the system is structured vice the way we look at it form the Western model.
The second question --
DR. STEINDEL: So that would be a difference in the hierarchical arrangement?
DR. BICKFORD: Correct, and labeling and different concepts that may not be present.
DR. STEINDEL: Before you ask your second question I would like to comment on that. This was one thing we did observe as we went through the multiple anatomy models that we looked at is anatomy, the person who develops an anatomy model tends to take a different approach to hierarchical arrangements and it's very, very common to see anatomy models using a different hierarchical approach, and an anatomy model can easily accommodate alternative medicine approaches and we would still be able to map them to the actual anatomical structures.
One comment that I would like to make just as an aside on this is one comment I made throughout these discussions is that when I look at an anatomy model is I like to be able to produce the body, walk up the tree and finally get a body, and SNOMED you actually cannot do that because of the way they've arranged their hierarchies and we're still recommending it so you can see we can accommodate multiple approaches. Second question.
DR. BICKFORD: The second question is in relation to the definition of the tables, who are going to be the experts that are going to be helping make those decisions? And I raise that as a question from the standpoint perhaps of those who, physical therapists or occupational therapists who may have an entirely different view of the physiology or the anatomy tables, the subsets that we're talking about. Or are we going to allow that to be an information system solution, which is a personal view, rather then being reliant on sub-tables?
DR. STEINDEL: I'm assuming you're asking that question in relation to the HL7 Site Table and whether the subcommittee feels this should be an HL7 Site Table or a subset of SNOMED, the wording would be that the charge to the SDO considers the inclusion in the process that they used to develop the subset table.
DR. COHN: Let me try to move things back around here, I actually hadn't realized you were going to move us off into the physiology area because I was trying to get closure on the anatomy, that's no problem, there's no way to know what the questions are going to come up. We've talked about a lot of things but haven't made any decisions on anything, or reached any conclusions on anything. I think what I'm hearing, and let's go back to the anatomy and talk about what we just discussed and see where we are.
Now first of all I've just looked through these recommendations, I find them a little confusing because it appears that we have recommendations excepting SNOMED CT, HL7, Site Table, and NCI Thesaurus, and then at the very end there's the conditions relating to two of the three, so it's probably, I don't know what the rest of them look like but probably conditions that relate to recommendations probably ought to be a little closer to the actual recommendations themselves lest someone such as myself think that we're actually recommending these three as opposed to really giving temporal conditions on two and then recommending one other. That's just an organizational issue.
Now having said that I think what we seem to be sort of basically agreeing to accept SNOMED CT, I didn't hear many objections to that. Now I am hearing, or at least maybe I brought up and never received a resolution on, this issue of what we're doing with the HL7 Site Table or is it the SNOMED micro glossary, and I think what I'm hearing is a general view that there needs to be some sort of a micro glossary, smaller set of clinically useful anatomic terms developed, and I think that we're trying to say is that this ought to be done in collaboration potentially with HL7, that there ought to be a discussion that occurs with them, that this should be used to populate the HL7 Site Table. I think that's what I'm hearing, or it would certainly be very convenient if that happened. And that we should encourage discussions between the CHI group, SNOMED, and HL7 to try to make that a possibility, I mean given that we don't, I mean we actually have two representatives from HL7 at the table but they both were recused from discussing this I believe, and Stan's nodding his head so I think that's probably right. So I guess the question is what do we want to specifically do about the HL7 Site Table. Cynthia, sounds like you have a suggestion? No, I saw some body language --
MS. WARK: No, I was just listening to your proposal about possibilities of the organizations working together and it sounds like --
DR. COHN: Well, it sounds like a phase two recommendation that as Karen was sort of talking about it. Steve and Kepa.
DR. STEINDEL: Simon, I would like to point out that the wording on the condition on the recommendation for the HL7 Site Table, if you look at the last sentence, it is further recommended that the present and added terms to the Site Table be closely coordinated with the corresponding SNOMED CT terms. So the subcommittee is already thinking, the workgroup that put this together is already thinking in the mode that you are thinking and it could be an NCVHS recommendation to make that sentence stronger, and I think that would --
DR. COHN: Okay, are we generally, Kepa, did you have further on this particular point?
DR. ZUBELDIA: Yes, and I would like to make that sentence not just stronger but more specific, and instead of closely coordinated I would say based on, and the added terms to the Site Table be based on the corresponding SNOMED CT terms, assuming or presuming that they will be able to manage some licensing arrangement where those terms can be distributed for free.
MR. BLAIR: Kepa, would you consider Clem's observation that it could be based on that, but it also would need to come up with some mechanism to accommodate the international community?
DR. ZUBELDIA: Clem, what's the concern from the international community, is it the cost or the licensing terms of SNOMED?
DR. COHN: Did we lose Dr. McDonald?
MR. BLAIR: Maybe you're on hold, or maybe on mute.
DR. COHN: I think we've heard his concern before, I don't know that he needs to restate it, and I guess I would be concerned us being too pejorative but --
DR. ZUBELDIA: The concern that I have is that we may end up with a close coordination that ends up being different, and I like to avoid redundancy that is not redundant, that is just different.
DR. MCDONALD: Kepa, I couldn't hear that.
DR. ZUBELDIA: Clem, a concern that I have, can you hear me now? A concern that I have is that with the HL7 Site Table being closely coordinated with SNOMED, that close coordination could end up resulting in a different set of terms that are close but not based on SNOMED and the proposal that I have is that instead of just saying close coordination is that the HL7 Site Table be based on the SNOMED terms and be a subset of SNOMED. And this of course presumes the ability to license those terms out into the public domain or something where the international community could use them freely.
DR. MCDONALD: Well, the question is if you set it up so that HL7 can't do international and you don't already have an agreement, that seems like a really tough statement. I think you either have to say if they can make the arrangement they wouldn't, because otherwise you basically break their international policies.
DR. COHN: I think we need to come up with wording here that gives people some flexibility, I mean we're, considering that we really don't have an HL7 representative here that's able to speak freely I think we're in sort of an awkward position beginning to, I don't think we're in a situation where we want to tell a standards development organization what to do, I think we want to basically get them working together to sort of solve this problem.
DR. ZUBELDIA: So how if we recommend that SNOMED and HL7 work together in achieving one common set.
DR. COHN: Karen?
MS. TRUDEL: I think the recommendation that was already put on the table does that, it suggests to the Secretary, recommends to the Secretary, that the CHI project approach both HL7 and SNOMED to do just that to the extent that it's feasible.
DR. COHN: And so this is additional wording to what we have in the recommendation or a suggestion from the subcommittee. Steve, were you able to capture that? Have you --
DR. STEINDEL: I'm not totally certain.
MS. TRUDEL: I've got it.
DR. ZUBELDIA: It may be that some of SNOMED terms have to change based on the HL7 international recommendations.
DR. MCDONALD: But what you're saying is the international countries then have to buy a SNOMED license unless something changes, so I just think it's dangerous to set something that can't maybe be done unless both parties agree. Plus you also have the problem, these are balloted things that may be very difficult to create a ballot if there isn't already a SNOMED term. So I think there has to be some flippage in the process to allow it to happen right.
MR. BLAIR: Would it be possible to finesse this by indicating that it is the objective, that the coordination between SNOMED and HL7 and the international community be done with an attempt to reach certain objectives, one is that the HL7 Site Tables be based on the SNOMED terms, that's an objective, not a requirement but an objective or a target. And the same thing is that the objective is also that it accommodate international requirements.
DR. MCDONALD: Well, it's more clear then that, I think if you're going to make it an objective it would have to be that they would be free for use internationally under the HL7 arrangements, otherwise you're really putting an independent organization in a bad situation with its members.
DR. COHN: I want to bring this back in on this one, actually I thought Karen's suggestion actually meets all the needs I've heard so far. Steve, do you have a question of clarification about that?
DR. STEINDEL: Can I hear Karen's wording one more --
DR. COHN: Why don't you restate because it sounds like --
MS. TRUDEL: The recommendation was that the CHI group approach HL7 and SNOMED to encourage discussions between the two to develop a micro glossary.
DR. FITZMAURICE: To the extent possible.
DR. COHN: To the extent possible. Steve?
DR. STEINDEL: Simon, the way the wording is right now in that area, it's the second sentence starts, this is after it says the table is not complete, HL7 has a mechanism to facilitate the addition of these terms through their vocabulary technical committee, so we're already saying that HL7 has the mechanism and that mechanism should be used, which addresses Carol Bickford's question and Clem's point about balloting, etc., that the HL7 mechanism would be the one that was used. As Kepa brought out and as Clem brought out, I'm suggesting that the next sentence be modified, it is further recommended that the present and added terms to the Site Table be based on the corresponding SNOMED CT terms and be freely available for international use.
DR. MCDONALD: For international use.
MS. TRUDEL: That's outside the scope of CHI, international. It's outside our scope.
MR. BLAIR: Is it really?
MS. TRUDEL: Yes. This is an electronic government initiative for the federal sector.
MR. BLAIR: I understand, but don't we have to, individuals travel internationally right now, they get care internationally, our troops are international, so I think that some consideration for international agreements needs to be at least considered.
DR. STEINDEL: Jeff, with regard to the comment about troops, etc., the federal SNOMED license actually covers our use in those situations and I can understand Karen's point and we can actually make the comment of international use, not in the change to the CHI recommendation but in the note that we put to the letter to the Secretary. That could be an NCVHS comment.
DR. COHN: I think similarly maybe, I think what we've gotten so far on the conditions, what we're actually now talking about is also an additional action, that we're encouraging CHI to work with HL7 and SNOMED to help make this happen, I don't know whether that's part of here or whether that's in the cover letter. Or do you not want to do that?
DR. STEINDEL: Well, I agree with Karen's, we have actually in a lot of CHI recommendations put steps that CHI needs to coordinate in phase two but we've never specifically used the words CHI should coordinate. What I'm hesitant about doing is just actually introducing those specific words in that we haven't done it in the past.
DR. COHN: Why don't we, isn't this part of the cover letter? Of our letter to --
DR. STEINDEL: Actually what I was making the changes was in the recommendation itself and we could add the statement that you were saying in the cover letter.
DR. COHN: Well, maybe that's really where it should be so we don't introduce it --
MS. TRUDEL: We can't change these recommendations without going back to the CHI Council and through our whole process again, so I would very strongly suggest that what we focus on today is what the NCVHS wants to tell the Secretary because if the NCVHS recommendation and whatever suggestions are made, that will need to be discussed back through the groups to work them into the reports.
DR. STEINDEL: Karen, it's been our process in the last set of recommendations to word the transmission over to the Secretary where there were felt that there were changes that should be made to word it that we agree with the CHI recommendation except as noted in the modifications.
DR. COHN: Well, we concur but we also have further suggestions sometimes in the cover letter.
DR. STEINDEL: Yes, in the cover letter, and then items that we felt were outside of the CHI scope we have added in that paragraph and the international use might be something that we feel is outside the CHI scope, and we would add it.
DR. COHN: And I think this next step issue is also probably in that note. Now so coming back and before we spend all day talking about anatomy and physiology, where are we with this? I think what I'm hearing is that we made some modifications to the conditions for the HL7 Site Table, I think we're recommending in the cover letter basically concurrence of the anatomy recommendations with a statement that we want CHI as a next step to begin discussions with HL7 and SNOMED about creating this really micro glossary to be utilized in the Site Table I think is what we're sort of saying here. And I think we're saying that we should further recognize that, do we want to say something about international in this cover letter at all or do we want to just leave that alone thinking that they will be talking and deal with it as they deal with it? I'm not sure there's really a terrible value to go there in this letter. Stan?
DR. HUFF: So, I mean yell at me if I'm getting into the area where I shouldn't. I think the process is in place that are going to take care of the concerns that the committee has, so a couple of things. We're aware of the SNOMED contract, the SNOMED content into the ULMS meta-thesaurus. HL7, the co-chairs of the vocabulary committee have met with Betsy Humphreys and members of the National Library of Medicine for the purpose of placing HL7 terminology into the meta-thesaurus as well. And what that would mean is that for a particular use in a message those set of concepts would be a subset of the concepts in the meta-thesaurus and because of SNOMED being there as well you have the cross mapping of those terms to the SNOMED concepts. So I think in the end we're going to end up with a unified set of things that are a known subset of SNOMED so I think the process is in place that the goals as stated in CHI are going to be achieved by that process anyway.
And so I guess I'm not so worried about the particular wording because I think all of the people are motivated to do what you're trying to have expressed here anyway. We do have to work through some of the international issues but again I think that will happen because of the central coordinating activity that's going on within the National Library of Medicine.
DR. COHN: Well, Stan, I think it's fine for you to say that and the question is is are we okay with this comment in the letter or are you speaking against comment in the letter because you think --
DR. HUFF: What I was saying, I'd be happy with any wording because I think this is going to get taken care of anyway, so whatever's most appealing to people I'd be in favor of.
DR. COHN: Well, I'm going to suggest we maybe not include the international issue in the letter unless, I don't know, Steve, what do you think?
DR. STEINDEL: I was just expressing the wishes that I thought I was hearing from the HL7 community and I also agree with Stan, I think it's redundant, it would be approached during the discussions. I do have a bit of a problem as Stan has pointed out in specifically, and this is putting on my NCVHS hat, as Stan has pointed out about including an actual statement that CHI should work with HL7 and SNOMED to coordinate the development of a new Site Table because as Stan has already noted the National Library has already started these discussions. And introducing CHI into this process --
MS. TRUDEL: HHS?
DR. STEINDEL: Well CHI is, I would just, I personally would just leave it as it says in the recommendation where it is open as to who is doing it because the National Library has been recommended by the NCVHS in the PMRI terminology letter as our point of contact for SNOMED and for this type of work. And CHI actually extends beyond HHS.
DR. FITZMAURICE: So then should the committee recommend that NLM do this? Do the coordinating part?
DR. STEINDEL: I just think we should, if we are already saying that it needs to be done in the recommendations and CHI as a body as their punch list from the first set of recommendations will note that this is something that they should follow up on and will talk with the appropriate members of CHI to see that it gets done. So I think they will coordinate it without being specifically told in this letter to coordinate it.
DR. FITZMAURICE: I'm more comfortable with that.
DR. COHN: So you just want a concurrence?
DR. STEINDEL: Actually it's concurred as modified because we have had, I think I've heard two modifications to the recommendation itself, the first modification was to the NCI Thesaurus, it will be released in March instead of Fall, the second was Kepa's modification that it should be based on and not coordinated with, he wanted to see stronger language there for the Site Table.
DR. COHN: What is the wishes of the subcommittee on this one? I think the question is how inferential we are versus how direct and I'm not sure I have a strong feeling one way or another. I see Kepa sort of agreeing.
DR. ZUBELDIA: After all I've heard I think that based on, or in conjunction with, is going to end up with exactly the same result. At some point I'd like to concur with some recommendation without any changes.
DR. STEINDEL: In that case Kepa since it was your recommendation the change in date is real, so that would occur, but that's basically a trivial modification and if you're happy if being close and coordinated with then we have --
MR. BLAIR: I would move, make a motion, that we would concur with the CHI recommendations with the one modification that Steve has mentioned with respect to the availability date for the NCI terms.
DR. COHN: Okay, and I think this applies to the anatomy, we'll move to physiology in just a second.
MR. BLAIR: Physiology, I'm sorry.
DR. COHN: No, this is just anatomy, I think I want to handle, I think there's an issue about physiology we need to discuss. But all in favor, I mean any further discussion on the anatomy piece? All in favor?
SUBCOMMITTEE: Aye.
DR. HUFF: In favor but abstaining.
DR. COHN: In favor and abstain. And I guess Clem, do you need to abstain or not?
MR. BLAIR: Clem, are you on mute?
DR. MCDONALD: Yeah.
DR. COHN: Are you in favor or did you abstain?
DR. MCDONALD: Me? Oh, I guess I didn't hear the request for a vote.
DR. COHN: We were trying to bring the anatomy to a close.
DR. MCDONALD: I missed some of the wording but I think I was agreeing with, it came out as what Steve was saying, we'd leave it kind of like it was and as Stan was saying?
MR. BLAIR: we have a motion that we concur with the CHI recommendations.
DR. MCDONALD: I support that.
DR. COHN: So passed then. Okay, now let's talk about physiology for just a second since that really is a separate domain. Currently the recommendation is, I think no recommendation with, I see on our slide here it says reevaluate in 12 to 18 months, I'm not sure that I even see that in the recommendation, I just see no recommendation as the recommendation on this. I guess to me the only question on the physiology is as part of any sort of reevaluation does there need to be a reevaluation of the definition of the domain? I mean it's really unclear to me even despite all the conversation whether we're talking really about physiology or physiologic effects or what here. And I think that part of the problem here is that there seems to me to be a certain fuzziness about what it is we're really talking about.
DR. STEINDEL: I think Simon you are quite correct in your observation and this was a problem that the CHI workgroup had in defining the domain. And I think you're also correct in your observation that this is not something that we really can decide in a few minutes, which is one reason why it is no recommendation. We did not see a clear area, we do not know if the extension that we placed on the physiology definition is acceptable to the community, and we should find that out.
DR. COHN: Well, I guess then what I would do is to move a, that we concur with the recommendation to further, and this is a cover letter issues is that we recommend that it be reevaluated in 12 to 18 months and that that work begin with a reevaluation of the definition of the domain.
MR. BLAIR: I'll second that.
DR. COHN: Does anybody else have any other comments?
DR. HUFF: Well, I would just concur with that as well, I mean I think the more explicit the roadmap of what needs to be done is the better we could address it, so I mean if there were a list that said we think pulmonary function tests are the most important, next thing are EKG potentials, the next thing are visual evoked potentials, etc., if we had a list like that's something that could be targeted by SNOMED or LOINC to address those areas and fill the gaps that are seen. So we would welcome I think that kind of refocusing of the work.
DR. COHN: Steve, are you comfortable with what we just came to?
DR. STEINDEL: Yes.
DR. COHN: Okay, any further discussion, comments? All in favor? Concur?
SUBCOMMITTEE: Concur.
DR. COHN: Kepa? Stan, do you need to abstain?
DR. HUFF: In favor but abstaining.
DR. COHN: Okay, Clem? Dr. McDonald? I guess he's not answering. Well, I think we have a majority of people here so we move forward.
Well, we've made it through the first two domains.
DR. STEINDEL: No, that was one domain.
DR. COHN: It counts two to me.
MS. WARK: The next one will be quick.
DR. COHN: Okay, the next one is billing?
MS. WARK: I think we're running a bit behind schedule, should we take a break now or should I go on with the billing one?
DR. COHN: Should we take a ten minute break then?
DR. STEINDEL: Why don't we do it?
MS. WARK: Okay, we go ahead with billing before the break?
DR. COHN: I'm hearing no objection, we'll go through billing and then we'll take a break after that.
Agenda Item: CHI Final Reports - Billing - Ms. Wark
MS. WARK: The billing team composition, as you can see we had very wide representation across the federal government with several representatives from CMS, we had two representatives from the Social Security Administration, we had four representatives from the Veterans Administration, and two from the Department of Defense.
The description of the billing domain is that these standards are used to implement electronic exchange of health related information needed to perform administrative functions in the federal health care enterprise. This was a validation effort assuming that the HIPAA transaction and code sets would serve as the basis for the standards.
The scope of the billing domain was defined as those areas already addressed under HIPAA and in addition the workgroup felt that it would be beneficial to look at appeals and certificate of medical necessity, and other areas that were out of scope were claims attachments, report of injury, non-claim payment electronic funds transfer, purchasing, such as medical supplies purchases, provider identifiers, unique patient identifiers, advanced beneficiary notification, and electronic signatures, all of those items that I just mentioned were out of scope. Employer identifiers was defined to be in scope because that is already covered under HIPAA.
The alternatives for billing were all of those already approved under HIPAA and in addition we added ICD-10-CM and ICD-10-PCS.
The recommendation is that the HIPAA approved transaction and code sets, both those currently approved as well as future updates are recommended for adoption, so essentially the aim here was to keep the federal enterprise in sync with HIPAA and that we not need to revisit any future updates, that we simply say that those are adopted as they are enacted.
In the area of claims attachments, those were considered out of scope due to the scheduled publication of the attachment, NPRM by HHS in 2004, work is underway between the HL7 attachments special interest group and the CHI staff to map and align the clinical standards with the proposed HL7 claims attachment standard, therefore until this work is evolved further the workgroup considers this out of scope and suggests the area be revisited in 12 months.
In the two additional areas that the workgroup looked at, appeals and certificate of medical necessity, there were no additional standards that are being recommended now for adoption, however it's worth noting that the X-12 837 transaction could be used for certificates of medical necessity, however, it is not a HIPAA approved transaction code set and it would need some work in order to be used by the federal enterprise. Currently there are no federal agencies using an electronic standard for data or structure related to certificates of medical necessity, therefore no standard for this function is being recommended.
And that concludes the billing workgroup recommendations.
DR. COHN: Questions or comments from the subcommittee? Kepa?
DR. ZUBELDIA: I concur with the recommendations but I would like to see in our letter to the Secretary to expand the domain and so the domain scope to include all of those parts of HIPAA that are currently not in scope, and you've already recommended to review the claims attachment in 12 months but potentially under HIPAA there's a requirement to adopt the transaction for report of injury and there will be provider identifiers, patient identifiers, and payer ID's that will also at some point, except perhaps the patient ID, become HIPAA standards. And at that point I think the recommendation will be that CHI should consider those also as CHI standards and that be revised as HIPAA adopts new standards.
MS. WARK: I believe that is consistent with our recommendation that for example the, initially we did not have employer identifiers as in scope because it's not been implemented yet, however we changed that because we do have the compliance date noted. It is the intent of the workgroup and the CHI Council that any future updates to HIPAA would be reflected in the CHI standards therefore we would simply need to update our standards list and our documentation to reflect that.
DR. ZUBELDIA: So my recommendation is to recommend to the Secretary that those future HIPAA standards be brought into scope for this initiative.
DR. COHN: Well, I think I'm reading this as the HIPAA approved transaction and code sets, both those currently approved as well as future updates are recommended for adoption. Is that sort of what you're saying?
DR. ZUBELDIA: Yes.
DR. COHN: So that's already in here. I guess the one question I would have, which I think is sort of along the lines of Kepa, that I think identifiers are sort of implicitly included but are not mentioned and I tend to think of them as something different then transactions or code sets, so I would expect that you would probably, I mean I think that's what you're doing though it's sort of hard for me to tell. I think that you're talking about including recommendations for HIPAA approved transactions, identifiers and code sets.
MS. WARK: Yes, that's correct.
DR. ZUBELDIA: Then I would add to that list, the plan ID is not on the list.
DR. COHN: That's an identifier.
DR. ZUBELDIA: Yeah.
DR. COHN: So I'm just sort of saying that maybe this first sentence ought to be modified where we would suggest the modification, it says HIPAA approved transactions, identifiers, and code sets as the first sentence, we should just add identifiers. Kepa, are you okay with that?
DR. ZUBELDIA: Yeah.
