[This Transcript is Unedited]
Humphrey Building
200 Independence Avenue, SW
Washington, D.C.
Call to Order - Simon P. Cohn, M.D.; Jeffrey S. Blair
Testimony of Public Health, Hospital and Others About Elements of the Proposed Process
Helen Stevens
Sandy Fuller
George Arges
Review of Testimony: Next Steps - Subcommittee
DR. COHN: I want to welcome everyone. I want to call the meeting to order. This is the second of two days of hearings of the Subcommittee on Standards and Security of the National Committee on Vital and Health Statistics. The committee, as you all know, is the main advisory committee of the U.S. Department of Health and Human Service on national health information policy.
I am Simon Cohn, chairman of the subcommittee and I want to welcome fellow subcommittee members. We actually have a quorum this morning. So, thank you for coming. HHS staff and others are here in person.
I also want to tell people that unlike our usual situation, the Internet is down. So, this is actually being broadcast over the Internet.
[Multiple discussions.]
The Internet is never down. There is not a live broadcast. However, we are doing our regular tape of the presentations and testimony and this will be transcribed. Obviously, I want to remind everyone to speak clearly into the microphone.
With that, let's have introductions around the table. We checked with Kepa and apparently he was asleep when we last called.
We should do the introductions. I also want to have people publicly recuse yourselves if there are activities you have outside of the committee that are any way of a conflict, with the topics or issues that we are going to be discussing today.
With that, let me start. I am Dr. Simon Cohn. I am the national director for health information policy for Kaiser Permanente, a member of the committee. In terms of my activities, I am a member of the editorial panel, National Uniform Claims Committee and ANSI Health Informatics Standards Board.
Jeff.
MR. BLAIR: I am Jeff Blair, vice chair of the Subcommittee on Standards and Security. I am vice president of the Medical Records Institute. I am on the board of ANSI Care Informatics Standards Board. I am a member of HL7. I am a member of ASTM and I am present.
MR. SCANLON: I am Jim Scanlon from HHS, Office of the Assistant Secretary for Planning and Evaluation. I am the executive staff director for the committee.
DR. SHORTLIFFE: I am Ted Shortliffe. I am professor and chair of the Department of Medical Informatics at Columbia University, member of the committee and a new member of this subcommittee.
DR. COHN: Want to introduce yourself, Kepa.
DR. ZUBELDIA: I am Kepa Zubeldia, Claredi Corporation.
DR. COHN: Kepa, we are asking you also to publicly announce or recuse yourself if there is any matters coming before the committee today, which have a conflict.
MR. BLAIR: Any SDOs you are a member of.
DR. ZUBELDIA: [Comment on telephone.]
DR. FITZMAURICE: I am Michael Fitzmaurice, senior science advisor for information technology at the Agency for Health care Research and Quality. I am liaison to the National Committee and staff to the Subcommittee on Standards and Security. My agency supports the meetings of ANSI Informatics Standards Board.
DR. YASNOFF: I am Bill Yasnoff, associate director for science of the Public Health Practice Program Office at the Centers for Disease Control and Prevention, staff to the committee.
MS. STEVENS: I am Helen Stevens. I am the senior project manager with McKessonHBOC in the information technology business and I am a member of HL7.
MS. FULLER: I am Sandy Fuller. I am the senior vice president at the American Health Information Management Association.
MS. GREENBERG: I am Marjorie Greenberg from the National Center for Health Statistics, CDC and executive secretary to the committee.
MS. FYFFE: Kathleen Fyffe. I work for the Health Insurance Association of America. I am a member of the committee. I am also a representative of HIAA, the National Uniform Claim Committee and the National Uniform Billing Committee and the National Council for Prescription Drug Programs.
MS. WILLIAMSON: Michelle Williamson, National Center for Health Statistics.
MR. HUFF: Stan Huff with InterMountain Health Care in Salt Lake City and the University of Utah and the chair of HL7 and a member of ASTM.
MR. MEYER: Chuck Meyer, McKessonHBOC, member of the board of HL7, co-chair X12N Task Group, member ASTM, NCPDP and IAAA.
DR. EDINGER: Stan Edinger, Agency for Healthcare Research and Quality.
DR. COHN: Let me just take care of one or two items on housekeeping.
DR. ZUBELDIA: I can just barely hear.
DR. COHN: We are getting lots of feedback also.
Let's take care of one or two pieces of housekeeping before we get into the testimony. Yesterday, we talked about upcoming hearings and it was the decision of the subcommittee that we would cancel the next set of hearings, which were scheduled for May 1st and 2nd. We are going to be going forward with the hearings that are currently scheduled for May 31st and June 1st.