DR. COHN: I think that handles that, Steve, are you, I think that's really editorial because I think you intended that.
Now one question, actually I have two sort of questions here and these are just more clarification. Obviously you single out certificate of medical necessity recognizing that we will in some point 2004 probably in January be hearing from DSMOs about updates and changes to transactions. Did this turn out to be the high priority item that you did not have currently for, in terms of HIPAA requirements? Is this in your view a sort of a very high next level item or is it just happens, just sort of mentioned?
MS. WARK: I think the workgroup felt an obligation to at least address areas that weren't covered under HIPAA and to see if there was an opportunity to standardize anything beyond the HIPAA approved standards, to take advantage of the fact that we had the workgroup together and we were doing the assessment. There was quite a bit of discussion around it, I wouldn't say that anyone, any organization felt that we needed to put this on the agenda, it was merely capturing what's being done in that area and essentially all reporting federal agencies said there really isn't anything in use. We did make an attempt to look at X-12 to see if there were something there that could be considered, there is something, it would need some modification and it was left at that. I wouldn't say it was on anyone's agenda.
DR. COHN: Okay, so I think we'll leave it then without change.
Now a final question and this is actually, I'm almost afraid to go here but just in terms of your, of how the, your comment about the HIPAA medical code sets and the recommendations, it says ABC codes for registered users, which I presume you mean ABC for registered users during the pilot testing period.
MS. WARK: Yes.
DR. COHN: Okay.
MS. WARK: And the full report is more specific about that, yes.
DR. COHN: Is there another full report that's, okay.
MS. WARK: When the standards are adopted under CHI we take all of the information that we gather and we post a report on the web, and so I was just merely commenting that yes, we had had a significant amount of discussion about what that meant, the ABC codes and it is under the pilot period for those voluntarily registered users.
DR. COHN: You might just want to wordsmith that bullet.
DR. STEINDEL: Simon, is that a recommendation that it be a modification in the --
DR. COHN: Yeah, I actually was going to suggest that maybe it's just, once again, I think it's editorial but it's just a, so we're concurring with --
DR. MCDONALD: Simon, I can't hear what you were just saying.
DR. COHN: Okay, Clem, I'm sorry, I think I'm speaking right into the microphone and I do apologize. What we're doing is recommending this domain with the two following modifications, one was that in the recommendation in the first sentence it would be changed to HIPAA approved transactions, identifiers, and code sets, so that's in the first sentence, which we think is just an editorial modification. The other is in the, there's a table where it says HIPAA medical codes sets and there's a bullet that says ABC codes for registered users, and we were basically just further clarifying that by saying ABC codes for registered users during the pilot period, is that --
MS. WARK: That's correct.
DR. COHN: That is correct, just to further clarify the meaning --
DR. MCDONALD: Sounds fine.
MS. GREENBERG: I think this should say and code sets, this recommendation, plural --
DR. COHN: That first sentence, code sets rather then codes set --
MS. GREENBERG: In one place it says codes sets, another place it said code sets, and this third place it says codes set, I think the terminology is code sets.
DR. COHN: Thank you, okay. I figure that's typographical. Okay, any other changes to this one?
DR. ZUBELDIA: I would like to add to this domain/sub-domain table that lists the domains that are in and out of scope the health plan identifier, I think that it's probably an oversight because it's one of the HIPAA identifiers and is not on the table in the front. Health plan identifier.
MS. WARK: I believe that is an oversight. When we changed the line to indicate that employer identifiers was in scope we originally had plan/ and I think it just got dropped off. Thank you.
DR. COHN: So are we, anything else or are we ready to vote on this? Jeff, do you want to make a motion?
MR. BLAIR: I'd like to make a motion that we concur with the CHI recommendations with the noted editorial corrections.
MS. GREENBERG: Can I just ask, in light of this language then where it says no for these identifiers it should say yes, is that a conclusion? They are in scope and to the extent that they're --
MS. TRUDEL: They're not in scope for phase one because the standards have not been adopted, and in phase one what the group did was to go back and do a validation process for the standards that had been adopted to date. So they would be in scope in future.
DR. COHN: Any further comment or discussion? Okay, in favor?
SUBCOMMITTEE: Aye.
DR. COHN: And Stan your hand up was yes?
DR. HUFF: Yes.
DR. COHN: Any opposed? Any abstentions? Okay, so that domain is passed. I think now it is time for a break, let's take a 15 minute break and we'll come back, I think we're one domain behind but we will make up time.
[Brief break.]
DR. COHN: Okay, our next session with a discussion on the supply domains, is that correct Brock? We're going to supply now.
Agenda Item: CHI Final Reports - Supplies - Mr. Hefflin
MR. HEFFLIN: So here we see the composition of the medical supply team, I'd like to thank --
DR. COHN: Brock, you need to get closer to the microphone.
MR. HEFFLIN: Here we see the composition of the medical supply team, I'd like to thank its members for their participation.
The subset of the medical supplies domain is any instrument, apparatus, appliance, material, or other article intended to be used by human beings to diagnose, prevent, monitor, treat, or alleviate the disease, injury, or handicap, or to replace and/or modify the anatomy or a physiological process.
The supplies domains includes all medical devices and supplies, and currently that's nearly 7,000 generic groups. The names for these products are used primarily to inventory and document their utilization by health services establishments, and to regulate their availability and utilization in the community by public health agencies, specifically the FDA.
The group believed that the terminology for medical supplies should be high comprehensive and sufficiently specific that as to the generic device for supplied group level to accommodate regulatory and inventory activities. We also believe that the terminology should provide definitions for the terms.
Here we see the alternatives considered, HCPCS, otherwise known as HICPICS, ICD-9, UNSPC or UNSPEC, and SNOMED CT were eliminated as candidates because they were either not sufficiently comprehensive or robust or specific or they lacked definitions. The FDA Medical Device Classification is comprehensive and has definitions, however it's structure is flawed, there are problems with its structure. So that left the Global Medical Device Nomenclature and the Universal Medical Device Nomenclature System.
Now the GMDN and the UMDNS are very similar in scope, that is each supplies names, definitions and codes for essentially all medical devices and supplies. The GMDN was recently created through the harmonization of six established medical device terminologies, it is including a previous version of the UMDNS and the terminology used by the FDA. It is an ISO and a CEN standard, so therefore it's being adopted by many countries.
The UMDNS is owned and operated by ECRI, a U.S. based health services research agency, it's been used internationally for a few decades. The UMDNS is available through the NLM UMLS and is supported by an established business plan.
The GMDNs business plan is still in formation, however as I mentioned previously it is an international standard and is strongly supported by the FDA for global communication and as an eventual replacement for its terminology.
Efforts are now underway to start to merge the terminologies, ECRI is being invited to participate in the GMDN maintenance agency, and in addition the FDA and ECRI are involved in a CRADA, Cooperative Research and Development Agreement, and the purpose of that is to, one of the purposes of that is to map the GMDN and the UMDNS to each other.
So the recommendation is to wait for the GMDN and the UMDNS to merge and to adopt the resulting terminology. Hopefully this will occur within three years time and its progress should be monitored. The terminology resulting from a merge of the GMDN and the UMDNS will enable the U.S. federal system components to utilize one set of medical device and/or supply names, definitions, and codes, and to use these same product identifiers to communicate with foreign establishments.
In addition I was recently asked to address device type identifiers for medical devices, device type meaning make and/or model, and I have a few points here. There's been primarily discussions surrounding this issue to date, there hasn't been a whole lot of action, it's been primarily discussion. Some manufacturers have produced device type identifiers as part of bar codes, it's been at their discretion, it's been voluntary, and the purpose has been primarily for ease of tracking of their devices.
Historically device regulators have had problems with inconsistencies in device type identifies due to identifiers not provided by all manufacturers, different identifiers reported on the same product, identifiers reported not reflecting final identifiers used to market the product, a large number of manufacturers involved with significant turnover or companies going in and out of business or merging, a high diversity of medical devices and the lack of a unifying set of data elements, for example for drugs you have drug formulation and dose in the NDC, you don't have that for medical devices.
Internationally many in the device arena recognize the benefits of more specific device identifiers, those being accurate and consistent device reporting, cross referencing, tracking, and improved ease of reporting. A consideration is to place the device nomenclature category code and device type on the device label, barcode or other industry led identification scheme that is used to mark devices to limit manufacturer burden. However, it's felt that this practice will require regulating to become universal.
The GMDN/UMDNS merge does not address device type identifiers, again without a unifying set of device, data elements for devices it would be difficult to make this information part of a device nomenclature. However, we can propose that this issue be addressed as a new project by involved parties.
DR. COHN: Well, Brock, thank you. I mean obviously the first piece I think we had heard when you came and testified to us in September and I'm not sure that I have any great comments to make other then it sounds like you're increasingly optimistic that there will be a merging of the two, is that correct?
MR. HEFFLIN: Right, we're headed in that direction.
DR. COHN: Obviously the second piece that is actually not anywhere listed here is actually I think in some ways the most interesting of the comments that you've made. I guess I'm looking to see what the full committee, subcommittee thinks about how we should handle this because I'm certainly willing to accept your recommendation without further, or concur with your recommendation but I really, when I heard the earlier testimony on the various nomenclatures it didn't seem like it really solved the problem, and obviously your discussions about the device type ID really begins to get down to knowing what machine we're talking about or what device we're talking about and the issues related to that device, which has been something that's been sort of lacking in the work in this area. But of course as I said it's not really listed in any of these recommendations. What does the subcommittee think? I mean I'm sort of thinking that we sort of concur with the recommendation and that we strongly encourage further work in this area around device type ID and that that be really the recommendation.
MR. BLAIR: I would second it.
DR. COHN: How do others feel? Kepa?
DR. ZUBELDIA: I'm a little bit confused on this one. Is the recommendation a contingent recommendation or is it just no recommendation because there's nothing to adopt at this time and then you'll review the recommendation later?
MR. BLAIR: This is no recommendation I think is what he's saying at this time.
DR. ZUBELDIA: Well, that's not what it says.
MR. HEFFLIN: Well, the actual recommendation, I think what I've written there is that we wait for the merge and to adopt the resulting terminology.
DR. COHN: I guess what I'm hearing actually is probably technically no recommendation, to be reviewed in some number of months or whatever, though it doesn't say it, they're right, I mean we're just looking for some commonality between these and other recommendations are the same.
DR. ZUBELDIA: Under conditions it says the recommendation is contingent upon the success of the merger, so it looks like it is a contingent recommendation rather then no recommendation, that's why I'm a little confused about it.
MS. WARK: Can I add something there that might be helpful? I think the intent was that we put this on our phase two work plan as something to monitor and it is contingent on the success of the merger and the continued work in that area. At this time there is no standard ready to adopt but that we would continue to monitor the work toward what we expect will be the outcome as long as things continue on the plan that we envision for the merged terminology.
DR. COHN: Marjorie?
MS. GREENBERG: Did Jeff want to say something?
DR. COHN: Jeff, I'm sorry, did you have a comment?
MR. BLAIR: Well, that's okay, let me pass for now. Go ahead, Marjorie.
MS. GREENBERG: I just wondered if you had looked at ISO 9999 and what its relationship is to --
MR. HEFFLIN: GMDN, ISO 9999 technical devices for disabled persons, it was one of the candidate nomenclatures that comprises the GMDN.
MS. GREENBERG: So it's included in it, okay, because it has recently been adopted as a related member of the WHO family of international classifications.
MR. BLAIR: My comment is purely a procedural one and it echoes I think what Kepa was trying to say is your recommendation is that you wait for the merger, and that's fine, and actually you could really do this any way you want to but the explanation earlier about what conditional meant was usually it was conditional upon a recommendation going forward, so it just appears a little confusing because the condition sounds like you are making a recommendation where in fact you're winding up saying you're going to defer the recommendation for a period of time. Maybe it's just an observation rather then asking for a change.
PARTICIPANT: We can clarify that.
DR. STEINDEL: I actually have a wording question and generally speaking what we have said in the transmission letter is the NCVHS concurs with the CHI recommendation for, in this case are we concurring with the recommendation for a conditional recommendation or should I word this to the effect of the NCVHS concurs with the CHI conditional recommendation?
MS. TRUDEL: The recommendation is not to adopt a standard at this time. So you can concur with that recommendation.
DR. STEINDEL: I could either add the wording or just leave it as concur.
DR. ZUBELDIA: My recommendation would be to remove the condition.
MR. BLAIR: -- for us to wind up saying we concur with the recommendation. And I don't think it changes anything in terms of your intent or what you're trying to do.
MR. HEFFLIN: Right, so the recommendation you're recommending to be removed --
MR. BLAIR: No, not the recommendation being removed, no, it's the conditional statement.
MR. HEFFLIN: The conditional statement, you're getting that off the final recommendation information sheet, is that what you're looking at?
DR. COHN: Well, that's what we would --
DR. ZUBELDIA: The bottom of the back page, it says conditions.
MS. TRUDEL: We can move that statement out of the condition section and into one of the other sections.
DR. ZUBELDIA: Or into the recommendation section.
DR. COHN: Well, it would be hard to move into the recommendation section since it is opposite of what you're recommending.
MS. TRUDEL: We can fix this easily, this is easy.
DR. COHN: This is easy, this is wordsmithing, so we can concur with the final recommendation as stated --
DR. ZUBELDIA: We can concur with the recommendation to not adopt a terminology at this time.
DR. COHN: Why don't we say that? And I think we're further, assuming that everyone is in agreement, further recommend that the federal government further explore the device type ID as an essential next step. How are we on that, does that sort of meet everybody's --
Okay, comments? Are we ready to vote on this one? Steve, do you feel comfortable with the wording?
DR. STEINDEL: [Inaudible.]
DR. COHN: I was afraid you were going to ask me about that one. I guess we recommend that HHS further investigate device type ID as an essential, I'm not sure the term is essential, maybe it's critical, item --
MR. BLAIR: Do you want to say HHS or CHI?
DR. COHN: I think it's HHS actually.
DR. STEINDEL: If I may, we recommend HHS further investigate device type ID as a component of device nomenclature.
DR. COHN: Is that okay? Brock, are you --
MR. HEFFLIN: Or some method of device ID, as I mentioned earlier it might be difficult to integrate that into the nomenclature but that it be utilized in some fashion. We can explore the nomenclature option but there are other options as well, for example, device type ID that's part of a barcode, separate from the nomenclature.
DR. COHN: So how should we wordsmith this one? Read what you have again Steve.
DR. STEINDEL: We recommend HHS further investigate device type ID as a component of device nomenclature. Would it be okay to just change device nomenclature to a device identification?
MR. HEFFLIN: That'd be fine.
DR. COHN: Great. Okay, so are we comfortable with this?
MR. BLAIR: I'm assuming that preceding that we're saying that NCVHS concurs with the CHI recommendation to not recommend a supply standard at this time, or you want to say supply or device standard.
DR. COHN: Medical device.
MR. BLAIR: Medical device.
DR. COHN: And supply.
DR. STEINDEL: Jeff, the way that's worded right now is the NCVHS concurs with the CHI recommendation to not adopt a terminology for the supplies domain at this time, just simply because supplies is what they have called it.
MR. BLAIR: That's fine.
MR. HEFFLIN: Right, and we tend to, in my neck of the woods we tend to refer to devices and supplies just as devices, and I know here we're using the word supplies but just so we understand each other.
DR. COHN: Let's call this medical devices and supplies.
DR. STEINDEL: Is that okay with CHI?
PARTICIPANT: Yes.
DR. ZUBELDIA: I have a question, perhaps out of scope but maybe more for my own information. How does this interface with the adoption of HCPCS under HIPAA for supplies? Or does this have anything to do with it?
DR. COHN: I think under billing we just accepted HCPCS.
DR. ZUBELDIA: That's right, we did, but HCPCS is outside of the scope of this one and they both talk about supplies, is there going to be confusion somewhere?
MR. BLAIR: That's a very good point.
MS. TRUDEL: This is something that we're going to have to tackle in a number of places and it has to do with the fact that the standards that are appropriate for billing and administrative transactions under HIPAA are not necessarily appropriate for clinical settings, especially with regard to nomenclature and we need to begin to talk about, well, we have begun to talk about how mapping would help us step from one to the other.
DR. COHN: Okay, further comments, questions? Steve, what do you have?
DR. STEINDEL: Just as an aside, during the PMRI terminology sessions we did have some discussion on the difference between the identification of devices for billing purposes and devices for other purposes, so we've already recognized the difference between the two and the need for mapping between the two.
DR. COHN: Great. Any further comments? Can we vote? All in favor?
SUBCOMMITTEE: Aye.
DR. COHN: Opposed? Abstentions? Okay, it's passed.
Okay, now are we back to nursing or are we, okay, Alicia, please.
Agenda Item: CHI Final Reports - Nursing - Ms. Bradford
MS. BRADFORD: I'm Alicia Bradford, team leader for the nursing workgroup, and we had representation from CMS, DOD, VA, and the Indian Health Service.
Our domain, the scope of our domain is terminology that is used to identify, classify, and name the delivery of nursing care. Sub-domains were derived from the Nursing Process and the American Nurses Association approved Nursing Minimum Data Set, emphasizing nursing assessment, diagnosis, interventions, and outcomes of nursing care.
So I just listed our sub-domains, assessment or observations, plan or goals, diagnosis, intervention, evaluation and outcomes. Intensity of nursing care and patient demographics were determined to be out of scope. Intensity of nursing care, although part of the Nursing Minimum Data Set, has --
DR. COHN: Alicia you need to get closer to the microphone.
MS. BRADFORD: The intensity of nursing care, part of the Nursing Minimum Data Set, out of scope, as no rating or vocabulary standard exists, or is currently widely implemented. Patient demographic data is out of scope because it has been covered by the demographics workgroup.
So went to the AMA to determine what nursing, what vocabulary support nursing practiced, there's 14 that they've approved, three are primarily administrative and were excluded at that point, leaving these alternatives, SNOMED CT, ABC Codes, NANDA, NIC, NOC, Omaha, the Patient Care Data Set, the Home Health Care Classification, Perioperative Nursing Data Set, International Classification for Nursing Practice and Clinical LOINC.
MR. BLAIR: I'm assuming that when you said SNOMED CT you were including the nursing terminologies that are part of SNOMED CT.
MS. BRADFORD: Yes. The final recommendation is for SNOMED CT, which contains nearly all of the source nursing terminologies that were listed there. It contains over 1,000 nursing intervention concepts, modeled from the Georgetown Home Health Care Classification, Omaha System, Nursing Interventions Classification. It has intervention concepts from the Perioperative Nursing Data Set, Nursing Diagnosis and Problem Concepts from NANDA, Perioperative Nursing Data Set, Home Health Care Classification and Omaha.
The convergent terminology working group with SNOMED continually works with the source nursing terminology owners to update the nursing content and model that according to ISO standards. NOC is going to, the Nursing Observations Classification will be integrated in the January 2004 release and additional from the Home Health Care Classification and the Omaha Systems will be in the July 2004 release.
So as attached to the back of my report I did a little table to kind of illustrate what nursing terminologies are for one, free to the public, the intent of them, and the ones that are mapped or the concepts are integrated into SNOMED CT. And all of which are other then the International Classification for Nursing Practice and the Patient Care Data Set, which is currently only used at Vanderbilt University. I spoke with the terminology owner and developer and she says that she has plans to have it coded according to Clinical LOINC and mapped soon to SNOMED CT. The discussions continue with the National Classification for Nursing Practice.
Obviously the workgroup could not examine the entirety of SNOMED CT so we examined it as it pertained to nursing concepts listed in our sub-domains. Essentially the nursing concepts fall in the SNOMED CT hierarchies of findings and procedures and I've illustrated an example of that hierarchy in my report.
So there's no conditions with this recommendation. The workgroup would like to see mappings between the source nursing terminology owners and SNOMED CT and for those mappings to be maintained, validated and distributed through the UMLS. The workgroup recognizes the importance of a collaboration of the source nursing terminology owners and the SNOMED CT convergent terminology working group for the appropriate inclusion and representation of nursing terms within SNOMED CT.
That ends the report.
DR. COHN: Okay, comments or questions. Jeff were you going to ask a question?
MR. BLAIR: Forgive me, this is the blind guy and I did get a chance to look at some of this but I just don't remember. I think you mentioned alternative billing, not, the ABC Codes, and I just don't remember how you indicated they would or would not be reflected. Could you remind me?
MS. BRADFORD: Actually the alternative billing codes, the nursing content of them are built upon NIC, the Nursing Interventions Classification and NOC, and also NANDA, and I took several of the terms and mapped them directly into the SNOMED CT using the clue browser. I discussed with the owners of the ABC Codes whether discussions are underway with SNOMED CT regarding mapping or integration and that currently according to them is not underway. But primarily the ABC Codes represent more of a billing and administrative then the documentation of nursing care.
DR. COHN: Comments? Michael?
DR. FITZMAURICE: So as I understand the recommendation it's SNOMED CT because it contains all of the nursing terminologies that you just went through, and you would adopt the January '04 release but you then would adopt the July '04 release when it comes out.
MS. BRADFORD: Well, part of the licensing agreement, I'm not sure if that includes the continued upgrades of SNOMED, so if it does, yes. We adopt SNOMED CT because it brings together all of the source nursing terminologies that have not converged to date but their inclusion in SNOMED CT is what makes it appropriate for the nursing work. So as they're updated we would continue to adopt that.
DR. FITZMAURICE: Okay, good. I just wondered if you wanted to put that in your final recommendation, which version that you're really adopting. Thank you.
DR. COHN: Other questions or comments? Does anybody want to move concurrence of this recommendation?
DR. HUFF: I would move concurrent.
DR. COHN: Oh, Stan is. Second?
MR. BLAIR: I'd be happy to second it.
DR. COHN: Okay, further discussion, comments? All in favor?
SUBCOMMITTEE: Aye.
DR. COHN: Opposed? Abstentions? Okay, passed.
MS. BRADFORD: We're scheduled for lunch but I have like a five minute history and physical one, if you want to get back on track with that I think.
Agenda Item: CHI Final Reports - History & Physical - Ms. Bradford
MS. BRADFORD: The team leads for this workgroup are not present, it's Dr. Viet Nguyen and Linda Nugent with the VA.
This workgroup came to you guys I believe in October and it was suggested just to throw the domain out but instead of that we have determined that there's considerable variability in the format and content of the history and physical and it's presently not standardized and is typically dependent on the clinical judgment of the practitioner, so the recommendations to defer work on standardizing history and physical format to align efforts that are going on with the standards development organizations and also with the other domains within CHI.
DR. COHN: And I guess you're also trying to wind up with the text based --
MS. BRADFORD: This is the same workgroup as the text based reports --
DR. COHN: Oh, so this includes that.
MS. BRADFORD: No, the same people are working on both workgroups, it became evident when they struck out that they needed to settle text based reports and then in the next phase look at history and physical and align with what's going on currently with HL7, delaying their --
DR. COHN: And by struck out you mean went out to evaluate the area.
MS. BRADFORD: Yes.
DR. COHN: I just want to make sure we didn't have both history and physical striking out part of the expression and text based reports striking out, thank you.
Kepa, did you have a question.
DR. ZUBELDIA: I move that we concur with this recommendation.
DR. COHN: Okay, second?
MR. BLAIR: I'll second.
DR. COHN: Jeff second. Okay, any comments or questions? Okay, all in favor?
SUBCOMMITTEE: Aye.
DR. COHN: Opposed? Abstentions? Okay, I think we're concurring with that.
DR. STEINDEL: Should I word that as we concur with the recommendation or we concur with the recommendation to defer?
DR. COHN: Okay, that sounds fine. Okay, well I think we're roughly back on schedule at this point, I think it's like 11:45 as I check. I'm going to suggest we maybe take a 45 minute lunch break rather then an hour, and that puts us complete back on schedule, so we'll start back at 12:30. Okay, thank you.
[Whereupon at 11:45 a.m. the meeting was recessed, to reconvene at 12:30 p.m., the same afternoon, December 9, 2003.]
A F T E R N O O N S E S S I O N [12:30 p.m.]
DR. COHN: Okay, well why don't we get started with our afternoon sessions. Now first of all I think we want to check to see if there's anybody who is on our call in line. Sir, would you please introduce yourself?
DR. ANDERSON: This is Bill Anderson and Dr. Lawrence Desi(?) as SSA.
DR. COHN: Well, welcome. Anybody else on the call in line? Is Dr. McDonald on the line? No, okay.
DR. HUFF: I had contact from him at lunch, I know he was planning to join back. I guess he was, had to take care of something, he came back and nobody was on the line so he wondered where we were. Anyway, so I expect him to join us.
MS. GREENBERG: Does he know we're reconvening?
DR. COHN: Why don't we start with the discussion on disability? Jennie? Welcome.
Agenda Item: CHI Final Reports - Disability - Ms. Harvell
MS. HARVELL: Hi, I'm Jennie Harvell and Stan Shipley and I were co-chairs for the disability domain workgroup that was convened as part of the CHI.
The disability domain workgroup consisted of representatives from the National Library of Medicine, Vivian Auld was extremely helpful in our efforts and you'll see as we go through our report Vivian's contribution to our effort. We also had representatives from the SSA, the CDC, and CHS, the VA, CMS, and DOD monitored our activities and we had other ASPE staff participating as well.
DR. COHN: And Jennie you need to make sure to get really close to the microphone because I'm sure that people from the Social Security Administration won't be able to hear you otherwise.
MS. HARVELL: When our workgroup first convened --
MR. BLAIR: Could we just ask for those folks that are listening in if you don't have comments to make and somebody is presenting if you put the mute switch on your telephone then we won't hear your rustling papers and coughing and making other little comments.
Participant: What's the mute switch?
MR. BLAIR: Don't you have on your phone, most phones have a little button that you could press that you could mute --
DR. MCDONALD: It's a key I think on most of these services, but I don't know what key it is.
MR. BLAIR: That's Clem, we didn't want you to mute --
DR. MCDONALD: Actually I'm on the phone, I'm not on a, never mind.
DR. COHN: Anyway, Jennie, please.
MS. HARVELL: When our workgroup first started meeting the first thing the considered was how to proceed with our work and the first thing we thought we needed to understand were how does the federal government use disability information in its current health care practices. And we identified several use cases, CMS uses disability data in its patient assessments and its various payment methods, as its internal quality assurance process, in its public reporting process for quality indicators, it uses disability data in its home health programs, in its skilled nursing facility and nursing facility programs, and in its inpatient rehabilitation facility programs.
CDC, the NCHS, uses disability data in several surveys that it sponsors and conducts and it uses the information it gathers from these surveys to guide policy development.
SSA uses disability data in eligibility determinations for both children and adults.
The VA uses disability information for internal quality monitoring purposes, for its service delivery, and it also uses disability information in eligibility determinations.
DOD uses disability information in various administrative activities, and it's used in the course of service delivery within the department.
The workgroup decided that it would sample some of the disability data that is used across the federal health care agencies that were participating on the workgroup. We sampled terms and concepts from the Nursing Home Minimum Data Set, the Home Health Oasis Assessment Instrument, the Functional Independence Measures used by the VA in its service delivery program. The FIM is also a core component of the CMS inpatient rehabilitation facility patient assessment instrument. We also sampled terms from the residual functional capacity form used by SSA in its eligibility determinations and we also sampled concepts from various surveys sponsored by the NCHS.
The workgroup initially decided that it would consider the coverage provided by SNOMED CT, the ICF, and other sources in the UMLS meta-thesaurus in examining the disability data needs that the workgroup had sampled.
As I mentioned Ms. Auld was extremely helpful and critical to our work. She used the meta-map transfer program in helping us identify sample terms that, coverage of sample terms that we provided to her and the coverage in the UMLS meta-thesaurus. The meta-map was able to parse the terms into noun phrases that she then mapped back to the meta-thesaurus and she used those noun phrases to run against the SNOMED CT clue browser and the ICF browser.
The workgroup applied its expertise with respect to federal disability data needs to Ms. Auld's findings of SNOMED CT and ICF coverage. We used the ICF manual as needed to review and modify Ms. Auld's determination of coverage provided by the ICF. We did not use a similar method with respect to SNOMED so therefore we did not use a SNOMED manual to review and modify if needed Ms. Auld's SNOMED coverage determinations. We also reviewed and modified as needed Ms. Auld's determination of full, partial or no coverage being found in SNOMED CT and the ICD for our sample terms.
The workgroup constructed several spreadsheets throughout our efforts and the spreadsheets reflected one term or concept per row, and across the columns we reflected for SNOMED and ICF and the UMLS in our initial analyses, whether or not there was full, partial, or no match found in either the SNOMED CT, ICF, or other sources in the UMLS meta-thesaurus.
Our initial findings based upon the work Vivian initially produced for us was that the UMLS meta-thesaurus could identify numerous vocabularies, and given that conclusion the workgroup thought that given the purpose of the CHI which was to identify largely a single vocabulary or a couple of single vocabularies if it were possible to slice the disability domain into bitable chunks, that further review of the UMLS meta-thesaurus as possible sources for disability coverage would not be productive at this time. We thought the UMLS meta-thesaurus could be useful in the future once a single vocabulary or terminology is identified to fill gaps that are found in that identified single source vocabulary or terminology. Thus the workgroup limited its further analyses to SNOMED CT or ICF.
We selected SNOMED CT because it is widely recognized as being the most robust clinical terminology and of course the recent license between the National Library of Medicine and CAP(?) to make SNOMED freely available in the United States. We selected the ICF because it is the only internationally recognized disability classification and there's been considerable development work towards that classification over many years.
In additional work, and let me just back up, in our first round of analyses the first spreadsheets that the workgroup produced, we sampled over 900 terms and concepts and Vivian ran many of those against the UMLS meta-thesaurus to identify these noun phrases. In our subsequent analyses given the fact that we were not going to continue focusing on the UMLS we also reduced our sampled terms from the MDS, the NCHS survey items, and the Oasis items. Our reduced sampled terms exceeded 300, I think there were about 312 final sampled items that were subsequently included in the additional analyses. And as I said we limited our analyses to only the coverage provided at that point for those 300 plus items to SNOMED CT, coverage provided by SNOMED CT and ICF.
So in our report the final analyses, and we have appendices in our reports that will eventually be part of a public document, our final spreadsheets include the 300 plus terms, a summary of coverage provided by ICF and by SNOMED, a description of whether that coverage provided by either SNOMED or ICF is full, partial, or complete coverage, and it tallies those coverage levels.
The general findings by the workgroup was that SNOMED CT and ICF do not include many of the disability data needs that the federal government uses in its current programs. Further, when coverage is available often it's only partial coverage. I should note that levels of partial coverage, there are shades of gray in partiality. In addition the workgroup observed that ICF bundles concepts, it is a classification system, and that the federal government often needs these concepts unbundled and needs discreet terms.
In addition neither SNOMED CT nor ICF adequately covers the disability scaling needs and the ability to crosswalk across --
MR. BLAIR: Excuse me, Jennie? That is a phrase that I don't understand and maybe there's others that don't either. What are the scale, disability scaling? Could you explain that?
MS. HARVELL: For example, on the Nursing Home Minimum Data Set, for every item a person is to complete an assessment of a nursing home resident's ability to perform a certain task. So for example, in the ability to ambulate, there is a skill included on the Nursing Home Minimum Data Set form that I believe includes five or six items and it rates, goes from either independent to needs a level of supervision, needs physical assistance, minimal physical assistance, needs moderate physical assistance, requires total assistance, or not applicable. That's one example of a scale on one of the federal use cases.
MR. BLAIR: Now in your recommendation you refer to the scaling requirements of SSA and is that different, is that the thing we have to relate scaling to?
MS. HARVELL: I'm trying to remember in our reports the scaling requirements, I think what you might be referring to with respect to SSA is that on our first round of analyses that we got back from Vivian she looked at the SSA sampled items and as we were going through and looking at Vivian's results with respect to the SSA sampled items we quickly realized that one of the places where we were either getting no coverage or often partial coverage was with respect to SSA scaling items. We then took into account scaling issues in other federal use cases, for example the Nursing Home Minimum Data Set, the Home Health Patient Assessment Instrument, the Functional Independence Measure used by the VA, etc., and we saw the types of concepts, the types of words that were showing up with respect to these scales are similar, but applied differently, across these federal use cases. And so we saw that these similar terms are not showing up or not showing up completely in either ICF or SNOMED.
DR. COHN: Jennie, since we're talking about scaling, I guess what I'm thinking of this as is basically permitted values in a data element. What we're talking about is basically permitted values for a data element is really what you were describing and I guess there's the issue, which I'm sure you're probably going to talk about, is one is, I mean we could talk about this one plus to five plus or whatever way you want to scale, but there's a question of exactly that terminology, exactly those words, and then there's a second piece of are there other substitute concepts that really mean the same thing but aren't said the same way. And I presume that was part of the evaluation also?
MS. HARVELL: In part. What we typically found, for example on SNOMED, SNOMED allowed observations, findings, so it was possible in SNOMED to capture a person was observed to talk independently one time, two times, three times, whatever, an algorithm would need to have been constructed to equate those SNOMED findings, the observations that could have been recorded using SNOMED, to the scaling use case in any particular instance. So if, and I just want to be clear, the workgroup was not at all recommending consistency across scaling, across the difference federal use cases.
What we were trying to do, trying to find if it's possible to have a robust enough terminology that would allow the various scales that are used to be captured is one thing, and then secondly assuming that that could happen, then to crosswalk across those different scales within an agency, for example CMS in its Home Health Nursing Home and Rehab Patient Assessment instruments uses different scales for similar functional activities or similar functional domains. Right now it is not possible to crosswalk or equate those different scales. So if there were an agreed upon terminology that was robust enough to capture observations it could be possible to equate or crosswalk across these different scales.
DR. COHN: If you could equate the questions.
MS. HARVELL: Correct.
DR. COHN: I'm sorry to break into your presentation.
DR. STEINDEL: I'd just like to comment that this question of scaling is not unique to the disability domain, it's universal in any assessment instrument that's using any type of scale or questions that determine assessment or observations that determine assessment, not just in disability but virtually any clinical area. I know one example where SNOMED actually did include the terms is for like the Glasco(?) Coma(?) scores, and they put explicit that but that's another type of scale. So it is a problem in terminology in general.
MS. HARVELL: Right. On the other hand, and this goes back to this concept of varying levels of partiality, so on SNOMED, SNOMED you could capture observations and we considered that only to provide partial coverage because somebody would still need to take that observation and create an algorithm, create a statement, that would link it back to whatever federal use case we were looking at.
On the other hand ICF was also in terms of scaling considered to provide only partial coverage because ICF had its own scaling concepts and I don't have the ICF right in front of me but they would say no impairment, mild impairment, moderate impairment, etc., providing their own definitions of what that meant so they had their own concepts of disability functioning that sometimes equated with some federal use cases and oftentimes did not equate with federal use cases so ICF was also rated as providing some partial coverage but for different reasons then SNOMED was rated as providing partial coverage.
DR. MCDONALD: Could I just make a comment? Can you hear me? It's Clem. I think there's getting to be confusion between questions and answers in the sense that you have to construct a bunch of answers to make a question. The attachment structure deals with many of these things using a question and an answer, the field and the value in the field, but you need two different things to deal with these survey instruments, that's what these are.
DR. COHN: Clem, I don't know that you created great understanding, it looks like people are looking at you quizzically. I'm presuming you're reflecting back on LOINC, which is that I guess the --
DR. MCDONALD: Whether we call it LOINC or BOINC, it's something different, I just conceptually we can't take a vocabulary and make a survey question out of it and make it right, you got to have two different things, you got to have one thing as a survey question and often it actually has to have a sentence that's exactly the way it's asked to get it to be a standardized survey and this is true of some of the nursing issues, it's true of some of the attachment issues, it's true the Hamilton Pain score, it's true of the Glasco Coma score. You have a series of questions and you have sort of standardized answers.
The answers might be actually codes that look like SNOMED codes or ICD-9 codes or any other vocabulary, discreet atomic elements, but what the confusion I think in the discussion is yes, you could construct the syntax and all this to make something that looked like it but in real life these things are all two different dimensions, you've got a field and a value, you've got a question and an answer, you have something and usually when you have them being asked about patient symptoms or scores in a formal way it's done with a formal question, like within the last two weeks have you ever, etc., etc., and how much, there's like hurts scores you get in there sometimes, sometimes it's just yes/no, sometimes it's exactly what the symptom is.
But I just want to clarify that I think the thinking should be developed a little more about this, that's all.
MS. HARVELL: Right, and just making sure I understand your comment, one of the things that the workgroup, particularly with Vivian's help, was able to do, as I said we sampled terms and sometimes we sampled concepts, and when we were looking at a sampled term it was a discreet unambiguous word, and Vivian's task was to go back to SNOMED, ICF, UMLS meta-thesaurus and see if she could find that exact word someplace. Alternatively, if it was a concept that included multiple words, Vivian would look to see if she could either parse it out into discreet words or somehow a concept was otherwise captures and you would see this on our spreadsheet attachments, she would include all the discreet covered terms that composed that concept and then the workgroup looked at the listing of discreet covered terms that Vivian told us were, for example included in SNOMED or ICF, and looked at whether or not those words when strong together equated the concept that we had sampled.
Just the last point on the slide that's up there before we change was that the workgroup also realized that given the lack of scaling, the lack of terms included in vocabulary or the ICF classification, that not only would that content need to be added to make a terminology or a vocabulary more robust, but once it is there then work would be needed to create the, to translate the scaling needs of the federal government within an agency as I said and across agencies.
DR. COHN: Actually can I ask a question back about that? Can you go back for just one slide? I guess I was just curious about that, we're obviously asking questions about your findings. You consider that the algorithms that you're talking about are those within scope of a domain or a CHI domain discussion? Or is that something different? Normally when I hear about things like this, and I'm thinking for example of the patient safety work going on in the federal government where they're trying to commingle or standardize across various databases, that typically is not so much a terminology discussion as in like standardized concepts. And not standardized on the terminology, it's like let's first standardize concepts. Is that what you're seeing here? It sounds to me like you're saying let's keep the concepts the same but let's figure out some wonderful mapping algorithm that will map everybody's unique way of doing it.
MS. HARVELL: Right. I think that's exactly right and you're right, this was --
DR. COHN: There's no reengineering going on here at all that you're describing.
MS. HARVELL: Right, but it was just a reflection of well, once we get the terminology straight, work is still going to be needed to map across --
DR. COHN: Well, or some would say that you need to get that done before you can get the terminology straight, some people would make that comment also. I'm not saying that as a statement of fact but some might observe that.
MS. HARVELL: Right. We did not because our premise was that the federal use cases are the federal use cases and the Nursing Home Minimum Data Set, the Home Health Oasis instrument, the NCHS survey forms, the SSA eligibility forms, the FIM items used by the VA, are those items and they are using those scales in their programs. And we did not --
DR. MCDONALD: Could you explain code scales?
DR. COHN: Clem, a code scale is a question with a range of answers.
DR. MCDONALD: And I think goals are like that, too, and I think probably outcomes are. If the outcome is to get the patient's diabetes in control it's a number at some threshold, so this is mixing apples and elephants I think.
DR. COHN: Okay.
DR. MCDONALD: I mean I think if we're talking about things that have, especially if they have numeric values but even if they have categorical values in their survey instruments, it's a different critter and I don't want to make the LOINC versus something else distinction but it's just a different creature. And if you look at how people send those around in HL7 how it's modeled, any kind of, they're observations, they're not discreet concepts, and the observation can have numeric, I'm being redundant but --
DR. COHN: Clem, hold the thought and when we get into the recommendations --
DR. MCDONALD: I mean some of these things are discreet concepts like diagnoses and interventions, but some are these structures that have two parts to them.
DR. COHN: Well, as I'm saying hold that thought and when we get into the recommendations we can reflect back on some of this stuff. Okay?
DR. MCDONALD: Okay.
MS. HARVELL: So the next slide you'll see a table that reflects the percentage of coverage, complete, partial, or no coverage provided by SNOMED CT and the ICF for the final 312 sampled items. You'll see that SNOMED provided complete coverage for 56 percent of the sampled items, partial coverage 39 percent, compared to the ICF that provided 35 percent complete coverage and 52 percent partial coverage.
Generally those findings are consistent just in terms of the trends with previous research that was conducted by the Mayo Clinic that found that SNOMED CT provided a complete match on 46 percent of the MDS sampled terms for three selected clinical domains and the ICF provided a complete match for two percent of the MDS terms that the Mayo study looked at.
The workgroup recommendations were as follows. The ICF is a classification scheme that's used in conjunction with or complementary to the ICD, but it is not a terminology. SNOMED CT was found to have significant gaps and would need to be mapped to classification systems as well as federal patient assessment instruments, therefore the workgroup did not recommend SNOMED CT.
The workgroup recommended that work would be needed to enhance a core terminology to include the needed disability content that is used by and needed by the various federal health agencies and following up on your comment, work would be needed to produce algorithms to crosswalk and equate these scaling concepts within and across federal agencies and maybe this is beyond the CHI scope comment.
So that was the work that the workgroup had completed by the time it had written its report. Following the submission of our report to the CHI the workgroup had a follow-up conversation with CAP to address an issue that came up during our workgroup discussions and the question was to what extent will SNOMED hierarchies support the disability content. And that was a useful conversation to have, the folks from CAP helped us understand some of the hierarchies and how some of those hierarchies would support at least some of the needed disability content. During that discussion CAP indicated that Denmark, Netherlands, and Wales had asked CAP to incorporate the ICF into SNOMED and CAP expressed an interest in receiving advice on refining their hierarchies to better capture needed disability content and also expressed an interest in working with the World Health Organization in incorporating the ICF into SNOMED.
Subsequent to the conversation with CAP some of the workgroup expressed concerns that CAP had organized its work so that the nursing workgroup would be used to identify gaps with respect to disability terminology, and I think the concern was that these were nurses and not persons with expertise in disability.
The workgroup report was presented the HHS CAHIT and during that presentation the CAHIT requested NCVHS input on the pros and cons of the items listed up on the screen, which are that while SNOMED looks the most promising it does have significant gaps and therefore do not recommend SNOMED CT as the endorsed disability terminology at this time. However, that future efforts should be targeted at enhancing or improving the disability content of SNOMED CT and further that when NLM and CAP are negotiating the scope of work to refine SNOMED CT that the disability domain be included as a part of those discussions.
And before we get your input on that page --
DR. STEINDEL: I'd like to point out that the NLM contract with CAP does not indicate any grounds for NLM to demands or even negotiate future work in SNOMED.
DR. MCDONALD: I'm sorry, I couldn't hear that.
DR. STEINDEL: The NLM contract with CAP concerning SNOMED does not provide for any way for the NLM to do anything but request CAP to do anything with SNOMED, there are no terms and conditions for future work with regard to the development of SNOMED. That needs to be required by the government in the contract, so that the last condition of asking I don't believe we can do under the contract.
MS. HARVELL: It's been a while since I've looked at the license agreement and maybe I am recalling some earlier version of it, but there were processes in place, either through participation on the editorial review board, to highlight areas for enhancement of SNOMED.
DR. STEINDEL: It is just the standard process that occurs within the development of SNOMED.
DR. COHN: Well, Steve, But CAHIT can make whatever comments it wants to, so I think this is not a statement of reality necessarily, this is a statement of what their comments were. I mean this is not their comments, this is CAHIT's comments, and the next time you see any of them tell them, please tell them --
DR. STEINDEL: I know about it because Betsy and I have spent a lot of time talking about that.
MS. HARVELL: Okay, so just before we actually get your comments on that slide I just wanted to let you know what the workgroup's next, or what our next steps will be. As I said the workgroup report was written prior to the conversation that we had with CAP so we will be modifying the workgroup report with an addendum to reflect the discussion with CAP, to reflect the CAHIT comments, to reflect whatever input we receive from you all today. If necessary we'll revisit the report recommendations and we'll present a revised report at least to CHI and if necessary I guess to CAHIT.
So that's where we're at.
DR. COHN: And maybe to us again, you never know. Any other, is this? Okay. Well I have some comments, I'm sure that others may have some comments also but maybe I'll, Stan, do you want to start out and then maybe I'll chime in and maybe comment a little bit along the lines that Clem was commenting.
DR. HUFF: This is a comment not just to you but I guess a comment in many ways to a lot of the work that's gone on here. It seems to me sometimes we focus so much on the terminology we forget about the grander scheme of things. That is if I understand what we're trying to enable, we're trying to enable unambiguous sharing of data, and that sharing of data is more then terminology, it's the structure in which you put the terminology, and this is a kind of a constant refrain, and that I think in a way this is reiterating what Clem said, that if what you're trying to achieve is an unambiguous representation of data you need to consider more then just the terms, you need to consider how you're going to combine those terms into a message.
And knowing that HL7 has already been accepted part of what you could do as your activity is say if we were to send this data from one agency to another what does the data look like in our databases, but what would it look like in a message that was sent between there that was an HL7 message or some other kind of message, and that would help clarify the issues of the relationships between the questions and the answers and would get you closer to the goal of unambiguous data representation.
The second thing I think, again, I think following on what Simon was mentioning earlier, and having dealt with this in my own institution I guess in a lot of ways, if you don't do, if I understand again the approach it was to let people have total independence in the instruments they're using to evaluate what they evaluate and that somehow magic would happen that would allow agencies to share data. And I guess the kind of questions that come to mind and the kind of problems that we've had, the similar kind of problem that we've had for instance is when we collect data in our institution and then we want to participate in one of the national studies that aggregate data, so for instance we participate in the National Myocardial Infarction Database and try, we collect data that says when did the person have chest pain, when did they have their surgery, what was the complication from surgery, and the NERMI(?) thing asks things like did the patient have any contraindication to surgery within one week before the surgery. Did the patient have shortness of breath two hours before, I mean these very, and if you don't think about what --
I guess what I'm saying is that even though the questions look very similar if you don't worry about justifying the concepts and the content ahead of time then you can't answer the questions you're interested in based on what the data asks. Or alternatively, you're going to ask the patient or whoever is filling in this data to answer 20 questions that are very similar and they're going to get really tired of did you have depression in the last two weeks, did you have depression in the last three weeks, have you ever felt depressed, and you get so many variations somebody says this is not rational for us to participate in this. But I may be misunderstanding but that's kind of some feelings I was getting relative to this activity.