This has to do significantly with the fact that the designated standards maintenance organizations will not be ready until the end of May to present their recommendations regarding revisions to the standards, the final rules.
We will also be checking with the committee members for upcoming hearing dates for later in the year. I am sensitive that many of you have busy schedules. I am looking particularly at one person. So, we are going to be checking for probably a date in the last half in August and then another set of hearings in October. Maybe we also will try to reserve some time in December.
We may or may not use all of them, but it has become apparent that we had better get them on the books or else we are going to lose the ability of everybody to participate.
MS. GREENBERG: I just want to note that the Executive Subcommittee is meeting in probably the last two weeks of August. That is, I believe, in Chicago.
DR. COHN: I guess I would urge the Executive Subcommittee to identify those dates as quickly as possible.
MS. GREENBERG: We are polling right now.
DR. COHN: Yes, I know. Polling and announcing dates are sometimes different.
DR. ZUBELDIA: It just looks like there is a loose wire that cuts out and it is out now. I don't want to be disruptive. So, I may just hang up.
DR. COHN: Okay. Well, Kepa, I am sorry. You did better yesterday, but, obviously, it is not working very well today.
DR. ZUBELDIA: Now it is working. It looks like there is a loose wire somewhere.
DR. COHN: Well, you know, with the time zone changes, I have sometimes been described as a loose wire, but I don't think it is my fault this morning.
DR. ZUBELDIA: I can hear you now.
DR. COHN: Okay. Well, we will keep going and if it gets to be too much of a problem we can -- you know, just hang up at that point.
Anyway, with all of that, let's move into the hearings. Obviously, the focus of the hearings this morning is to identify next steps for the committee related to patient medical record information standards. I think we made very good progress yesterday and this is just continued discussions on that topic.
I think Jeff Blair, the subcommittee vice chair, Mike Fitzmaurice and Susie Burke-Bebee -- not well today -- basically for their help in putting these hearings together. Really with that, I want to turn the hearing over to Jeff for any introductory comments and to lead the session.
MR. BLAIR: Good morning everyone. I will be relying on Simon to help me as we go through this.
Kepa, are you able to hear me now? No.
Okay.
DR. ZUBELDIA: Yes, I can hear you now.
MR. BLAIR: For our testifiers, first of all, welcome, thank you for coming. Sit back and relax. Your role will be to critique our work. So, the questions that we will have for you as we go through things this morning will be to really understand your perspective, your insights and your wisdom with respect to the right way for us to do our job. Specifically, there were a couple of things that happened yesterday that was a little confusing to some of the testifiers. So, I want to try to put things in context and maybe that will help you as you begin to testify.
We on our Subcommittee on Standards and Security of NCVHS are under the requirement of one of the elements of the Health Insurance Portability and Accountability Act, which specified that we are to study uniform data standards for patient medical record information and the electronic exchange of that information and make recommendations to the Secretary of HHS by -- when you would up looking at the calendar, it turned out to be by August 17th of this last year.
That report was delivered. And in that report, we wound up instead of making recommendations for the adoption of specific PMRI standards, we pointed out a process and a framework, plus a number of other recommendations. We are proceeding now with this hearing right now on the first two recommendations.
The other recommendations are being pursued within Health and Human Services. I won't get into that. The two that we are focusing on are with respect to which specific PMRI standards we might be able to set forth for adoption within the health care community. You should not expect that this would necessarily be in the same fashion that the other financial administrative transactions had been set forth. It is not a given that we would go through an NPRM process. It is not a given that there would be a mandate.
We are trying to find out the best way to be a catalyst to move the ball forward with respect to the acceptance of standards. Now, the other pieces in that report set forth guiding principles for the selection of standards, PMRI standards. We have taken derivatives of those guiding principles and tried to make them pertinent to the very first step for the very first phase because we are envisioning that we may be going through several phases in making recommendations for PMRI standards.
The first phase we are focused on is some low hanging fruit, so to speak, which would be some of the message format standards. The criteria in the framework that was sent out to you gives you a draft of our thinking for what the criteria might be. It gives you a draft questionnaire that we might, if you feel that it is appropriate, that we might send out to SDOs in a month or two to be able to ask them for feedback on specific message format standards.
Also, an idea of the scope, which transactions we might solicit feedback on, plus open-ended questions at the end of that framework document. So, I would ask you as you begin to give the testimony to us, to please focus on the critiques of this process, these frameworks and the draft questionnaire. Maybe I will just wind up asking among you, which of you would prefer to testify first.
MS. FULLER: Well, Helen pointed at me. This is Sandy Fuller. So, I will go ahead.
MR. BLAIR: Thank you, Sandy.