DR. COHN: And Stan maybe I will sort of jump on what you're saying a little bit. I guess as I'm looking here, and I'm not even going to address the CAHIT comments or questions for the moment, but it sort of struck me that there's sort of two very different activities. And I've looked at others to see if I'm mistaking this because some of it is based on what Jeff was reminding us about earlier today.
The work that the NCVHS has done has been trying to create a vision where in the course of regular clinical practice clinicians are addressing issues about a patient, are describing them, are making comments, are dealing with the world of health care. And then that data can be pulled back and used and conglomerated and structured and whatever in various ways. Now that's the world of clinical care.
Now then there's the world of questions and answers and questionnaires, and clearly the way that you've embarked on this has not been to look back and say well geez, a nurse looks at a patient or a physician looks at a patient, geez, they have a colostomy, they have some disability with that colostomy, what might they say about it, what concepts they would use. You've been really going directly into the questionnaires and have been trying to replicate them. And that's really sort of a different set of activities.
Now you can standardize that set of activities but I think as Stan was sort of commenting that probably there needs to be discussion at some point between all the various agencies about geez, do you want to one, do you want to ask about shortness of breath in one block or two blocks or three blocks or every five minutes or what do want to ask of them before everybody goes nuts in terms of asking questions, even before you come up with something more.
Now there are ways to standardize the representation and answers, and I think that's been work for example Clem has done around claims attachments with some aspects of LOINC as I understand, where you ask a question and you have a range of answers and you codify them. Now you don't get cross organizational reliability and, I mean everybody has to ask the same question for you to figure out what to do with the answer, so there are limitations with that approach. But if you're trying to analyze it from that direction it becomes very hard to figure out what terminology to use I think, or even what makes sense. Clem am I making bad comments here?
DR. MCDONALD: No, no, you're addressing, you're saying what everybody says. But I think that the key thing is if you're going to, terminology is useless by itself. That's maybe to take an extreme position. And the goal isn't to find one terminology, the goal is to find a way to convey. So if you were going to a history and physical, you're just using terminology, you're just talking. And if we do narrative text we might want to have a terminology for automatic, then you just think about a flat terminology that would just automatically sort of encode.
But if you're trying to quantify something, if it's done with a lab test and I'm not talking about the terminology, you want to define how hot the beaker is, what the re-agent, you have to do a lot of standardizing of the process to make it work. And when it gets to clinical questions we as clinicians we typically just do narrative and that's fine, that's good, I like it.
But if you're going to do something like how much shortness of breath do you have and try to use it widely, to quantify it you want to say some kind of context, after climbing stairs or when you do a treadmill, you really do quantify the hell out of it. So what we're aiming toward is those questions that are kind of discreet things that we use for purposes, the Hamilton Depression score, some of the nursing scores, you want to have it standardized so that it will have general meaning I think. You need to have, maybe it's a long question, a lot of words to it.
DR. COHN: Clem, are you, I guess I want to find out whether you're agreeing with me or disagreeing with me. Do you think that the issue of having questionnaires and having questions and answers is the same or different then the general terminology question?
DR. MCDONALD: They're connected because sometimes forget that just having a whole bunch of general terminology doesn't get you all the way there without the thing you just mentioned first. Because people think that's all there is and the reason why you have to bring up the other one is because it's not all there is.
DR. HUFF: I think my view is that they're two ends of the same spectrum, one is just an extreme end, so at one end of the spectrum you have serum sodium and nobody cares how you ask the question, they want to know what was the milliequivalent for milliliter. Somewhere as you get closer and closer to very important things about disability and social functioning the representation of the question becomes much, much more important. So a term that's in SNOMED or in LOINC or any other terminology, the nuances of human language become important in the interpretation and so the social scientists work and they say the meaning of this, they validate the question and the questions' validity is dependent upon the context of the questionnaire that its asked in and all that sort of stuff.
So we want standardized codes for those questions but their representation is substantially different then the usual kinds of quantitative data or just asking did somebody vomit or not vomit or something like that. Are you depressed, which might be the simplest question that was implied by a diagnosis code of depression is wholly insufficient if you're trying to elicit the different shades of whether they're ready to commit suicide, all that sort of stuff.
So anyway, it's a spectrum and we want coded codes and well understood meanings for all of those things but I don't think it's, it's just sort of goes from white to black with a huge gray middle area and it's just different ends of the same spectrum whether you're talking about answers to questions or whether you're talking about specific findings in the patient I think.
This is a very philosophical discussion but I don't know that we're going to change anything they do so much, maybe just a suggestion how they focus the activities that you're --
MS. HARVELL: Stan, your first comment about the thing that we really need to focus on is the sharing of data and how are you going to send message, I mean not as part of this workgroup but as separate work we are working on in a relatively small project looking at is it possible and if so how is it possible to send messages about disability content related to quality measures and quality indicators and care panning information in the nursing home world using the minimum data set, which is the CMS required instrument. So we're beginning to go down this path of looking at the feasibility, the ability to share information using HL7 --
DR. HUFF: I guess the impression I was getting, and correct me if I'm wrong, is that you took those as immovable boundaries, that because it was adopted by the state and this one was a standard and this other one was a standard you couldn't change that. I guess what I'm saying is if you take that as an assumption then there are lots of kind of data sharing you'll never be able to accomplish. We may have to go back and examine the assumption and say is it really rational that we ask this kind of question in a nursing home or this other kind of question in the Social Security Administration on different, because there's certain kind of data correlation you will not be able to accomplish if you take those things as immutable requirements.
MS. HARVELL: I think in part that's true and so maybe this is Pollyannaish, but I think one at least my hope of having a, if there were a robust enough terminology that could be used in electronic health records, and then there were electronic health records that spanned a continuum of care, the fact that CMS slices and dices that clinical content in one way and nursing homes in a different way and home health, that I think minimizes the provider burden but gives CMS the flexibility that historically it has wanted and used in slicing information in different ways for different purposes.
DR. COHN: Gail, maybe you can help us with this.
MS. GRAHAM: I doubt that very much but I'm just trying to reflect on, because we do of course the FIMs and we do the MDS, but I think if I'm getting Stan's point and others is if you would ask those, if there would just be an agreement that the outcome you're looking for is X and you're going to ask it the same way, then I could pull from the same question on the nursing admission form or whatever, I could pull the same data for multi purposes but you would have to make, there would have to be an effort to standardize the instruments and have an agreement. I really don't think that much of this can be done after the fact, no matter how robust the terminology is, without going through that up front to say when we ask about X and we're trying to find out we will all agree to ask it in the same manner. I think that exercise has to preface the terminologies as it relates to all the data set scaling.
DR. COHN: Other Comments? I guess I'm, Marjorie, please.
MS. GREENBERG: Well, just I think an example in the survey in the international world is that there has been an effort to sort of back code or survey disability surveys that ask questions different ways, etc., to some kind of common terminology but, and it gives you, it's informative but it doesn't provide really true comparative data and there now is an international effort to try to agree on one or more general questions to be asked on surveys or in censuses and then more detailed questions because unless, I think there's been a conclusion, unless they really ask the same questions, and even asking the same questions as Stan said won't get you identical either because of course, well, international you get into cultural issues, language issues, but you also get into the setting in which the question was asked and whether you allow proxies, so I mean this is an evolving area but I think you can only go so far with back coding to the terminology or using the terminology.
DR. COHN: Other comments? I guess, I apologize, I'm sitting here thinking of what we can advise you and obviously you'd like comments from us as well as pros and cons. I'm sort of, I was recently on the IOM committee that dealt with patient safety recommendations, patient safety data, in some ways it seems like sort of similar problems that you have many currently existing databases, many maintained by the federal government, asking similar but not exactly the same bits of information and all this stuff and what do you do to begin to move towards increased standardization. And I think we'll be hearing some of this coming at a future NCVHS, hopefully response from that.
But on one hand you throw your hands up and you say well, we can't do anything until we completely reengineer the entire situation, and then of course we look at how many more years any of us have until retirement and knowing how quickly these activities occur, and yet you don't want to do nothing before perfection. So we're dealing with this funny situation which I'm sure you're finding yourself in where different agencies for many very, very good reasons are all asking either different questions or very similar questions in slightly different ways with slightly different scales and obviously what you'd like to do is to be able to behind the scenes normalize everything.
And I think what we're all sort of saying is that that may be a little unrealistic. Plus on top of everything else what you'd like to do is to have physicians in the routine care make notations using terminologies that could be easily abstracted without additional effort that would immediately also be the same as the questionnaires that you ask and that would be really where you'd like to go.
Now I don't think we're quite there today and I don't think Stan, who I think is a lot younger then I am, I don't even think he has that long in this career to wait for this to happen.
Now I guess the question I have for the subcommittee is what sort of, I mean given that we're not dealing with perfection but we're trying to help deal with next steps, what is it that we can advise CHI in terms of this and the other questions is what else do we need to do? This is a little unusual in the sense that typically the disability areas have been a lot held by our Subcommittee on Populations and obviously they aren't at the table for this discussion right now. So we may very well, I mean the question is are there things that we just want to say today, are there things that we want to have them come back in a month, do we need to get the Subcommittee on Population involved in this discussion knowing that I'm sure we'll be hearing a lot of this at the end of January. What do people want --
DR. HUFF: To summarize after all those other comments I would just say that in doing the things that you've listed here and it sounds like you're already doing this so it may just be, but if you focused on saying what the data and information were that you were hoping to be able to communicate and say how that would look in a message and let that be the guide to how you ask for terminologies and codes to represent that I think that would help, that would help focus the activity in a manageable way.
MS. HARVELL: Can I ask a clarifying question about that? Because I'm still learning this whole area, so if we undertake some activities to understand the information, the terminology that we need to send messages so we can identify where we are sending messages, but now it's identifying the terminologies, the content of what those messages are going to be, and just looking at it only with respect to disability because for example in the nursing home world we send messages beyond disability. But where do we go to look for the content for disability that will be the content of these messages that we want to send? I think that's the question that the CHI workgroup has been struggling with.
DR. HUFF: Well, there might be different answers, I mean you look where you can find them but ultimately you need everything that you need, so your need is not circumscribed by whether it's present or not present in a given terminology. What your goal is to make what you need available someplace where everybody can use it and so I would say that your goal is to either get it into one of the core terminologies, you'll get it into the UMLS meta-thesaurus, so I think part of your job is in fact to expand the universe, I mean again, use everything that exists, whether it's in SNOMED CT or in ICF, but if it's not there then the goal should be a plan of how to get those things into either SNOMED or ICF and both of them into the meta-thesaurus so it can be managed as a common resource.
DR. COHN: Jeff and then Steve and then I have another comment.
MR. BLAIR: I agree with Stan that it is very important to think of this information within the context of messages, but maybe I disagree with Stan that it's also broader then just thinking of it within the context of messages. And I'm not a clinician so I always feel uncomfortable when I start to step into these areas but in my mind the use for these codes is multiple. The codes for disabilities have to be useful for a caregiver, for a therapist, for somebody in rehabilitation that is endeavoring to either improve or help someone recover from a disability, and I think the needs for a terminology or some way of representing disabilities in that context is probably rather different then someone who's trying to get an overall view from the CMS or a federal government or a public health standpoint as to the disability status of a certain population. And while you might want to be able to do some cross checks or mappings between the two I think the information requirements of both is very, very different, so I think you have to consider that the different uses may require different answers.
DR. STEINDEL: What I've been taken about looking at the disability domain is a lot of the analogous nature of it to the laboratory domain, which of course in some laboratory and everything relates to the laboratory domain. But what I actually mean by that is in many cases we're looking at question/answer as Clem brought out at the very beginning of this discussion, and the disability domain, like the laboratory domain, is defined by the question you ask and the type of answer you need, and it varies. And we noted that when we wrote up the laboratory result content report.
There were some very specific questions that you can ask and since I don't know the disability area there may be some like this that have an extremely specific answer. Can the patient do, yes/no, so there are those types of questions. There's other types of questions that as you point out require a scale, an ordinal answer, a, b, c, d, or 1, 2, 3, 4, 5. And then there's the third type where you ask a question where you're describing something and you want a little bit more textual type content and I assume those three types of questions exist within the disability domain as well. And those are the types of questions that exist in the laboratory domain.
And in that area we very carefully crafted, and this is not something that we did today, there's been over ten years worth of work in crafting the way the laboratory domain question and answer area goes in that we've developed this set of question to ask for laboratory tests that exist in the laboratory area of LOINC and then we've developed a set of answers to go to those laboratory tests. Some of those are very straightforward, they're like as Stan mention the milliequivalents per liter of sodium where we just put down a number and a unit and others where we have textual information where we've called for SNOMED as the way to describe that textual information in messages.
And what I might suggest is that the disability domain look at this type of question and answer, structure, and right now the questions don't exist in any form of terminology but the ability to add the questions to a terminology exists very easily because you have as a minimum all these questionnaires that you're currently using that could form the basis to put those questions in some type of terminology very, I would say very quickly.
And one area, while we have two people that have to recuse themselves, one area where this is built for is the area of Clinical LOINC, where the questions can go in there and LOINC has the ability of having the specific numerical answers, the ordinal answers, your scale values, and the textual answers, and then we can look at SNOMED and make sure it provides the types of textual answers we want. And this gets to the point that Simon and Jeff were mentioning of can we use this at the point of care because we can build those questions into the various devices at the point of care and we can also address the granularity that Jeff was pointing to in the continuum of care.
MS. HARVELL: One of the issues that has come up with respect to for example some of the patient assessment instruments, which in many instances were developed 20 years ago in a totally different environment then today, for different purposes then what these forms are presently being used for today, that as we hopefully move forward into an electronic health record world, do we want to imbed these forms in the electronic health record world, or conversely, do we want to have the electronic health record world, do we want clinical information that's being collected electronically at the point of care be able to be used by for example the federal government in various uses. Which may mean that the forms that we use will change over time to be consistent, more consistent, with clinical data capture. So this is something that we are, we don't want to codify in an electronic world forms that may be, capture information that's not recorded at the point of care, and instead we'd like to be able to change our forms over time to better reflect what's happening at the point of care.
MR. BLAIR: There's another consideration that you may want to explore if you feel it's within your scope and your resources and your time. If an individual is blind it isn't just their acuity, it could be their field of vision, and the way you represent field of vision is totally different then you represent their acuity, and the cause could be different, and that effects how you deal with the disability. The other piece is if somebody has a broken leg you could wind up using video films to measure progress in rehabilitation in terms of measuring the degree of functionality on how people walk, the degree of their limb, or their gait. So I don't know if you consider that within your scope but I guess I'm making it more complex for you because I'm winding up saying, and part of this is coming out of the fact that one of the things that has been extremely detrimental to the disability community has been trying to lump all these things together.
I was on the State Commission for the Blind in New Mexico for a couple of years and fortunately we were able to remain independent. Many other states combined disability services for the deaf with disability services for the blind. Not only are the needs completely different, the problems, the psychological problems, the coping skills, everything is so completely different. So if you have the ability, the time, the resources to be able to look at the disabilities independently I would encourage you to try to do so in terms of your coming up with a measurement system that could be more valuable, more useful, for helping persons with functional disabilities.
DR. COHN: Thanks, Jeff. Steve, I'm going to let you talk but I do want to begin to, we need to come to a resolution on this one. I want to remind, other then for Steve and Marjorie we are actually not part of the workgroup, I just want to remind you all. It is not our job to solve our problem but more to reflect on what they have to come to us with.
DR. STEINDEL: My comment was with respect to the last comments Jennie made, if you do decide to embark, if the disability community decides to embark on a question and answer type approach, number one it gives a chance as Gail brought out earlier to harmonize the questions because the community would come together to work on what types of questions to submit to the various, whatever body you choose to submit the questions to so that gives a chance to explore the harmonization question. And in areas where we're well aware that there are regulatory reasons where some antiquated questions need to remain for whatever reason they need to remain, in those cases they might remain but they might remain in such a fashion that a hierarchy that you design of the questions can make roll-up of those questions to a more modern one easier, I mean that's one approach.
The second approach is as we get to the point where the world is more comfortable working in a unified environment, which we have to admit we're not today, but as we can now go back and look at those older questions and retire them, say they're no longer usable in the community and use the modern ones. So an approach like this would give you a way to build something that could exist today and transition into the world that we're looking at tomorrow. And into the world that you're conceiving of where we do have a continuum of care and different types of questions can be used throughout the continuum, but we can move through the continuum in the same fashion.
DR. COHN: Marjorie?
MS. GREENBERG: Well, I think everyone has made some excellent points to reflect on and I agree with a lot of what was said but I just wanted to make a few additional points. One is that whereas I definitely agree with Jennie, which has been the goal has been all along to try to capture the information that's collected or that's recorded, that's recorded at the site of care and then in a granular enough form that it can then be used for multiple purposes, that's been the whole PMRI mantra.
But I think one of the things in this area that we have to be cognizant of and then see how we can influence that was found by the committee in its report and in a number of articles as well in the literature that not that much and certainly not much standardized is being collected routinely, or being recorded routinely, on functional status and on some of these aspects of disability, as it is being captured on diagnoses and labs and procedures and all that. But often when you look in acute records or whatever, and probably you'll find it more in rehabilitation or therapy records, but it isn't being captured now so it's not even having the terminology that people can use, it's you've got to, there's work that has to be done on even getting this information recorded that's relevant.
And I think that relates to the second point, which is actually the first point I wanted to make, that I think Stan has alluded to it when he mentioned social participation, Jeff has alluded to some of this as well, but in the disability and functioning area you do go beyond what might be considered the medical model, you're looking at social participation, you're looking at environmental factors, you're looking at not just what can they do in the rehab facility but what can they do, can they get back to work, can they be in school and all of that. So I think this is also an underdeveloped area for the concepts and terminologies so there's clearly work that's been done. So I think those are important distinctions to be made here.
DR. COHN: I'm just trying to think of, we're sort of sitting here with a, I mean just to sort of try to figure out what we're going to do with this. We have CAHIT comments, we have disability workgroup recommendations actually on a page here that include the need for more research and certain I'm, I guess given where we are it's hard for me to imagine that we have an answer today, but maybe somebody does. I know Kepa was going --
DR. ZUBELDIA: Simon, I think that we have given a lot of feedback to CAHIT on how to address this and it's not just for disability, the same problem applies in other areas where you have a request/response and I think at this point what we need to do is say okay, there's the recommendation is not to adopt any of the existing standards because they don't satisfy the need and keep looking, and maybe bring this need to LOINC or to somebody that can set up a request/response type of standard.
DR. COHN: Okay, well I guess I'm wondering, maybe I'm maybe mishearing but it seems to me that there's sort of two issues, and I absolutely agree that there's a need for that response, question and response type issue. I'm actually sitting here wondering whether or not there's also a need for further work to improve the disability content in SNOMED because I'm not sure that there's, it seems to me that's sort of like you need to hit it from both sides, I mean Stan talked about it as a continuum but in reality we're sort of, a single terminology approach doesn't really equally handle both sides of the continuum --
DR. ZUBELDIA: It's a two dimensional problem, depending on the question you ask you're going to get a different continuum of responses, and I don't think that has been addressed with SNOMED or with ICF. And at this point I think our only recommendation is to say okay well let's wait for something like that to happen, the discussion is the feedback we're giving them, and I think that's probably all we can do today.
DR. COHN: Well, do we need to write anything down or do we want to wait until they come back to us in January?
DR. STEINDEL: I'm putting on my crafter hat. When we drafted the original template letter that we're using for the CHI letter itself we did note, and it's something that we haven't used before that we may want to modify and use in this case, and just to read the way it was in the original draft letter as an example, we note that the CHI Council has done extensive investigation into domain area blah, however we still feel that the area is immature and we recommend either a delay in implementation or implementation only on a limited experimental basis. We might want to use that as something to craft the actual, we did foresee something like this occurring.
DR. COHN: What wisdom, please, Jeff.
MR. BLAIR: One of the things is, Steve, if we use that wording --
DR. STEINDEL: No, we have to use different wording.
DR. ZUBELDIA: We don't have anything to implement.
MR. BLAIR: My thought is why don't we give credit to the work that's been done and the way they articulated their words in their recommendation section, which I think was that we are in a position to concur with the CHI position at this stage, which is that, and I don't have the words right in front of me, is that they are, no terminology will be recommended at this point and additional research is needed, I'm paraphrasing it but I think it says something like that and I think we can concur with that conclusion. No, yes?
DR. COHN: I think we are, the question is whether we agree with a through f of the additional research or whether we want to even deal with that at this point.
DR. STEINDEL: I think the problem is we may not agree with a through f.
DR. COHN: I guess the question is, this one is, I mean I think we all knew walking in today this one was going to be a little more problematic then many others and I'm wondering if maybe what we need to do is to sort of leave it potentially at this point, that this being the sense of the subcommittee, I know the workgroup is going to presume have other sessions between now and the end of January, is that, no? The workgroup doesn't exist anymore?
MS. HARVELL: We will modify the report to add the addendum, the subsequent information that happened since the report was initially written --
DR. COHN: You aren't going to review the recommendations?
MS. HARVELL: Feedback from you all? Yes, if you all give us feedback we will take that under consideration and meet to discuss that. If there is no feedback coming back from you all then we'll just, we have an addendum, review that report, and then pass it along.
DR. COHN: Gosh, I thought we were just providing a lot of feedback. Just be aware of the process, Jennie, our process is we're obviously hearing your domain recommendation, we would take a recommendation from the full committee in late January for final approval. And I guess what I'm trying to say is it seems to me that there's an additional cycle that needs to occur here where I think you've heard a fair amount of feedback from us. I think it does impact these recommendations, I don't know how it impacts the CAHIT comments or otherwise, but I think it may cause you to come up with a revised set of recommendations that I was going to suggest that you bring back to us in January on probably the 27th or 28th for us to review and have further discussion on was really my, what I was suggesting. Karen?
MS. TRUDEL: I heard an awful lot of discussion but I'm not sure I heard specific recommendations, and I want to ask Jennie and Sam whether they feel that they're hearing enough of a consensus from the subcommittee in terms of recommendations, that you know what to go back and talk to your workgroup about.
MS. HARVELL: Well, one thing that I heard is work on algorithms to equate within scope or out of scope, that's a question, and certainly part of our recommendations dealt with that. So we can go back to our workgroup and talk about that and communicate the sentiments that we heard form NCVHS about that.
One thing I haven't heard, and I think this is the kind of core difference between the workgroup recommendation and the CAHIT request for feedback, which is the workgroup was agnostic, was silent on suggesting where further work in terms of enhancing a terminology or a vocabulary should be directed, we just said further work is needed to build, to identify and build in needed disability content. The CAHIT recommendation, which is now not up there, was to direct future work at enhancing the disability content of SNOMED. So the CAHIT was a much more specific direction in terms of research efforts, work that will be done.
DR. COHN: I see, I'm sorry, and the current recommendations you had before us in writing sort of said you wanted to do an evaluation before you made that decision.
MS. HARVELL: The workgroup's recommendations that you have in writing, the a through f, said first let us explore the adequacy of the SNOMED hierarchy because by the time this report was written we were uncertain whether the SNOMED CT hierarchies would support the needed disability content. We had one conversation with the folks from CAP, and some offline conversations with others, that suggest at least in part the SNOMED hierarchies will support the disability data needs of the federal government. I think additional work is probably needed in the review of those hierarchies, and perhaps, I don't know if additional hierarchies would actually even be needed. With respect to social context and family issues, the representations by the people from CAP is that they in fact can capture that. I'm not sure how completely or adequate that capture is so that's a refinement or actually an expansion, that's something I think that should be explored.
The second recommendation from the workgroup was that a more complete content coverage analysis be conducted, with more time basically to identify whether or not SNOMED, ICF, and sources in the UMLS meta-thesaurus actually cover needed disability terms and concepts.
And then to enhance core terminologies I believe is what we said and my impression from the workgroup was that if SNOMED was found to provide the structure to support disability content that the work would go forward in enhancing SNOMED, to more completely cover disability terms. However, if SNOMED structure would not support the inclusion of needed disability terms then some other work would have to be done on some other terminology to make that other terminology or vocabulary more robust.
MR. BLAIR: Jennie? I think there was several other things that the committee thought you might consider as you go forward, one was Stan Huff's comments that you look upon the role that these things will play in terms of messages and that looking at it within that context is important.
An additional thing is that the information requirements for disability may differ significantly if you're looking at it from a patient care or therapy vantage point versus a public health standpoint, that probably needs to be considered.
And then a third area that I was suggesting is that if you're looking at it from patient care or therapy standpoint the needs of different disability areas can have very different measurement scales, even if you just pick one example within one area, just within the blind community the example I was giving is that a measurement scale for acuity is totally different then a measurement scale for field of vision.
So I think, I'm not saying you have to do this, I'm simply saying that consider the differences that the different disabilities have in terms of the measurement or information representation requirements. That grouping these all into one area and saying SNOMED could take care of it or International Classification of Dysfunctions and Disorders can take care of it, maybe that's from a public health standpoint but from a patient care standpoint I'm not sure that that's going to meet the needs.
DR. COHN: Steve, Kepa, Steve and Kepa.
DR. STEINDEL: Jeff, a clarification question for you. When you say consider the messaging needs do you mean the question and answer type format?
MR. BLAIR: I was really trying to make sure that Stan's comments and guidance was not lost, although I expanded on them it isn't just, from my viewpoint it isn't only messaging but I felt, and maybe Stan rather then my answering for you, maybe you could restate the gist --
DR. HUFF: Well, I think, messaging is an example. What I said is you need to focus on the fact that your goal is to unambiguously communicate data and information and that communicating that information is more then just a stack of codes, you have to consider the structure, and the message is one structure that allows you to focus that. But I think that goes hand in hand with some things that Steve was saying and what I mean by that is that will force you in fact to consider the question/answer paradigm because that's the most flexible and capable way to transmit this information. And then I think to just, to say, there are two different sort of spaces here, the question space and the answer space, and SNOMED may be entirely appropriate for representing sort of the answer space and what was seen in the textual descriptions. Unless there's a philosophical change I don't think for instance SNOMED is set up to handle survey questions per se, and so you might need to talk to LOINC or talk, and again, I'm in an area now where I need to recuse myself and not propose that solution but I think you might have to go someplace else if you really want to accurately represent survey questions and just as a factual case, I mean we're working with, LOINC is working with CDC and others because we have an interest in that and I'd be glad to share further information --
DR. COHN: Kepa, and then let's see how we come around to how we deal with this.
DR. ZUBELDIA: Since I don't need to recuse myself I would like to be very direct about it and say that in addition to looking at SNOMED and ICF, which are single dimensional, even though they're hierarchical but still single dimensional nomenclatures, that the recommendation would be to also look at LOINC, even though it may not have the terminology that you want in it already, that it provides the infrastructure that you need to develop a terminology that will meet the questionnaire type of environment that you need. So the recommendation is specifically to look at LOINC in addition to what you've looked at to provide the environment to support that question/answer --
DR. COHN: Kepa, I guess I would ask, I'm obviously looking at the recommendations here and I guess I'm wondering whether or not we want to be that specific as opposed to introducing, and I'm just looking at the recommendations on page four of the sheet we have here and under 2D they talk about algorithms, and maybe we really mean that they need to really be focused on the specific issues of sort of these question and answers and how it fits in, how the questionnaires can be specifically dealt with and leave it a little more open ended as a research component for them to figure out. Now I bring that as an option, I don't know how you feel about that. The question is what role does the question versus the answer, what terminologies, how does that all play out?
DR. ZUBELDIA: The fact that they've identified that none of these two terminologies that they have looked at have met their needs, and they're at the point they can't recommend either one because they don't meet their needs, and in the discussion it has become very clear that they have this question/answer bi-dimensional problem that needs to be addressed with bi-dimensional terminology. So one that provides that sort of infrastructure is LOINC, and there may be others and they look at other possibilities, but I think they need to look outside the box because they've only looked at three, SNOMED, ICF, and UMLS, and perhaps it's time to look at more then those three.
DR. COHN: Well, the reason I guess I'm asking is that even in one of the frameworks, you use one terminology for the question and another terminology for the answers, and that was really what I was trying to get to.
DR. MCDONALD: That's hardly ever the case, actually Simon. That's hardly ever the case and it should be avoided. It's not a question if you have, I mean you could say that I guess no answer could always be applied to a single concept but if you have anything more complicated then that there's other qualifications that go on the concept name that make it not be any confusion.
DR. COHN: Marjorie you had a question?
DR. MCDONALD: So you could take a whole bunch of SNOMED codes, or a whole bunch of IDX codes, and you could just declare by convention that these are questions whose answers are yes or no, and that's the one place you can get this confusion. But for practical purposes that's hardly ever, I mean I guess that is something, that's the one you have to worry about. But beyond that if you get to a survey instrument with a Lichert(?) scale one through five you've got other qualifiers on that symptom you're talking about that makes there be no confusion. They never say headache, ever, in any official survey. They'll say have you had headache within the last two weeks that was bothersome, or it will be qualified by a much longer thing. Is that making sense? It's only in sort of the simple case there is this issue you could take a single yes/no, like an ICD-9 code or any kind of code and say the question is present or absent.
DR. HUFF: But I don't think that's what Simon was saying. I think what you're saying is true, Clem, but I don't think, Simon was saying something different then that.
DR. MCDONALD: I thought he said there's the problem of the same code becoming the question or the answer.
DR. HUFF: No, I think he was saying that the solution may be to have the question come from one terminology and the answer come from a different terminology.
DR. MCDONALD: Never mind.
DR. COHN: Thank you Clem, I was beginning to wonder if I was completely missing it here. Marjorie?
DR. MCDONALD: I guess I'm like that Saturday Night Live thing with the bald eagle and eagle rights.
MS. GREENBERG: Am I correct that if you, if the LOINC approach, and I think someone put it well, it may provide the infrastructure, does provide infrastructure for this question/answer type of framework, but you still need terms, still need terminology. LOINC doesn't independently have the terminology, does it? Is there also the LOINC disability terminology or functional?
DR. HUFF: The thing that's happening within LOINC, and you guys just yell at me if I'm being inappropriate, in the usual case LOINC provides only the questions and so if you're talking in a usual clinical domain, for instance it might say something like heart rate or type of abnormal heartbeat. And then you would expect to find within SNOMED, PBT, atrial fib, all of that sort of stuff. And so it's this combination but when you get into survey questions then we're putting into place an infrastructure so that in fact the answers themselves are part of LOINC also. And so it's, but usually now it's not, LOINC is still not trying to incorporate the situation where it would say what is the reason for disability. The reason for disability would be a question in LOINC but the answer would be ICD-9, ideology codes or SNOMED codes or some other code system. The kind of codes related to this are usually the ones that are Lichert scale like or some other part, so those answers are but the other parts you would expect to find in some other terminology, in SNOMED or the meta-thesaurus or somewhere else, so I think that's --
MS. GREENBERG: I guess as a member of the workgroup though I'd defer to the chairs, but I think it's reasonable to ask the workgroup to go back and sort of absorb for the conversation here and I think we'll get, well, we'll get the transcript in ten working days I guess of this discussion and also the things that Jennie said she was planning to do also. But obviously it's not reasonable to ask the workgroup to come up with a solution by the end of January or whatever so it's a question of what you think is reasonable to recommend or to ask of the workgroup, or of the CHI process between now and the end of January and maybe in what we've referred to as the next phase.
DR. COHN: Sure, and I'll let Karen address that, I thought that maybe a revision to the recommendations as opposed, I didn't think we had a solution to give you that you could handle in four weeks. Karen?
MS. TRUDEL: I think it's reasonable for the subcommittee to say yes, we agree that, either you shouldn't be making a recommendation at this point, which is essentially agreement, and you may want to come back and say and we think that in addition to the points of study that you've already talked about you should add some more, and here there are, and the main one is the notion of the fact that SNOMED CT may not be well suited to the question and answer issue and that in phase two or whatever the group needs to cast the net wider on that, and also some of the notions about in looking at this in the future, at least addressing the possibility of disability specific consideration is again something that the workgroup might want to look at in a subsequent phase. And I think that's a really good recommendation and it gives the group something to work with and it does validate the main recommendation which is nothing at this time. And we can go back to the CAHIT and say that perhaps they were going a little bit too far in terms of SNOMED because we have this additional notion of the question and answer and that need to look for something else. I think that's, I would feel as if I had adequate feedback with that.
MS. HARVELL: I think one thing that would be helpful, and I don't know if this is something you would write down or an offline conversation, is to get more information about LOINC and the suitability of that because that's something that the workgroup did not consider, it's not something that this workgroup is pretty, is versed in at all in terms of what it can handle and what it can't, and where we could go to find out even more information about that would be useful.
MS. TRUDEL: We can help with that, too.
DR. COHN: I think we'll point you at Karen or maybe other people after they return home or whatever. Well, Steve, so what's happened with the wordsmithing as we've gone along here?
DR. STEINDEL: I'm not finished yet but basically the way I have it worded now is it starts off the same way with supplies, the NCVHS concurs with the recommendation to not adopt the terminology for the disability domain. I'm going to add a sentence, which I haven't constructed yet, which will say something to the effect of we advise CHI to consider adding the question and answer type format to the list of research areas they had in their recommendation.
DR. COHN: We might want to say we agree with the list, that there needs to be a future research agenda and we further recommend that it also include x, y, and z.
MR. BLAIR: Is it the sense of the subcommittee that we would not indicate you should also wind up considering the terminology specific to patient care and within the patient care area that disability specific areas might be different? Audiology different then blindness and different then physical therapy, and that ought to be looked at? Or did you decide that you don't want to take them in that direction?
MS. GREENBERG: You're absolutely right, they are different, but I'm not quite sure what the implications, and it certainly has implications for policy and the questions you want to ask and all of that but --
MR. BLAIR: Well, maybe it gets to whether or not --
MS. HARVELL: Well, it certainly raises the issue and that we ought to think about it, whatever the implications might be.
MR. BLAIR: Maybe it gets to whether or not how you define the scope of your disability terminologies, if the scope is expanded to including patient care and therapy then I think this would be appropriate.
MS. HARVELL: Again, one of the other things and I think this is on point with what you were talking about is that we've just begun thinking about is trying to define in the long term care arena the information needs and the information management needs in that area. Because right now what we have are these forms that the federal government said we need these bits of data for payment or quality monitoring for whatever reason we need these forms. As opposed to starting kind of from a blank slate and saying what is the information need in the clinical setting in long term care. And we've just begun thinking about that question and how to go about identifying those information requirements but if we actually sponsor some work in that area I think we could identify what our information needs are, we could identify what they are by disability domain condition, and that could inform the federal government how on point or not on point our disability data requirements are and whether we should be perpetuating those requirements or changing them and if so changing them how, and how to make these comparable across these different, or at least consider comparability across these different agencies.
DR. COHN: Okay, I think we need to begin to wrap this up. Now where are we?
DR. STEINDEL: I'll read what I have right now for the second sentence, the first sentence we agree with the recommendation not to recommend remains the same. We agree with the need for a future research agenda, concur with the basic list present, and also recommend addition of investigating the question and answer style format for this domain.
DR. COHN: Well, for aspects of this domain.
PARTICIPANT: [Inaudible.]
DR. STEINDEL: I was just going to ask that question, I've gotten to this point and now I wanted to see if we wanted to include the comments that Jeff was making and the comments that Jennie was just making about harmonization. I'm asking the subcommittee.
DR. COHN: I'm not sure I have a strong feeling on this one, Jeff, I guess I have to ask, while I agree that clearly different disabilities are very different, I also think that diabetes is a different disease then an acute MI and I don't necessarily think we need a different terminology now. The question is is what you're asking for is that the, that additional research be done to assure that the domain, the coverage is expansive enough to cover people with different types of disabilities? Is that really what we're talking about here?
MR. BLAIR: Close, and I'm not trying to be prescriptive, this is why I used the word consider in your research, because if you just check with ophthalmology and audiology and physical therapy groups, and if they happen to have measurement scales and they are there then that may meet needs, and if so maybe they'd be appropriate for adoption. I guess all I'm asking is that you consider and include that in your research and whatever you wind up deciding is what you decide.
DR. COHN: Jeff, just to restate this, so you're basically recommending that as part of the next phase of research that the workgroup activity survey a larger group of professionals and others for additional disability questions scales, etc.
MR. BLAIR: And I wasn't sure whether it's the next phase or whether it's this phase, if you can't do it in the timeframe for January then it's the next phase.
DR. COHN: Well, I don't think we'll have it for January.
MR. BLAIR: Okay, fine.
DR. COHN: Jennie, you guys think you can have it for January?
MR. BLAIR: Then your wording is appropriate.
DR. COHN: Steve, I think you're asking a question.
DR. STEINDEL: Yeah, I would like to just make an observation on Jeff's comments. While I totally agree with Jeff's comments I think what we're basically recommending is that we feel that there isn't a terminology that's ready for prime time at this stage and that even with your basic recommendations we realize that the terminology, whatever terminology that's out there needs to be further developed. And in the aspects of terminology development it behooves, for us to accept a good terminology is to involve all the people that are going to be using that terminology, so I think just in the course of developing the terminology Jeff's needs would be met, and I don't think we have to spell it out. At least right now, we might in the future but right now I think we should just assume that it's good practice in terminology development.
DR. COHN: Well, Jeff, I'm looking to you to see what you think of that.
MR. BLAIR: I've offered my suggestion and I'll accept whatever the subcommittee decides to do.
DR. COHN: Marjorie?
MS. GREENBERG: I hear what Jeff is saying and I think in this disability area it is somewhat different because it involves not just different medical specialties but as I said before a broad range of users and domains that actually go beyond the health domain but are treated in the health domain, such as transportation needs, accessibility issues, benefits, all sorts of stuff, so that maybe just, which is a little different then what he's saying but I think is reflective of the fact that this is a very multidimensional domain that involves a lot of different perspectives and a lot of different users and needs. So maybe just saying that in further terminology development as well as the question/answer development the broad range of aspects of disability and people who deal with disabilities should be consulted.
DR. COHN: It's just a recommendation, I think that's fine. Is that okay with you Jeff, that sort of, okay.
MS. GREENBERG: I think that sends a message, too, to terminology developers, which is good.
DR. COHN: Now let me just ask, is everybody comfortable with this? Stan's nodding his head, Kepa, you're okay? Jeff? Okay, I'm okay, Clem, are you okay with this? I heard him say yes --
Now I think we're seriously behind schedule but we will forge ahead anyway. I'm actually going to suggest that maybe we take a ten minute stretch break at this point.
[Brief break.]
MS. GREENBERG: Did you vote on that by the way? Did you actually vote on that or that you had consensus.
DR. COHN: I think I heard, all in favor say aye.
SUBCOMMITTEE: Aye.
DR. COHN: Opposed say nay. And I guess I'm hearing from one person that we ought to probably forge ahead because I think, I've been reminded we do lose a quorum in about 40 minutes.
DR. STEINDEL: Can I request that I make certain what we just approved?
DR. COHN: Sure, please, do you want to read it?
DR. STEINDEL: Yes.
DR. COHN: Please.
DR. STEINDEL: The NCVHS concurs with the recommendation to not adopt a terminology for the disability domain. We agree with the need for future research agenda, concur with the basic list presented, and also recommend addition of investigating the question and answer style format for aspects of this domain. Request that the future research be inclusive of all domain stakeholders.
DR. COHN: Sounds good. Let's move on to Gene and Protein.
Agenda Item: CHI Final Reports - Genes & Proteins - Mr. Sorace
MR. SORACE: I'm here for the gene and protein group. Basically we had a pretty good representation and we had a lot of very fun and spirited contributions from people and both the CHI, NCBI, FDA, VA, and CDC. And basically our main goal was to allow the federal health care sector to exchange information regarding the roles of gene and proteins in biomedical research and health care.
It was an interesting thing just to even try to scope out because it's a very broad field and we really broke it down into five domains, and actually the top three here are the more interrelated domains. That is we were thinking more in terms from the medical perspective, again looking back into a basic science background and that health care professionals would like to know inherited genetic variations that may contribute to diseases, and this is your classic gene susceptibility traits, known genomic diseases, or genes that are associated with responses to pharmacological agents like the mutations that are found in cancers that may break cancers down diagnostically or prognostically. Another one I think which has actually turned out to be one of the most interesting was infectious diseases, and that is it's not just human genes that are important but it's the genomes of those things that like to give us diseases, and that includes issues such as genes involved in bacterial or viral pathogenesis, drug resistance or actually just pathogen identification, so figuring out what the patient has.
And then we had two really narrow domains and actually they're more just what do you name these things so we were looking in protein nomenclature and gene nomenclature --
DR. HUFF: Question. Are you considering the names of laboratory tests that determine the presence or absence of the gene, is that in scope or out of scope or was that part of laboratory considerations or how --
MR. SORACE: Well, I'll answer that two separate ways, it was definitely in scope in terms of gene names. In other words not so much laboratory tests but that you can link laboratory tests to the gene that it identified somehow, so at a minimum we wanted a genomic nomenclature that allowed you, that could then be linked to like LOINC codes, for example.
We had some debate and we actually think that it may actually fall more into a gap in terms of how you expand these concepts in various other areas. But I think that our way of looking at it was not necessarily that we'd want a code that filled both the test function and the gene identification function in one code, but rather you'd want something that would allow you to unambiguously specify a gene that could then be linked to say LOINC codes for laboratory purposes.
And we looked at several alternatives, the Human Gene Nomenclatures, which is supported by HUGO(?) and the Gene Ontology Nomenclature and SNOMED CT. And basically we recommend, at least for human genes, that we adopt the Human Gene Nomenclature, sponsored by the Human Genome Organization, which is directly funded by NIH and the UK. But we're aware that when you look at the rest of the things which we said that we'd really like to be able to do with genes and proteins we couldn't find an adequate protein nomenclature, the field's just to young and everything is very developmental. We couldn't find a good nomenclature for non-human genes that was somehow comprehensive and would allow you to put all the non-human pathogen genes that you may need to express. And then in these more integrative questions we felt that there was a lot that needed to be done in phase two and for example the field of infectious disease actually represented a very important area where a lot of work needs to be done. There needs to be more work in not only the nomenclature for pathogen genes but also just in terms of nomenclatures for annotating pathogen genes functionally, so like which gene confers ampicillin resistance say, or this gene confers adherence factor that's important to urinary tract colonization, that those basic vocabularies, it was very hard to find something that was satisfactory in all those roles.
But it wasn't clear, and it's something that we need to look into more later, whether or not something like SNOMED or some other nomenclature could in fact actually be expanded to fill some of those issues. Or whether LOINC could. So I think the main view of the committee was that in many cases the current vocabularies that are already adopted by the CHI effort could actually fill these gaps, or could be tasked to, and that it might not be wise to go and try to recommend one now because it's too early.
Another thing we noted was that just more active coordination to cover known agencies would be very useful for translational research, and also for disease surveillance in bio defense. And that in terms of infectious disease CDC input and in coordination with NIAID would also be very, very useful.
The National Cancer Institute is trying to adopt and develop nomenclatures for a lot of the acquired genomic tests that you can find in a cancer example and they think they may have something ready in about six, seven months, but they didn't feel that their efforts were ready for recommendation at this time. And I think in terms of genomic diseases a very useful exercise would be to go back and take some of these well known genomic databases like OMIN(?), say or Gene Tests and go back and see to what extent they could be SNOMED coded and see to what extent we could fill in with existing vocabularies and then consider in what areas those may need to be rounded to meet the effort, versus going back and starting to try to find if there's a new vocabulary de novo.
So we came up with I think some useful recommendations in terms of human genomic nomenclature, fairly narrowly focused recommendation, I hope we've sort of provided a framework for future analysis in this area.
Any questions?
DR. COHN: Jim, thank you. Comments, questions? After the last one, simply silence, Michael.
DR. FITZMAURICE: It says that the HUGN(?) contains names for approximately one half of the expected number, oh that's right, this is for protein, not for genome.
MR. SORACE: Well, the issue is what's a gene and it's known that you, there's about 35,000 expected protein coding genes but they keep finding all these other sequences that have physiological significance.
DR. FITZMAURICE: So I guess the question is HUGN still a work in progress?
MR. SORACE: Yes.
DR. FITZMAURICE: And you're saying it's good enough to be adopted for the uses for which the government would use it today?
MR. SORACE: It currently is, I mean it's basically being used by NCBI and Locus(?) Link, and they all use HUGN notation almost exclusively. If they do assign a name outside of that nomenclature they only do it temporarily until they get a name assigned.
DR. COHN: Jim, just to make sure I understand, I mean basically HUGN is state of the art, it's that the other half hasn't been discovered yet. And people are basically, if you discover them it's hard to assign a number to them.
MR. SORACE: And there's nobody that I'm aware of who's done one percent.
DR. COHN: That's fine, that was what I understood from the reading. Kepa?
DR. ZUBELDIA: You say here that it's usually is free for non-profit uses, does the government have to pay to use it or is it considered commercial use, non-profit, and what's the license for commercial use?
MR. SORACE: It's free to the government. Somebody like Solera(?) or somebody who wants to do it commercially might have to go them for a license.
DR. ZUBELDIA: Do you know what the terms of the license are?
MR. SORACE: We haven't been able to get that information, we're still trying --
DR. STEINDEL: Is a for-profit hospital considered commercial?
MS. WARK: I think probably the best way to answer the question is that we're trying to get information from the organization and we have some preliminary information that basically indicates attribution to the organization but essentially free for use, I mean that's what we're hearing, we're trying to clarify a few issues because we don't want to ultimately put something out in the report that isn't accurate but we're still working on it.
DR. STEINDEL: And of course my question does not pertain to federal to federal exchange but the idea of CHI being a tipping point.
MR. SORACE: I don't think there'd be a problem for a non-profit hospital --
DR. COHN: Well, what do we want to do, Steve?
DR. STEINDEL: I do notice that in, I think the report is excellent and agree with the conclusion but it's very targeted as you noted to the human genome and the domain was genes and proteins and I don't think it really specifically notes the other areas as either gaps or no recommendation or something like that and maybe we should specifically close the loop on non-human genomes and proteins.
DR. COHN: So you're recommending no recommendation for those areas.
DR. STEINDEL: I think there's no recommendation at this time and maybe a note for research or development.
MR. SORACE: I think we're in agreement, we didn't really find anything out there, it will have to be created.
DR. STEINDEL: I'm just wondering if the report should note that --
MS. TRUDEL: We can make a specific statement to that effect, I mean the lack of a recommendation essentially gets you there by the back door but we can make a specific statement that says we are not making any recommendation for anything other then human genes.
DR. COHN: And maybe what our recommendation says is that we recommend, we concur with the recommendation for HUGO for human gene nomenclature and with the lack of a recommendation for terminologies for, and I think what is it, acquired gene, inherited genetic variation, acquired gene changes, infectious disease, gene proteins, and protein nomenclature, that would basically just sort of, and oh yes, the remaining four, exactly.
I guess the only other question I had, which Kepa was bringing up, is there a condition that we need to note having to do with clarifications of licensure? It's hard to imagine it's going to be too much since it's, well, actually you never know, I would say it's jointly funded by the UK and the NIH but do we, our letter need to recommend that or did you want to, what are thoughts? Kepa?
DR. ZUBELDIA: Perhaps we need to recommend that the license, a low cost non-discriminatory license be secured, or be provided or something.
MS. WARK: This would be for outside CHI, is that correct?
DR. ZUBELDIA: Well, interfacing CHI with hospital chains or for-profit hospitals, for instance.
DR. FITZMAURICE: For example, suppose VA or DOD sends some of their patients to a for-profit hospital because they have the expertise and they have the genetic information coming back. If they don't have it coded in the same way as their own genetic information --
DR. ZUBELDIA: Or you need to run a generic test somewhere else, you need to contract, farm it out somewhere.
DR. COHN: Well, what are the thoughts of the subcommittee on this one? How exactly do we want to frame it? Do we want to not approve it if there's not a low cost license negotiated or we're asking more for clarification of the commercial license fees?
DR. FITZMAURICE: With a recommendation that they get the specific information that they've described already about the licensing, that way the recommendation goes up to the Secretary that yes, we find it suitable, but there is a concern. Doesn't seem to be anything in second place.
DR. COHN: Well, that's true.
DR. ZUBELDIA: It's not a big concern because HUGO is a non-profit and it's partly funded by NIH, so they're not going to be having a huge license fee, it's not that big a concern but there could be some discriminatory terms that you cannot use it for certain things or whatever the terms are.
DR. COHN: Karen?
MS. TRUDEL: Could I make a suggestion? It may just be appropriate to say that we recognize that there's still work being done on the terms of the licensing agreements and the subcommittee does have concerns about how licensing fees could impact the use of this in the private sector, period. I mean that doesn't direct us to do anything, it just suggest that we continue to look into the concerns as we have already said that we are doing.
DR. COHN: Does that address everyone's needs? Kepa?
DR. ZUBELDIA: How about, a concern that I have is that you haven't been able to find out what the license terms are.
MS. WARK: We have some initial response from the organization but it is not clear to us that all of the questions about the licensure and any fees associated with it for commercial use is clear to us, so we have referred the matter to NIH as the co-funder of the organization to seek clarification on that and its just pending.
DR. ZUBELDIA: What if that's a requirement? That whatever the license terms are that they be explicit? I mean it's a little scary --
MS. TRUDEL: The suggestion you're making is that we not adopt this even though there is nothing else there for the federal health enterprise because we don't understand the consequences on the commercial market. I'm not sure I agree with that.
DR. ZUBELDIA: No, that's not what I'm suggesting, I'm suggesting that HUGO, I'm suggesting that it be adopted but that HUGO explicitly state what the license terms are.
MS. TRUDEL: And we're working on that. I'm not getting it.
DR. COHN: I'm not sure what the issue is here. What do you have at this point?
DR. STEINDEL: I actually have to agree, what I have so far is as the first sentence, the NCVHS concurs with the recommendation for the genes and protein domain as noted for the human genome, and note the need for explicit comment on the lack of terminology for the remaining sub-domains. The second sentence, which is pertaining to the discussion that we have, what I've gathered so far is that we recognize that work is still underway on clarification of licensure terms for use outside of the federal government. And I don't really even see the need for that sentence.
MS. GRAHAM: Is it that this committee wants to be apprised if there is a cost so they can readdress this if there is a cost associated?
DR. FITZMAURICE: For all the others we've known the cost. This is the first one we haven't really known the cost, right?
MS. TRUDEL: But we know the cost to the federal health enterprise.
DR. STEINDEL: We do know the cost to the federal health --
MS. TRUDEL: We do know the cost to the federal health enterprise, it's zero.
DR. FITZMAURICE: But if the federal enterprise uses the private sector to do its testing, to send patients out there, there is a cost bourn by it --
DR. STEINDEL: Maybe, maybe not.
DR. FITZMAURICE: Maybe, maybe not, so we just don't know until we see the terms.
DR. COHN: So our question is is do we leave it like that or do we say, do we want to encourage the Secretary to consider negotiating the low cost license for use by the private sector to encourage more widespread implementation? Or do we want to hold that end since they don't know, I mean we're in a situation where they don't know, they will know next month I presume. Will you know by late January? Okay, can we make a note on this one to based on what we hear back from them we can decide whether to put a sentence in there or not?
DR. STEINDEL: Simon, that brings me up with just a basic question I was going to ask later, are we going to send two letters or just one letter in January?
DR. COHN: We're sending one letter at the full committee --
DR. STEINDEL: At the end of January. Then I think what we should just do is for me to bold the sentence so we know to go back and revisit it in January.
DR. ZUBELDIA: And then in January remove it, remove that sentence altogether.
DR. STEINDEL: Hopefully yes.
DR. COHN: Or some other action. Okay, it is now 2:38, people want to keep, I think it would be nice to take a break. I'm sorry, you can tell I need a break. Are we, Kepa, do you move acceptance?
DR. ZUBELDIA: Yes, I move that we accept this recommendation with a sentence that Steve --
DR. COHN: Jeff, are you seconding that?
MR. BLAIR: I'll second it.
DR. COHN: Any further questions, comments? All in favor?
SUBCOMMITTEE: Aye.
DR. COHN: Opposed? Abstentions? Okay, this one passes. Now should we take about a ten minute break and reconvene at ten minutes to 3:00. Is that okay? Great.
[Brief break.]
MR. BLAIR: Just so that you're aware, we have five more of them to go through, two of them are going to be final recommendations with votes and we wanted to make sure we got through that while we still had, before Kepa had to leave, and we have three more preliminary ones, which are informative. So at this stage let me ask the testifiers who would be from CHI who were going to be discussing the diagnosis and problem lists, to introduce themselves and to take us through.
Agenda Item: CHI Final Reports - Diagnosis/Problem List - Ms. Acker
MS. ACKER: My name is Beth Acker and Karla Porter is also here with me, we are the co-chairs of the diagnosis problem lists CHI group.
Our group was composed of members from HHS, DOD, VA, and basically what we did is we had a really tough time narrowing down our scope for this and so what we did was we described the diagnosis and problem list as a series of brief statements that catalog a patient's medical, nursing, dental, social, preventative, and psychiatric events and issues that are relevant to that patient's health care, for example, signs, symptoms, and other defined conditions.
From that we took our scope and we decided what was going to be in scope and out of scope. We considered clinical diagnosis and problems, subjective symptoms, observed findings, and synonyms to be within scope. We considered nursing diagnoses, modifiers and descriptions, dental, and alternative medicine to be out of scope for this particular realm.
Now that we had narrowed down our scope we identified alternatives, those were SNOMED CT, ICD-9-CM, ICD-10-CM, ICPC, MedDRA, MEDCIN, and DSM-IV.
Now I'm going to read an excerpt of our summary for the basis of our recommendation. The team used as a starting point the criteria for PMRI put forth by the Subcommittee on Standards and Security of the NCVHS, that being concept orientation, concept permanence, non-ambiguity, and explicit version IDs. The CHI workgroup then collected information on additional items, which are content coverage, scope, settings, ownership, cost and availability, usage, mappings, and other relevant considerations to evaluate all candidate terminologies. The main distinction between the analysis conducted by the CHI workgroup is that the NCVHS Subcommittee on Standards and Security used the criteria to identify terminologies that would be considered further for recommendation as a core PMRI terminology and CHI considered all criteria and information gathered for additional items to make our recommendations. This resulted in identification of what we identified as the best solution for the government.
The next slide shows that our recommendation is SNOMED CT --
MR. BLAIR: Could I just get that clarified? When you say best solution for the government, what are the uses that you're thinking of when you say best solutions for the government? Are you thinking reimbursement, public health and research but not patient care? Or does that include patient care because of DOD and VA?
MS. ACKER: It includes everything, patient care also.
MR. BLAIR: Thank you.
MS. ACKER: The diagnosis and problem list code sets that were evaluated by NCVHS, Subcommittee on Standards and Security, did not meet all of the essential criteria for patient medical record terminologies, whereas SNOMED CT and MEDCIN referenced terminologies met all of the essential criteria. Content coverage in MEDCIN and ICD-10-CM is limited to diagnoses used by physicians for the purpose of generating bills. SNOMED CT has the most comprehensive content coverage including diagnoses and problems identified by nurses and allied health care providers, therapists, social workers, dieticians, etc. The MEDCIN referenced terminology is available for a small fee whereas the SNOMED CT is now available to all U.S. users without additional charge.
Therefore the overall analysis of available alternatives resulted in SNOMED CT being selected as the best choice for use in federal government as a PMRI terminology. This will need to be used in conjunction with administrative code sets to be used for billing purposes under HIPAA and with interface terminologies developed by a host of other parties such as vendors, government agencies, and so forth.
The workgroup anticipates that the SNOMED CT concept groups of disorders and findings will be the primary source for the diagnosis and problem list entries, although it also anticipates that most, if not all of the other concept groupings, such as body structure, organism, social context, will also be used instead or coordinated with the former to accurately express a diagnosis or problem list entry. The entirety of SNOMED CT is not being recommended, only the content that contains the diagnosis and problem list concepts as modeled as integrated from the source PMRI terminologies.
Therefore what we did was we recommended that there needs to be mapping done for ICD-9-CM, ICD-10-CM, MedDRA, and DSM-IV, and those are examples, they aren't exclusive to just those, and we said that these needs are being referred to the NLM to be handled.
And basically in the interest of time we condensed our presentation to that so if you have any questions for us --
MR. BLAIR: Are there questions from the subcommittee?
DR. STEINDEL: There is being raised, not necessarily on CHI but in general, about primary care classifications and you had ICPC as one of the terminologies you were considering. I was just asking if you would consider expanding your mappings to add ICPC.
MS. ACKER: That's certainly fine, again what we had named were just an example thereof, not exclusive to, so we can certainly add that one in there.
MR. BLAIR: Any other comments? Carol?
DR. BICKFORD: Carol Bickford, American Nurses Association. The group that was taking a look at the diagnoses and problem list excluded alternative medicine, complementary therapies environment, what is the intent about making sure that those practitioners using that frame of reference are supported in their care delivery process? And the second question that I have is if we're looking at health promotion and wellness how are those concepts integrated to make sure that some of those similar to complementary alternative medicine concepts are incorporated? So it's a two pronged question.
MS. ACKER: You might want to stay at the microphone for a second, I may need to ask for further clarification if we can't respond adequately to your question.
Basically what we consider those types of things is to be very specialty oriented, and that would be where the mappings would come in, that if there is something that is specifically needed to describe say alternative medicine, the ABC so forth, something of that nature, then that be referred as a mapping type category. And it's not a core terminology, again, what we tried to do or focus on what's got the basic overall where can we start from concepts, and that was how we made our recommendation is that this is what we feel has our overall basic concepts covered for all of these things.
MR. BLAIR: The only observation that I would have is you did recognize the ABC codes for billing but not in nursing and not in diagnosis and problem lists. And if the data isn't captured in that form during the patient care nursing activities, then it becomes pretty difficult to use it for billing. How do you wind up accommodating that?
MS. ACKER: I don't think that we meant to intimate that we exclude that, exclude anything, as far as a mapping type situation, again, we saw that as something that was specialty related that could be referred to as a mapping issue. We were trying to sort out, say this is what we feel is our core base and we feel that there are many other specialty type segments that need to be identified and mapped to and as those are identified and brought to the forefront those would be referred to NLM for mapping purposes.
MR. BLAIR: Is there any problem with including a phrase that you exactly, what you just said, that you are assuming that there will be other specialty related diagnosis and problem areas that would be mapped including alternative codes?
MS. ACKER: I don't think we would have any problem with that, we just said for example and we named a few, again, it wasn't an exclusive list.
MS. TRUDEL: Right, and we did in the write-up specifically define both dental and alternative medicine as out of scope for phase one.
MS. ACKER: Correct.
MR. BLAIR: Any other comments? Carole Bickford? Marjorie?
MS. GREENBERG: I just want to clarify that when you said they would be candidates for mapping you're suggesting that the mapping would be done and then if gaps were found they could be added.
MS. ACKER: Correct, exactly, because as we, we recognized that the timeline that we had to do these investigations, and it's not possible to identify everything. We wanted to identify their base and leave it open that if additional gaps were identified they could be addressed.
MR. BLAIR: No more comments? Anyone from the audience? If there are no comments is there a motion that would reflect the subcommittee's --
DR. STEINDEL: Jeff, if I can read what I captured. The NCVHS concurs with the recommendation for the diagnosis and problem list domain and requests the addition of ICPC to the recommended terminologies for early mapping efforts.
DR. ZUBELDIA: I move that we rule that wording.
MR. BLAIR: Is there a second?
DR. HUFF: I'll second that.
MR. BLAIR: Okay, any other discussion? Okay, shall we vote? Concur? Okay, the NCVHS concurs with the wording as Steve described it and thank you very much for your testimony to us, and Jorge Ferrer you're our next testifier.
Agenda Item: CHI Final Reports - Non-Laboratory Intervention and Procedure - Ms. Wark
MS. WARK: I am going to be presenting the recommendations for the non-laboratory intervention and procedure workgroup on behalf of Dr. Jorge Ferrer from CMS who was the lead for this workgroup and the team composition included representatives from NCHS, VA, DOD, and Indian Health Service. And as the subcommittee may recall Dr. Ferrer came in May and presented the preliminary report so this workgroup was convened in the spring time and the report was actually finished some time ago.
The domain description is that the standard will be used to describe specific non-laboratory interventions and procedures performed or delivered in the provision of health care.
The scope includes procedures by site, method, intent, focus, device, and regimen therapy. It excludes again, very similar to the diagnosis and problem list workgroup, dental, alternative medicine, and the laboratory procedures were addressed workgroup for laboratory test order names and administrative and management procedures were also excluded.
The alternatives were SNOMED CT, MEDCIN, ICD-10-PCS, and CPT-IV.
The final recommendation is that SNOMED CT be adopted and that the specific location in the SNOMED CT hierarchy that forms the basis of the recommendation is the procedure access, and that is excluding the administrative and laboratory procedure areas.
The content coverage was assessed across all of the alternatives and it was found, again, a very similar analysis was done on both the diagnosis and problem list workgroup and the non-lab intervention and procedures looking at first the PMRI criteria of concept orientation, concept permanence, non-ambiguity, and explicit version IDs, and then the workgroup went on to look at additional factors, content coverage, settings, scope, ownership, availability cost, use, mapping, and considerations.
And after taking into consideration the information gathered in all of these areas the workgroup came to the conclusion that SNOMED CT would best meet the needs of the government due to its extensive content coverage compared with the other alternatives, and the fact that the scope includes content for multiple disciplines involved in health care, it can be used in multiple settings, again, due to the free availability to U.S. users through the UMLS, and in terms of gaps there was noted a lag in the timeliness of emerging procedures incorporated in SNOMED CT.
The workgroup did look at some specific procedures that are relatively new and did not find those to be present in SNOMED CT and so is making a recommendation that every attempt be made on the part of SNOMED and NLM to address the need to incorporate those terms as quickly as possible. Also noted was that mappings are needed to be maintained between SNOMED CT and CPT-IV and ICD-10-PCS via the UMLS.
And that concludes the workgroup's report.
MR. BLAIR: Okay, questions, comments from subcommittee and staff?
DR. STEINDEL: Cynthia, the workgroup is making a point about the lag time, is that just an incidental comment? Because it's actually, it's a bit worse then they may let on because there's a lag time between when SNOMED includes the term, and then there is a further lag time as to when it might appear in the UMLS and there's a great deal of concern about shortening that lag time to as much as possible but it can never be reduced to zero. I think I, I agree with the emphasis but I would just like to make sure that it's just an incidental comment, that it's something they would like to see improved, and doesn't have any real bearing on the recommendation.
MS. WARK: It's not a condition in the sense that the terminology could not be used, however, the workgroup felt it was important to stress the need for timely incorporation of concepts associated with newer interventions and they did discuss the fact that it takes some time for CAP to respond by incorporating into SNOMED and then into the UMLS version and then the release of that version. I don't know that the workgroup had any direct involvement in the discussions that were taking place between CAP and NLM in terms of how they can improve the timeliness, I'm not sure that the workgroup didn't have any specific ideas about how that could be done quicker but they did feel it important to note it in the report.
MR. BLAIR: Other comments?
DR. ZUBELDIA: Yes, I have a question. With all this piecemeal adoption of SNOMED where specific access is adopted by one of the workgroups and a different access by another workgroup and there is some exception, has anybody looked at what's left? Is there something SNOMED that has not been recommended for adoption?
MS. WARK: Well, specifically in the case of this domain we are excluding laboratory test order names because LOINC was recommended for adoption in that area. It's probably like you say, as we proceed with each of the domains and we begin to knit together the totality of what is being adopted in SNOMED it will probably be more the exception that something isn't included then the rule. But because we haven't completed the entire portfolio yet, no, that work hasn't been done, it seems like we could take that back as a recommendation for phase two and the harmonization efforts that we begin to do that and specifically look at all of the components that have been adopted and those that haven't.
DR. ZUBELDIA: I would like to see the identification of those things that are excluded, not necessarily for this specific workgroup, at some point CHI I think ought to take that --
DR. STEINDEL: Kepa, just looking at the axis structure of SNOMED we have probably recommended maybe a third but less then half of the axis of SNOMED for adoption. In terms of actual concepts my off-hand guess, though I haven't really counted them, is there's about 300,000 terms in SNOMED, we've probably recommended about, concepts in SNOMED, there's more terms, we've probably have recommended about half for CHI purpose. And the things that we've excluded are things like physical force, physical object, the special concepts, social context axis, items like that is what's been excluded. I would say in terms of the actual clinical axis we probably have touched all of them.
MR. BLAIR: Any other comments, questions? Anybody from the audience. Carol?
DR. BICKFORD: Carol Bickford, American Nurses Association. In light of the federal government's significant initiatives in the tele-health arena how is that going to be encompassed in the interventions coding world? The focus has traditionally been on one to one interaction between clinician and the patient, so I'm raising that as a question in the discussion as we've moved forward, particularly to support the federal government, with their great diversity in their tele-health initiatives. Is that part of the intervention component, is it just an incident, how is that being accommodated?
And secondly, in relation to the interventions, are we going to be, the non-laboratory ones, was there a look at the public health component, the significance of that work in relation to the emergency preparedness and some of those initiatives? Looking at the patient as being the population not the individual.
MS. WARK: If I understood those questions correctly the first one being how are we addressing tele-health, and the one specific area where that has come up and is not being reported on today but there was a preliminary presentation in October is clinical encounters. And there has been some effort to look at terminologies for clinical encounter information that can be applied across all settings including tele-medicine.
On the second question, in terms of population health, there is, Steve Steindel will be offering a report, a preliminary report, from the population health workgroup later this afternoon and that may offer some insight into your question.
MR. BLAIR: Other comments? Okay, if there's no other ones, before I indicate whether or not, or call for a motion on this particular one I failed to mention in our last vote that in addition to the members that are here, Stan Huff, Kepa Zubeldia, and myself on the Subcommittee on Standards and Security, Simon Cohn instructed me to also add his vote on diagnosis and problem lists as concurring so for the record there were four committee votes, subcommittee members voting for it the last time, and he also instructed me to add his concurrence on non-lab interventions as well. So is there a motion for non-lab interventions?
DR. STEINDEL: Jeff, if I may read what I've captured because I haven't heard any other comments, is the NCVHS concurs with the recommendation for the non-laboratory intervention and procedures as presented.
MR. BLAIR: Any second?
DR. ZUBELDIA: I second it.
MR. BLAIR: Okay. Any discussion? Questions?
MS. GREENBERG: I'll have to check into it, as the executive secretary I should know the answer but I'm not quite sure whether we permit proxy votes but on the other hand given that, if they're given in advance and then there's discussion etc., but given that there's unanimity it's okay but I'll check into whether that, that's a little atypical.
MR. BLAIR: The only thing I'd add then to add to your examination of that is that Simon did read through the information before he rendered his assessment.
MS. GREENBERG: I'm sure. As long as you don't change it he'd agree with it.
DR. STEINDEL: And we didn't.
MR. BLAIR: Okay, then I think --
DR. ZUBELDIA: Let's take a vote on this one.
MR. BLAIR: I think we're prepared to take a vote, and can I hear the votes? I concur as well and Simon instructed me to give his concurrence, and so the motion is passed.
We have concluded all of the CHI recommendations, I'd like to thank everyone from CHI who has done the work to prepare and present that to us. I hope that the NCVHS comments and review and suggestions has been constructive and helpful.
MS. TRUDEL: We'd like to thank the subcommittee for giving us all of its time and for plowing through reports at very short notice and providing us with very useful feedback, so thank you also.
MR. BLAIR: We have three additional preliminary CHI recommendations and let me just discuss before we head into those. Karen, you had made the suggestion that those three be presented all together, however, we've gone through more expeditiously with our time, very good. So who will be our first presenter of the three?
PARTICIPANT: We have NIBIB, which correct me, National Institute of Biomedical Imaging and Bioengineering?
MR. BLAIR: Yep.
PARTICIPANT: And it's Dr. William Heetderks and Dr. Richard Swaja.
MR. BLAIR: And this is the media --
PARTICIPANT: Multimedia workgroup.
Agenda Item: CHI Preliminary Reports - Multimedia - Dr. Heetderks
DR. HEETDERKS: I will just walk you through briefly the preliminary work that we've done on this, first focusing on the scope and in particular domains and sub-domains. Our overall domain is the incorporation of a broad variety of multimedia information into patient records and the sub-domains that we identified were images, audio, critical care patient monitoring data, validated instruments, and by that we meant things like neuropsychological tests where one gets a collection of data, which taken together represents a validated measure of some patient information, and video information.
In terms of the alternatives that we're looking at, and we're actually here hopefully to get some additional suggestions if you have any, the ones that we identified early on were SNOMED, which was the required one, DIG35, I'm revealing my level of sophistication in this one, I'm not sure, I think it's actually DIG35, but that's actually a set of standards that's metadata standards being developed by the International Imaging Industry Association, I3A, and that's multimedia format that we're looking at. DICOM, which is probably best known as sort of the emerging standard for medical image information, especially among radiologists. And IEEE 1073, which is IEEE, the Institute of Electrical and Electronics Engineers, they are developing a set of standards which is really focused on data streams, in particular coming from various instruments and intensive care monitoring.
So those were the standards that we were looking at and additionally we're talking with a couple of organizations, NIST, and in fact we went to one of their workshops earlier in November on biomedical data integration, and RSNA, the Radiological Society of North America. They in addition to having a strong interest in DICOM also have this activity called, let me get it straight, the IHE, the Integrated Healthcare Environment, which is not really a set of standards but is looking at how standards if you will interact with a variety of other systems, so to be integrated into a health care system, so we've been talking to them initially.
And in terms of our initial thoughts, I don't think I need to go through those, but we just have some initial thoughts on those topics and I think I'll quit there and leave this open for questions at this time.
MR. BLAIR: Other questions, first?
DR. HUFF: From the description I'm not sure I understand the scope, in other words could you say just a few names of things that you would expect to see in terminology to cover this area? I mean is it the names of x-ray modalities or is it the names of the image series itself as you would find in DICOM? It's just not clear.
PARTICIPANT: [Inaudible.]
DR. HUFF: Oh, this is the standard.
DR. ZUBELDIA: This is not just terminology, right?
DR. HUFF: Maybe I'm more confused then I thought.
DR. HEETDERKS: Well, or it's also possible that we are, but I think that what we were taking as this was if you will a vocabulary that was agreed upon to incorporate this kind of information into patient records, so in a sense we're sort of interpreting this as a broader use of the word vocabulary if you will. So we were sort of picturing DICOM as a standard vocabulary for exchanging image information between records between different sources, so it's a vocabulary in the sense that, and obviously in the case of images so that in some sense one machine can understand another machine as well as one human understanding another human, so I think that that's a significant part of how we were interpreting vocabulary here. Now if that's not what's being looked for then --
DR. HUFF: No, I'm just trying, CHI is standardization in all kinds of things and so I guess DICOM was previously accepted as a standard right?
MS. WARK: As a messaging standard.
DR. HUFF: As a messaging standard, and so this is specifically looking at terminology that would be used in DICOM or in, I'm not sure I understand the context, or you've mentioned that it's in the patient electronic record but for instance, there are, Dean Bitgood(?) who helped create the DICOM standard also worked with LOINC to create 5,000 names for specific radiology procedures. Is that the scope of this work or is that something different?
DR. HEETDERKS: Dick may want to add to this, too, but my understanding was that the scope of this was if you will a vocabulary to actually incorporate the images, that is to incorporate multimedia information into the system as opposed to incorporating if you will vocabulary. So that's how we have been interpreting our scope is actually not integrating the words but integrating the information itself into --
MR. BLAIR: I struggled with the same difficulty that Stan is struggling with and I sort of feel like CHI can organize its standardization effort any way that it wants, but it looks to me as if most of your standard work activities were focused on medical domains and information content, this one seems to be multimedia, which is a technology, and it's a different paradigm or perspective.
I would just simply say that it's a way to do it that I feel uncomfortable with and the reason that I feel uncomfortable with it partially reflects on some of my comments earlier when I talked about disabilities as well, but I think Stan's example of radiological image is also applies. If you're going to look at radiology there could be vocabularies, there could be imaging specifications, I wound up trying to indicate that if you take a look at a particular medical domain in some cases you need a nomenclature, in some cases you need a message format standard, in some cases you need some way to represent questions and answers, in some cases you may need a standard for video, such as with physical disabilities. In some cases you may need a standard for images, such as fields of vision, and I certainly would feel much more comfortable if the terminologies and the standards were grouped according to the medical area rather then the technology.
So I've kind of said my opinion and I really don't know whether the rest of the subcommittee would agree with me, to make that into a recommendation or not, but I'm voicing that because to me there's a consistency and you avoid, you could avoid a lot of problems at least if you're consistent in the way that you define the groups. Do you have a comment or thought on that? Is there other reasons why you thought of what I said and you felt that it just didn't work somehow? Other considerations?
DR. SWAJA: We think the sub-domain is based on technologies because the variety of the technologies that have to be combined. We didn't focus, we didn't even consider the fact that within imaging domains, within imaging capabilities that there would be a variety of types of images that would have to be combined, but the point is put them all into one record that would be accessible throughout the patient's lifetime or maybe even available for purposes, access purposes and maybe possibly population research purposes in the future. That was why we approached it from the technology standpoint as opposed from the sub-domains of images, sub-domains of video, sub-domains of audio --
MR. BLAIR: Since this is preliminary what if you pursue just exactly the perspective you have but when you look at imaging, for example, you include both the needs of imaging from a radiological standpoint, maybe from a documentation standpoint, maybe from a visual field standpoint. If you made the effort to consider all of the different medical domains that require imaging so that you don't let things fall through the cracks when you do this then you get the best of both worlds.
DR. HEETDERKS: Can you expand on that because I'm trying to picture, and it may be --
MR. BLAIR: Well, I saw the value of what you're saying, you looked upon it from a technological standpoint and you said if I have a single imaging standard for use in a patient record, and that makes sense, so I was simply accepting the basic premise for why you're attacking this from a technological standpoint and then I was trying to marry it as well to saying okay, let's try to look at a particular standard, technology standard for imaging, and then wind up saying given that then let's make sure that we're also trying to see that all of the different medical domains that have imaging requirements are over time migrating or evolving to that standard, whether it be an ophthalmology field of vision, whether it be, and I don't have a clinical background to mention all of them, whether radiological images, whatever. So I was thinking if you do that then you'd have the best of both worlds and you wouldn't have things fall through the gaps, or at least you'd be able to identify what gaps there are.
DR. HEETDERKS: I think I see where you're going, I'm sorry I'm sort of slow on picking up on that. I think that yes, I would very much agree that that's certainly where we were thinking also, that in essence all of these, and in fact we were just coming back from the radiology meeting and it's certainly true that a field like radiology cuts across all of the other traditional areas of medicine and --
MR. BLAIR: Back up, the field of imaging. Imaging does, radiology is one example, representing from an ophthalmologist standpoint, field of vision is another image, and I'm sure there's many other areas where you have images, probably dermatology as well, so, yeah. Marjorie?
MS. GREENBERG: I was a little mislead when I saw validated instruments but that's only if they include several media, questionnaires wouldn't fit in there, they have to include images or photos or something of that type.
DR. SWAJA: Once again, addressing your comment, the point of the initiative is combining the data from several media, not standards specific to the media that has generated the data to be combined, so we're addressing the standard of how to combine all the information as opposed to the standards within the various domains.
MR. BLAIR: -- all domains correspond to the standard that you feel should be the overall technical standard for imaging or for video or for sound, if you don't look at those you may wind up selecting a technical standard that only addresses the minority of the clinical domains instead of something that is a good sound basis for commonalty because it addresses most of them.
DR. ZUBELDIA: I'm still concerned about, I'm confused, sorry, about the scope of the team. Is the scope the electronic exchange of images that can be stored and exchanged among the different domains? Or is the scope a terminology that will be supporting that electronic exchange that would happen in parallel with the terminology?
DR. SWAJA: I would say the second.
MR. BLAIR: Carol?
DR. BICKFORD: Carol Bickford, American Nurses Association. I have two questions, one is is the imaging associated with the genomics component of our health care delivery system encompassed in this? Particularly in light as we're looking at the construction of biomedical and genetic medications or interventions or whatever. And the second item I ask about is in my specialty for nursing as we're looking at our patient care delivery process there's a new approach in defining our care delivery initiatives called concept mapping, which includes imaging as well as some of our traditional care delivery thinking of a diagnosis intervention and looking at outcome sorts of things, and so it includes an image content. Does this fit in this area and would you be looking at that?
DR. SWAJA: Well, the overall scope is patient records. We haven't, beyond the scope of this is, or outside the scope of it is anything that would have to do with research, research data for example, and so as far as you have data and images at the genomic level that would be classified as patient data, yes, this would encompass that. And the second question, I don't really remember what that was.
DR. BICKFORD: I was asking about, in nursing we have our care plan, are care therapies for the scenario, like critical staff would be another example for physicians. And some of that now includes concept mapping which has an imaging component where you're actually representing a graphic as well as --
The second portion was in relation to what we're doing in nursing as we're looking at new ways to represent our care delivery process and our diagnosis and our interventions and our outcomes looking at concept mapping, which includes a graphic representation of those thoughts, those strategies, and that's an important part of our patient care documentation. In the past we've had outlines and text based, now we're looking at images and graphics to support that. It might hold true in the critical pathways as you're looking at representations of that in a graphic format, so I was asking if that's included in this work. It is related to patient care, it is included in patient records, it's serves as documentation of your decision making process, it represents the patient's input in some of these settings.
DR. HEETDERKS: I'm going to struggle with an answer to that question a little bit because I would guess that that could be incorporated but I don't quite, I'm not sure that I necessarily see how that would fit as multimedia data, because I don't quite, and perhaps it's because I'm not quite following how the images are represented but I think certainly if there is graphical information that would need to be represented in the chart I think that absolutely would be incorporated. And if there's icon kind of information, and that's where I'm confused a little bit about what you're referring to, I think that might or might not fit within this domain.
MR. BLAIR: I was going to kind of come back to you because I felt like I kind of interrupted you when you were winding up indicating how you were initially confused and I wondered if you wanted to follow-up or expand upon your comment and question.
DR. HUFF: Yes. So from the discussion that we've had, you're looking at standards that would allow you to incorporate multimedia data and information into an electronic health record. Is that correct?
DR. HEETDERKS: That's right. So if that's not what we're supposed to be doing we need to be hearing about it.
DR. HUFF: Oh, no, you can do anything you want, I'm just trying to understand what it is. So in that, given that use case then the question that I, so related to that activity there are lots of things that jump into my mind and I don't know how to relate it based on how you've laid it out here, so let me just explain and then you tell me what you're doing.
If I incorporate that image there are things that I want to know about it in terms of its format of storage, whether it's JPEG(?) or TIF(?) or that sort of stuff. There are things that I want to know, I mean in some sense that's a very limited but standard terminology. There are all of the things that are related to mime representation as to whether this, there's essentially the mime type and the mime sub-types, which correspond to whether this is an image of sound byte, and then for each of those whether, if it is an image whether it's TIF, so there's terminology related to just the physical representation of that.
Then I can imagine that there's terminology related to what is this thing, what was ordered in order for this to be created, is this a CT scan of the abdomen, is this four views of an ankle, what is this. There would also be terminology related to what was seen in the image, that is terminology related to what I can see in this image is a fracture or I can see a lesion in the left bronchi or I can see a maculopapular rash where I can see, so there's terminology related to the content and then there's all of the complexity of course that comes from DICOM about regions of interest and what, which are essentially overlays on the image that you might want to represent in here and the region of image is not terminology related but has more to do with coordinates. But within that region of image then you have all of the sorts of things that essentially within this region of image there is a particular lesion or thing that you're trying to focus on.
And so I guess what I'm, at the heart of my question is are all of those things within your scope? Is there a particular one of those that are in your scope? And if they're all in your scope then it would be useful I think to break this out and say these are the candidate terminologies for this use, for describing the physical format of these, these are the terminologies that would be useful describing the lesions that are seen by these images, this is the terminology that would be useful for ordering these things. There are even more complicated things that would have to do with collections, this might be a collection of things related to, because part of this for instance would seem to me would get close to claims attachment because you might request in fact that this, an image be sent as part of adjudication of a claim because you wanted to see the x-ray that showed the fracture or the lesion, that sort of thing.
Could you clarify using sort of my own paradigm about this, can you clarify what you see as the scope of your work here?
DR. SWAJA: Well, you just said a lot.
DR. HEETDERKS: Let me take a stab at it at least. I think that we saw as the scope essentially what you have just described with the possible, or with the probable exception of the, if you will, the domain of diagnoses or identified --
[Phone interruption.]
DR. HEETDERKS: So in terms of vocabularies, I mean I think that there would be, if you will, fields where that information would be contained within the multimedia record that is getting into incorporated but I don't think that we had necessarily been thinking about the specific vocabularies for the content of that field. I think in terms of the domain that we were looking at I think that it was more focused on how do you represent this information, so to try and say this a little differently, I think where we were looking was everything up to the final point that you got to which is okay, let's just take an example, for example, of an x-ray of a bone as a simple example --
MR. BLAIR: Could I interrupt for a sec? What you're doing is fine and you're going down the path of trying to resolve the issue, and I think Karen has a comment and an observation which might be appropriate given the direction of our conversation and questions so far.
MS. TRUDEL: It seems as though very many of the comments and questions have to do with the scope and what's in and what's out, and rather then maybe try to grapple with some of that on the spot perhaps if you could take that back to your workgroup and possibly if there is any assistance that the CHI Council can provide to you in terms of things that we've already addressed in other domains, where we could kind of help you work that out, we'd be glad to do that. But I think most of the input that we're getting from this process has to do with the fact that there's some uncertainty about the scope and what's in and what's out.
DR. HUFF: And that's a good comment but I'd still like to hear you complete what you saw as the scope of what's in and what's out.
DR. HEETDERKS: So we were approaching this more from the point of view of what is in is the multimedia and going back to the earlier discussion, especially the vocabulary to exchange the multimedia information, we were conceiving as part of multimedia there being if you will descriptors of what is represented in this multimedia be it related to the eye, related to the skin, related to the x-rays, but we were not focusing on what that vocabulary would be. We were I think presuming that the vocabulary that would be represented within that domain would be represented by another group.
DR. HUFF: So just see if I understood what you said then, what you're really focused on is the terminology that would all you to know what this multimedia thing is in terms of whether it's an image and that it's a JPEG image and that part of it and not focus so much on whether this is a CT of the abdomen, or both of those, or neither of those.
DR. HEETDERKS: Both of those and --
DR. HUFF: But not necessarily the diagnostic part.
DR. HEETDERKS: Right, and would have a way of associating with that image, the radiologist report of that image, and if you will that there would be a way of putting all of those pieces together in an agreed upon fashion that could be communicated widely. But I think what would not, what I would say is not part of it would be, if you will, a description of what are the vocabularies of all of the --
DR. HUFF: What you could see inside there.
DR. HEETDERKS: All the things that you could see.
MR. BLAIR: Other comments or questions?
DR. FITZMAURICE: In looking through the technologies and looking through the domains of faxes and material that could be incorporated in electronic health record, I notice that you don't have an example of imaging software such as Adobe Acrobat and PDF files, yet as I search through there if you have something come up you'll need to describe what is in such a file. Is that not medical, not clinical, or outside the scope of your work?
DR. SWAJA: No, the things that you just mentioned, PDF, for example, HTML, XML, those kinds of things, yes, they are part of this and they would be considered. We've concentrated here, the comments are focused on images, but those types of information, the computer generated data, the digital data, is all part of this also.
MR. BLAIR: Steve, have you begun to craft a phrase?
DR. STEINDEL: There's nothing that's required for it. We don't comment on --
MR. BLAIR: Well, if there are no other comments or questions then I think we could go on to our next category of preliminary, I can't even say it's preliminary recommendations, it's preliminary information about our next topic, population health.
Agenda Item: CHI Preliminary Reports - Population Health - Dr. Steindel
DR. STEINDEL: Thank you, Jeff. I'd like to start my comments on the population health domain by repeating the comment I just made, we don't report anything on preliminary reports. That's putting on my NCVHS hat and taking off my CHI hat.
The population health domain is a domain that started back last spring but was put off several times, basically because we thought we could get other people to do the work in front of us. Unfortunately late this fall we found out that we probably couldn't, so the workgroup has met, it's looked at the scope of its domain, it's decided to crawl into a rabbit hole and hide. So that is our recommendation.
Now to be a little bit more serious, we've looked at our domain as public health reporting, population health statistics, administrative data statistics, and institutional health statistics. Of those they are the first two, public health reporting and population health statistics, we've preliminarily reviewed as being in scope. We've viewed administrative data statistics as being out of scope. We have noted that institutional health statistics are both in and out of scope, we consider them to be out of scope in the sense that we're not going to make recommendations on what an institution should do but we note that it's probably a good idea for them to do what we're doing outside of the institution as well, but that's not a firm recommendation.
The approach that we have taken to this domain is that we have identified those standard terminologies that are currently being used for population health reporting, we did this by asking the members of the workgroup, which constitute I'd say about a third of the HHS agencies, which are the major agencies that are involved in population health reporting, we do have a representative from DOD on it, Nancy Orvis(?), as a commenter. But we've asked those agencies to just give us a list of the types of population health reports they produce and within that list the standard terminologies that are used in those reports. We note that that list is extremely non-exclusive, we made no attempt to find every population health report that's being used in HHS and definitely not in terms of the country, or even all federal agencies. But in that list we have given in the preliminary worksheet the standard terminologies we have found that are being used. They include ICD-10, ICD-9, ICD-9-CM, MedDRA, CPT-IV, LOINC, SNOMED, COSTART, DRGs, DMS-4, HCPCS, the Eindhoven Classification-Medical Model, HL7 Terminologies, HL7 Vaccine List, ICD-10 E-Codes, ILD Classification, the NAACCR terminologies, NDC, RxNorm, Units of various types and the VAERS codes for Vaccine Adverse Event Reporting. So you can see this is an extremely diverse set of terminologies and we do not feel that we've hit every one of them.
Our preliminary thoughts on this is the diversity of terminology needs found in the workgroup's investigation of population health reporting is leading to the conclusion that a specific CHI recommendation is inappropriate at this time.
I'm opening the floor to questions and comments.
DR. FITZMAURICE: A question, Steve, is there something that has to happen before a recommendation would be appropriate? Or is the scope so large that you think maybe it ought to be cut into segments and bitten off like you'd eat an elephant, a bite at a time?
DR. STEINDEL: I don't know how we can approach it because actually we do not know the scope of the domain, and one of the thoughts that the workgroup is proposing is that a study be done, that we find out the scope of the domain, and then maybe we can answer the question that you just asked. And if we can't recommend specific terminologies perhaps an investigation into how to do some type of mapping or similar type of relationship between clinical terminologies and the present population health reporting terminologies. And also when we look at this from a population health point of view we've also identified a need that has not been previously discussed with respect to mapping and that is the requirement to relate new data to old data, and that's a new aspect.
DR. FITZMAURICE: I can see the problem and I can see why you'd want to know the whole scope because in a lot of the standards that have been proposed in the other domains you probably have a lot of overlap, it's just that it's hard to define the overlap until you can get a handle on the domain.
MR. BLAIR: Carol?
DR. BICKFORD: Carol Bickford, American Nurses Association. In light of the previous presentation on imaging has the population health group talked about the GIF representation of data or are you strictly looking at alphanumeric text?
DR. STEINDEL: I do believe that when we do a complete list we will find needs for GIF. In the preliminary list we did it did not come up, but that does not mean it's non-existent.
MS. LESH: Kathy Lesh, the Kever Company. Steve, I know you're aware of this study out of CDC that looked at a concept coverage study of your website at CDC and found out that there was a 50 percent gap in public health concepts in the standardized vocabularies in the UMLS, so I think that would support that you really need to go back to the drawing board and find out what the scope is and where, the vocabulary doesn't exist at this point in time in public health and most of that would be encompassed in the population.
MR. BLAIR: Marjorie, do you have any suggestions or thoughts on terminology standards for this area?
MS. GREENBERG: Well, I'm actually on this workgroup --
DR. STEINDEL: That will allow you to remain on the workgroup.
MR. BLAIR: Oh, I'm sorry, I didn't really know.
MS. GREENBERG: And I guess this is preliminary because we're taking something back to --
MR. BLAIR: I'm sorry, this is conflict of interest and you'll have to recuse yourself? That was intended to be a joke, please Marjorie.
MS. GREENBERG: I think we are going to have some recommendations on next steps and studies or investigations that need to be done, so I think that's, we're not just throwing up our hands but as Steve said this is a very broad area and as was suggested by the other speaker, too, that we know that there are a lot of gaps in the vocabulary and the terminologies that cover this area.
I wanted to ask Steve something because I guess hadn't, sometimes when you see a whole report and then you see the condensed version something strikes you but didn't before, and that was that administrative data and statistics are shown as out of scope, I think billing data obviously is out of scope, for example, some ways like vital registration data are seen as administrative data and I don't think they're really out of scope.
DR. STEINDEL: I totally agree with you, they're not out of scope, it's just that we were, I think we were thinking about administrative data in the sense of billing data, and we consider obviously vital statistics one of the most important parts of population health statistics.
MS. GREENBERG: Because there are a number of populations statistics that are --
DR. STEINDEL: Are derived from administrative systems --
MS. GREENBERG: -- administrative systems that aren't necessarily billing.
DR. STEINDEL: Right.
MS. GREENBERG: So we might want to kind of clarify that.
DR. HUFF: I mean at the level I'm just sort of thinking again about my day job at IHC and sending this kind of information, exchanging this kind of information, it would seem like one area that would be very useful would be standardization of the names of population, so I mean I know the name, people are working on the names of codes for all the diagnoses that can be in all of the serum sodiums and measurements and findings and stuff like that. And in all that you're doing those names are going to be similar and I guess what I would see as being very useful is to now if I were going to report that information on a population how do I do that? One way to do it is you take all of the things that exist in ICD-9, you say this the sum of people who had that in my population, and so you would end up basically replicating all of the ICD-9 codes with a code that said this is the count of people with that diagnosis or something like that.
But I suspect what, I mean the focus that would be very useful is to say how do you do this population data, and it might be that what you're really doing is deciding the structure that will allow you to say what was the, you're making a small information model that says what was it I found and how many people had it or whatever the descriptive characteristics are that need to be taken together to describe what it is you're reporting. But I don't know if that's something you want to do or not. Again, at a very superficial level not being an expert at all in these population things I would have said oh, well, HEDIS should be here because HEDIS is reporting population statistics about how many people had mammography, how many people had something, and those HEDIS codes to me would have been something that were not being covered in a comprehensive sense by SNOMED or LOINC or by ICD-9 codes because it's the count of people who had a particular thing, not the name of the particular thing that this patient had. Is any of that relevant to what you're, to the scope of what --
MS. GREENBERG: It this a question and answer issue again?
DR. HUFF: No, I don't think so, it's just what, it is a domain question again. I assume that the goal of this was to be able to transmit in an unambiguous way population data so there are two people aggregating data and one agency wanted to know the population statistics, and is very similar to reporting on clinical trial data, I mean that's one of the things that's different about clinical trials then regular patient data, is that what I'm communicating actually how many people had this complication that were taking this drug, how many people had, and it seems that is a bounded problem where you could make, you could form a strategy about how you wanted to do that or whether it's too early and we just can't do that yet.
DR. STEINDEL: I think, Stan, the workgroup is looking at this as more as how do we derive population health data from clinical data.
DR. HUFF: So it's more process oriented.
DR. STEINDEL: Yes, is there a terminology and that was one reason why we put this off for a while was because we wanted to see what the recommendations from the other workgroups would be so for instance for diagnosis we could point to the diagnosis recommendation, and I was going to be sneaky as the population workgroup chair and just say ah, for diagnosis we can just use theirs. It turns out as we looked at it and we looked at the problem more and more we realized that was very naïve.
MS. GREENBERG: In some cases we can do that.
DR. STEINDEL: In some cases, and we do note that there are cases in reporting that it is specific, it's a one to one relationship and usually that's fairly clear. But in a lot of other cases like for instance the relationship between ICD-10 codes and clinical codes. In some cases it's a one to one relationship, in some cases it's a relatively simple map two codes combined, in other cases it can be quite convoluted.
MS. GREENBERG: Another issue, too, that I guess is in the report, but I mean it definitely requires some thought is there are cases where clinical data could be reported but you might have some people then reporting using say SNOMED, but other people because they are using SNOMED, but then other people like on causes of death or something, but other people are just reporting the terms that they use, not these standardized terms, but terms. And then do you end up, and then do you put both of them into the automated system to then create the ICD-10 codes, which is how we do most of it now? If they're different, data is coming from a different way, like coming at the term level or coming at a standardized term level or at a non-standardized term level, or maybe just at the classification level, and then do you have comparable data. I mean there are questions here so these are all ones that need to be explored.
MR. BLAIR: Any other questions or comments? Okay, we could go to the last document and get a chemical reaction from this, Steve, maybe you could take us to our next one.
DR. STEINDEL: I was just noting, why did we leave chemicals for last? I'm wondering if that's what we need at this time, at least caffeine.
Agenda Item: CHI Preliminary Reports - Chemicals - Dr. Steindel
DR. STEINDEL: The preliminary report is on the chemicals workgroup. The chemicals workgroup was assigned to identify a terminology that can be used for chemicals other then drugs. If you'll recall the FDA terminology is being used to identify those chemicals that are being used for drugs, so we have the world of chemicals that are non-drug that as the workgroup discussed this really do not appear that often in medical records, but they occasionally do, mostly associated with environmental conditions or poisoning, toxic events, but generally do not appear that much in medical reports. But this was a domain that we were charged to look at, we did look at this domain, we have made two conclusions with respect to it.
The first is the chemical abstract service number is essentially the number used in the United States and perhaps the world to identify a chemical. There are approximately 22 million CAS numbers assigned. The problem with the CAS number is that they have restrictions on its use for non-regulatory purposes. The CAS number can be freely used for regulatory purposes, we are viewing medical exchange of information as non-regulatory, and consequently the CAS number is ruled out. The intellectual property considerations are vast and the National Library a number of years ago tried to negotiate an agreement to use CAS numbers more freely and have advised us that it was very difficult to get that negotiation going, so we have excluded the CAS number from consideration.
There are two other terminologies we are required to look at SNOMED and the other terminology that commonly has chemicals is MESH, they both have approximately just under 20,000 chemicals listed, a large number of them are drugs, but they still have a large number that are not drugs. You have noted that I said there's 22 million chemicals identified by CAS so you can see that there's a large number that are not identified in SNOMED and in MESH, and our reaction to the chemicals that are listed in both SNOMED and MESH is that both of those terminologies tend to be reactionary, and by that I mean when a chemical has been shown to have importance in medicine it appears in one of those terminologies and we felt it was important to select a terminology that was proactive, that the chemical would already be listed so that if the person was doing messaging they would not have to have it added, and there are numerous terminologies that do have a more exhaustive list. As I note there are numerous federal and private lists that do have chemicals.
The workgroup is focusing on a list that's provided by the Environmental Protection Agency that's used in the regulatory realm and has about 80,000 chemicals in it, we are investigating its potential use for medical purposes.
And that is where we're standing right now, I'll take any questions.
DR. BICKFORD: Carol Bickford, American Nurses Association. Does chemicals include some of the complementary and alternative therapies? For example herbs and botanicals and some of those concoctions?
DR. STEINDEL: To a certain extent, the only thing I can say to that is it's a definite maybe. We're looking at chemicals as the actual chemical and in many cases these herbs and botanicals are associated with specific chemicals and in that case they would be listed. But if it's basically a preparation or something like that mostly likely it would not appear in the list.
I have to note that there might be in phase two of CHI kind of a bit of a creep in this domain because some of those do start appearing for medical purposes when they are found to have toxic events associated with them. So in that case they might need to appear.
MR. BLAIR: Other comments or questions?
MS. LESH: Kathy Lesh, the Kever Company, I have a question about the MESH, were you looking at the supplemented chemical list or just the straight MESH list?
DR. STEINDEL: The straight MESH list.
MS. LESH: You're aware of the supplemental chemical list?
DR. STEINDEL: Yes.
MS. LESH: Because any time it shows up even once in the literature --
DR. STEINDEL: But that gets to the point what we feel it is reactive and not proactive.
MR. BLAIR: Other comments or questions? Well, I'm not going to belabor or linger on. WE have completed all of these, is there any comments that anybody would like to make before we adjourn?
DR. STEINDEL: No, we don't have adjournment, we're an hour early for the next part of the agenda, which is supposedly the draft CHI letter.
MR. BLAIR: In which case we could wind up probably taking a break before we do that I would think and wait for Simon to return.
DR. STEINDEL: Simon is supposed to be back at 4:30.
MR. BLAIR: I think that's what, ten or 15 minutes from now? Okay, any other comments on what we've gone through before we break?
DR. STEINDEL: The next meeting site.
MR. BLAIR: Okay, well thank you every one and let's go ahead and break until 4:30 and this has been a marathon, I can't believe we ate the whole thing.
[Brief break.]
Agenda Item: Draft CHI Letter - Subcommittee
DR. COHN: -- probably just as well that we handle them, actually do we have one, two --
MR. BLAIR: We did our voting.
DR. COHN: Just the wordsmithing --
DR. STEINDEL: And I gather we're going to not submit it until after the January session, so this is just very preliminary wordsmithing.
DR. COHN: Exactly. I think we'll probably want to sort of see where we are and it's probably something we'll want to send out to the full committee as we prepare for our full committee meeting, so at least I'll have this as background.
MR. BLAIR: Steve, are you aware that there were two finals that Simon was not here for?
DR. COHN: I voted. I think we sort of need to look through and see where we are here. Steve, do you want to read?
DR. STEINDEL: The first paragraph is our standard paragraph, do I need to read that?
SUBCOMMITTEE: No.
DR. STEINDEL: Thank you. Now I do have a problem with the second, the next paragraph, the NCVHS has the following comments on the attached set of CHI domain area recommendations. The problem that I had with that was we actually only modified two of the recommendations and I thought I could get away with only attaching two, but the way I read it is I needed to attach them all. But I think we actually have to.
DR. COHN: I think we need to, and I also think it will be important --
DR. STEINDEL: I agree with that --
MS. GREENBERG: -- then people have access --
DR. STEINDEL: This was basically my laziness, but it didn't work, so they are all attached.
DR. FITZMAURICE: I noticed that all of them start with NCVHS concurs with the --
DR. STEINDEL: That was the style we used before and it really worked when we had like two or three but now we have close to, it will be about 12 --
DR. COHN: Actually, just wordsmithing if you basically put a period at the end of the sentence where the NCVHS has the following comments on the attached set of CHI domain area recommendations, period. And then it say the NCVHS: --
DR. STEINDEL: Yeah.
DR. COHN: Do we all, are we concurring with all of them? If we're concurring with all of them --
DR. STEINDEL: We do concur with each one of them but I agree with the style. The first one is with the anatomy and physiology domain as modified, and then we note the need to revisit the physiology domain in 12 to 18 months and begin with an evaluation of the domain definition.
DR. COHN: And actually and that begin I think.
DR. FITZMAURICE: How about instead of we note that just say we also concur?
MR. BLAIR: It's there, isn't it? Oh, you're saying in the next sentence, we also concur with. I see.
PARTICIPANT: And that this process begins, or this examination --
DR. COHN: Oh, you mean that process begins --
PARTICIPANT: Examination, I have --
MS. GREENBERG: When you say as modified then you will attach the --
DR. STEINDEL: Yeah, there's an attached report with the modifications.
MS. GREENBERG: Which is the CHI report --
DR. STEINDEL: In mark-up format.
MS. GREENBERG: Marked up. And then does it go back to the CHI Council? How does that work? Does the report get modified --
DR. STEINDEL: I actually think that's probably in the domain of the CHI people, in this particular case the modification was simply change of the date.
MS. TRUDEL: I think if they're minor changes we can at least let the council know what the changes were and find out whether they feel the need to re-ballot, which for the most part I don't think they will.
DR. STEINDEL: So we have this, do people agree now with the wording of the second sentence as its up there? We also concur with the need to revisit the physiology domain in 12 to 18 months and this process begin with an evaluation of the domain definition.
DR. FITZMAURICE: How about and that this process begins --
MS. GREENBERG: Beginning with an evaluation, that's good.
DR. COHN: And maybe it's beginning with a reevaluation of the domain definition is really what we're talking about.
MS. GREENBERG: I don't think you need that and that, if you just put a comma after months and then say beginning with a reevaluation of the domain definition.
DR. STEINDEL: Is everybody happy with that? We also concur with the need to revisit the physiology domain in 12 to 18 months beginning with a reevaluation of the domain definition.
Okay, the next bullet is concurs with the CHI recommendation for the billing domain as modified and there are a couple of modifications in there, mostly having to do with identifiers. I can go to the report and specifically show them.
DR. FITZMAURICE: I'm thinking if the Secretary is specifically receiving this letter he's not going to know what those modifications are, might they be modified by the end of January so that the full committee can agree with it without the as modified?
MR. BLAIR: I think at the level of the Secretary my thought is that just the fact that there's concurrence is about all he cares about, I don't think he's concerned about the detail.
DR. FITZMAURICE: Or if he hands it to somebody they're going to say alright, what were the modifications.
MR. BLAIR: But there's an attachment there that shows the modifications.
DR. FITZMAURICE: Okay. Otherwise it's like it's hanging out there.
DR. STEINDEL: The first sentence says has the following comments on the attached set of CHI domain recommendations, and we're referring to the modification as appearing in the attachment.
DR. FITZMAURICE: So they will see the original modification and they'll see, they'll see the original and the modification of it?
DR. STEINDEL: It's done in mark-up format, word document.
DR. FITZMAURICE: I've never seen something go to the Secretary in mark-up format.
DR. STEINDEL: The last one did. Okay, the next one, concurs with the CHI recommendation to not adopt a terminology for the medical devices and supplies domain at this time. We recommend HHS further investigate device type as a component of device identification.
DR. COHN: I agree, I'm just trying to think of whether we need to say additionally we recommend, just to somehow distinguish it from the fact that it isn't really a CHI recommendation.
DR. FITZMAURICE: We independently recommend.
MR. BLAIR: We did something by ourselves.
DR. STEINDEL: Do we use HHS or DHHS?
In the next one, concurs with the CHI recommendation for the nursing domain as presented.
Concurs with the recommendation to defer the history and physical domain until the next phase of CHI.
Concurs with the recommendation to not adopt a terminology for the disability domain. We agree with the need for a future research agenda, concur with the basic list presented, and also recommend addition of investigating the question and answer style format for aspects of this domain. We request that future research be inclusive of all domain stakeholders.
MR. BLAIR: Two pieces, where you said to not adopt, I think you have to say not adopt it at this time, that's number, you see what I'm saying because otherwise it's almost as if cancel the domain and that's not what we're saying.
DR. FITZMAURICE: Maria, what about split infinitives? Are they acceptable these days? I'm thinking about not to adopt versus to not adopt.
MS. FRIEDMAN: I can live with it.
MR. BLAIR: To not adopt --
DR. STEINDEL: Or not to adopt, we'll make Mike happy.
MR. BLAIR: Could you read again where it said inclusive of?
DR. STEINDEL: We request that the future research be inclusive of all domain stakeholders.
MR. BLAIR: Is it a problem if we say inclusive of the differences among?
MS. GREENBERG: Well, if you include them all, all the stakeholders and their various --
PARTICIPANT: Does that mean you only want them to look at the difference or do you want to look at all of them and consider the difference?
MR. BLAIR: It's the latter, I want to look at all and consider the differences, and consider the differences.
DR. FITZMAURICE: Do you mean instead of investing you mean investigating?
DR. COHN: Somehow we need to say that a little better, that particular sentence.
DR. STEINDEL: Which one? The second sentence?
DR. COHN: Investing the question and answer style format for aspects of this domain, doesn't sound quite right.
DR. FITZMAURICE: I would take away addition and say we recommend investigation the question and answer style format, question style format used by such systems as SNOMED, I mean such systems as LOINC? That may be too specific.
DR. COHN: I'm just trying to think of what, the question and answer style format --
DR. FITZMAURICE: The use of, investigating the use of the question and answer style format.
DR. COHN: Okay. I guess I, I mean it's question and answer style format, should that be in parentheses? I mean that's not a usually way that, I've not seen --
DR. FITZMAURICE: I think that would be good in quotes.
DR. COHN: Good in quotes.
DR. FITZMAURICE: It exists, it's not just something we're saying off the cuff, it exists.
DR. COHN: I guess I'm just wondering, we could describe this as aspects of this domain, for parts of this domain, to solve --
DR. FITZMAURICE: For appropriate parts of this domain? Kind of weasel word the question away? Investigate the appropriate use of the question and answer style for this domain.
DR. COHN: -- aspects of this domain. Well, I guess we were primarily talking about the questionnaires, but we really didn't want to start tying their hands in terms of how they might apply it.
PARTICIPANT: [Inaudible.]
DR. COHN: No, actually I don't think it is in the modifications, I don't think we made modifications to this report.
DR. STEINDEL: No.
MS. GREENBERG: I think we need two sentences here.
DR. STEINDEL: We did not make modifications to this report.
MR. BLAIR: I think Marjorie is going down the right track, she's saying we need two sentences.
MS. GREENBERG: Say we agree with the need for a future research agenda and concur with the basic list presented, period. We also recommend that the workgroup, or CHI, investigate how --
MS. TRUDEL: And it's not the use, it's really the appropriate application.
DR. FITZMAURICE: That's a good word.
MS. GREENBERG: Or investigate how a question and answer style format could facilitate, could be used, to facilitate the capturing of information for this domain, something like that.
PARTICIPANT: I think that's good because that gives the flavor --
MS. GREENBERG: They're somewhere else over there, but definitely we need a second stand alone sentence about --
MS. TRUDEL: And then I think it ought to be in addition we request --
MS. GREENBERG: You could say investigating how the question, I wouldn't say style --
DR. FITZMAURICE: -- add specificity to this domain?
MS. TRUDEL: Because we're not saying that we're looking at whether questions and answers are good or not, we're saying that they're inevitable, we have to figure out how to make it work.
MS. GREENBERG: That was not actually something that was explicitly explored by the workgroup, now the committee is telling the workgroup to think about that.
MR. BLAIR: For my benefit when you are finished Steve could you read it?
DR. STEINDEL: I'm not finished because I haven't heard what to put down.
MS. GREENBERG: I think and facilitate capturing information for this domain, or something like that.
DR. FITZMAURICE: Facilitate exchanging information. They can facilitate exchanging information.
MS. GREENBERG: Well, it's even the capture, let alone the exchanging. Maybe the representation of information for this domain.
DR. COHN: [Inaudible.]
MS. GREENBERG: That leave it open as to whether it's the exchange of information, the collecting of it, the representation --
DR. STEINDEL: Okay, let me read it for Jeff then. We also recommend investigating how the question and answer style format can facilitate representation of information in this domain.
MR. BLAIR: And then the rest of it?
DR. STEINDEL: Is still we request that the future research be inclusive of all domain stakeholders, Marie just added that --
MS. FRIEDMAN: And consider their differences?
MR. BLAIR: Yes, thank you.
MS. FRIEDMAN: Or while considering their differences or --
MS. GREENBERG: -- different needs and perspectives?
MR. BLAIR: Yes. Thank you, very much.
MS. GREENBERG: I would say, how about the word request? It almost sounds like CHI resisted and they didn't. I would say recommend, we recommend that the future research and is it clear that that's also the investigation? Because now we've got research in one sentence and investigation in the other. Just say that the future work, research and investigation, I wouldn't want that to only refer back to the second sentence and not the third.
DR. COHN: Actually this isn't too bad.
DR. STEINDEL: Future studies?
MS. TRUDEL: I think we liked activities better then studies.
PARTICIPANT: It's still not clear to me coming from a research environment whether you mean research like you normally think of it in that sense or whether CHI --
DR. COHN: I guess, I'm looking at this one, I guess I'm just trying to think in my own mind what does, I mean I understand what the last part of the sentence means I just don't know what the first part means. What does be inclusive of all domain stakeholders mean exactly in this context?
DR. FITZMAURICE: -- federal stakeholders, those who use the terminology? The federal programs?
PARTICIPANT: Wasn't that only to include the ones that we don't normally think about like the environmental and the --
MR. BLAIR: Well, I think that when I was saying all stakeholders, the point I was trying to make when we added that was that it be the health care providers and therapists in addition to the needs of public health and research.
MS. GREENBERG: Well, it always was though, that's where I think it wasn't clear, we were never just focused on public health, in fact probably the primary focus is on --
DR. COHN: I guess I'm misunderstanding because I'm reading this almost as being, since it has to do with their different needs and perspectives I was presuming that we were talking about different people with different types of disabilities --
MR. BLAIR: Yes.
DR. COHN: And so I was thinking well geez, we'd want, I mean are we --
MR. BLAIR: That's correct.
MS. GREENBERG: As well as people who are involved, who work with people with different types of disabilities.
MR. BLAIR: And different uses of that information as well.
MS. GREENBERG: So it's the whole range of stakeholders for disability data.
DR. COHN: Okay, so what does inclusive mean then? I apologize, I don't mean to bang on this one, I'm just, I can just imagine a large auditorium filled with representatives of every sort of whatever and I'm just wondering do we really want that be inclusive of all domain stakeholders? Or --
MR. BLAIR: Do you want to say consider the needs of all stakeholders?
DR. COHN: Consider the needs of all domains?
PARTICIPANT: There are a lot of different stakeholders --
DR. COHN: Or do we want to talk about consider the needs of all domain stakeholders? I mean that makes a lot more sense to me then having to have 100 people on this committee, which I could imagine.
PARTICIPANT: [Inaudible.]
DR. COHN: Do you agree with that, Jeff?
MR. BLAIR: Yeah, the only piece I don't want to lose with that is the tack on phrase including their differences, so I don't want to lose that.
DR. COHN: Let's see what we've got here.
MS. GREENBERG: Well, you've already got inclusive, consider the different needs and perspectives of all domain stakeholders.
MR. BLAIR: Okay, when you say it that way, there you go, you've got it in there, consider the different needs and perspectives, that's, you've got it.
DR. STEINDEL: Okay, the way I have this worded right now is we further recommend that future activities consider the different needs and perspectives of all domain stakeholders.
MR. BLAIR: Excellent, thank you.
MS. GREENBERG: I think that's good.
DR. STEINDEL: Okay, should I reread the whole section or should we just move on to the next one?
DR. COHN: I think we can go on to the next one.
DR. STEINDEL: Concurs with the recommendation for the genes and protein domain as noted for the human genome, and note the need for explicit comment on the lack of terminology for the remaining sub-domains.
The next sentence is just in bold for me more or less that we've got to come back to it, we recognize that work is still underway on clarification of licensure terms for use outside of the federal government, and the notation to revisit in January for clarification.
DR. COHN: And I presume that first sentence is going to get cleaned up also.
DR. STEINDEL: I'm ready to clean it up. The first sentence I believe is the one that we should focus on.
MS. FRIEDMAN: You can take out that second noted, as noted for the human genome as well as the need for --
DR. COHN: That's actually a placeholder, we're not wordsmithing that sentence. I mean basically we need to list the other domains that are not, I mean basically we basically concur with the recommendation for the human genome and we basically also need to concur separately that there is no terminology for the other sub-domains I think is what --
DR. STEINDEL: No, actually what we're referring to here, and I don't agree with the way I just reworded it, what we were commenting on was that the document that we were presented with didn't really make an explicit statement that we are lacking terminology in the other sub-domains. And we were requested that an explicit statement to that effect appear.
What I've said as a separate sentence, we recommend that an explicit comment on the lack of terminology for the remaining sub-domains be added. And I would prefer that we don't try to do any crafting on the sentence in bold at this time.
DR. COHN: Okay, that sounds, we don't have the information.
DR. STEINDEL: What they may do is they may actually add the information to the report in January and we'll delete the sentence entirely. Cynthia noted that they were investigating it.
Concurs with the recommendation for the diagnosis and problem list domain and request the addition of ICPC to the recommended terminology for early mapping efforts.
DR. COHN: I think it's actually a period after domain. We further recommend the addition of ICPC to the recommended terminologies for early mapping, is there a list of --
DR. STEINDEL: Yeah, there was.
DR. COHN: Where?
PARTICIPANT: It wasn't meant, it was just an example of a list but they brought that up several times, it was just for example and then Steve asked that they specifically name it in that list.
DR. COHN: Well, how about if maybe we further recommend that ICPC be considered for early mapping or priority mapping to, I'm not sure, what is it being mapped to?
DR. STEINDEL: SNOMED.
DR. COHN: To SNOMED, okay. Further recommend --
MS. GREENBERG: I mean if there is a map between SNOMED and ICD-10, although it's not an international map, it's a UK map, and there's also ICPC has been mapped completed to ICD-10 as well. You've got some foundational work done there.
DR. STEINDEL: The last one is concurs with the recommendation with the non-laboratory interventions and procedure domain as presented.
And that's it.
DR. COHN: That's it. Okay. Are we comfortable with that?
MR. BLAIR: The dirty deed is done.
DR. COHN: A lot more before it's done. Okay, well Steve I want to thank you. So basically we are on time here, it's actually we're 15 minutes early, so tomorrow morning we start off at 8:30 and Maria I think you're on board and we'll hopefully be read to adjourn by 12:30.
[Whereupon the meeting was recessed at 5:15 p.m., to reconvene the following day, December 10, 2003.]