MS. STEVENS: I thought I would be able to get away with pointing at her.
MS. FULLER: Dr. Cohn, members of the Standards and Security Subcommittee and guests, good morning.
I am Sandy Fuller, senior vice president of the American Health Information Management Association and on behalf of the association, its more than 40,000 members and myself, thank you for the opportunity to provide input into your efforts on advancing standards, which apply to patient medical record information. Some of you are very familiar with the AHIMA, I know, but for those of you who are not, I will give just a very brief overview.
AHIMA is a professional association representing more than 40,000 members, who on a daily basis managed health information and provider plan in government facilities throughout our nation's health care industry. Our members work with the challenges of managing information in today's health care environment that consists not only of paper, electronic records, data, multiple coding systems and numerous transaction requirements in the private and governmental sectors.
AHIMA has been an active participant in a number of the standards development organizations, including HL7 and ASTM. In addition, AHIMA has actively pursued initiatives to improve the quality of documentation and advance the electronic capture storage and dissemination of health record content.
In response to the current challenges in health information management, AHIMA is creating a new definition of the legal/medical record, a draft of which is attached. We are promulgating best practices in health information management, commissioning studies into the future of coding and its relationship to documentation and we have in conjunction with other groups begun the process of assuring that new and emerging health-related terminologies will address the demands of existing and future classification systems.
You will see examples of the involvement in my further testimony and in the attachments I have provided. This morning, of course, we are here to address the questions you have asked about standards for patient medical records, including the appropriateness of your criteria, any questions about the draft questionnaire, the proposed list of transactions to be considered and I will take the opportunity to make a few other comments.
First, the overall approach. Given the plethora of recordkeeping systems extant in the health care industry today, messaging standards are absolutely critical to the implementation of electronic patient records, whether it is deploying a comprehensive health care information system, creating a best of breed record solutions or coordinating care across multiple delivery systems, the transmission of information from one system to another is an absolute requirement in today's environment.
The development and deployment of standards already has decreased the number of individual interfaces, decreased the time of installation and improved the availability of data. We believe it is logical to assume that collecting data at its most logical point, then transmitting it as required to other systems not only improves the transmission and sharing of data, but the quality of that data as well.
The groups which have developed and promulgated these standards, ASTM, DICOM(?), HL7, IEEE and NCPDP, among others, have done a tremendous amount of work to advance the health care industry's ability to deal with the reality of the new electronic world of medical records and information. The subcommittee's approach to messaging as described should yield a description of the current state of standards development and deployment and the extent to which these standards developed in organizations are succeeding in their goal to create a standard; that is, to have one messaging protocol, which can be applied to a given set of data and used without required modification.
To the extent that the SDOs have or can acquire information about the implementation of their standards, they should be encouraged to address perceived barriers to adoption of their protocol. AHIMA believes that the creation or formalization of such a feedback loop between SDOs and vendors or other implementers may be an area NCVHS can affect in the future.
As Dr. McDonald noted in committee documents from last November, regulatory oversight of the application of standard messaging protocols will be extremely difficult. The questions that remain are what are reasonable sanctions for inefficiency. In most other areas of our economy, the marketplace has demanded standards that result in decreased cost and increased service.
We believe that rather than using penalties and sanctions as a motivating factor, that the National Committee on Vital and Health Statistics and the Secretary could better assist the promotion of standards by a number of activities, encouraging formal return on investment studies, promoting case studies and best practices, adopting official or approved standards for certain classes of messages, requiring that SDOs monitor implementation and update standards based on feedback and identifying overlapping standards and working to reduce competition between the SDOs.
Your next question addresses the criteria for selection of patient medical record information message format standards. Certainly the eight criteria -- and you know what they are -- are appropriate for consideration while developing standards for messaging. AHIMA respectfully suggests that the committee may want to consider establishing measurement standards or guidelines for each of those eight criteria, which summarizes the SDOs response to the criteria.
Examples might be do the standards address data quality? Do their standards promote interoperability? Do the standards require consistency with other standards? The committee may even want to ask the SDOs to self-grade their responses to the criteria and then use their detailed answers of the supporting questions to document their grade.
The next question addresses the questionnaire itself. Comments related to the draft questionnaire are probably best answered as you heard yesterday by the SDOs, which are expected to complete the proposed document. However, to AHIMA, the questions seem appropriate to the categories, even though we believe the information may eventually turn out to be difficult to assemble.
Again, by having the SDOs self-evaluate and complete the assessment grids the committee discussed in November, the amount of analysis interpretation required after the questionnaires are completed may be reduced. We realize that such an approach begs the question. Will the outcome of the process justify the work required the work required to complete the questionnaire.
That value proposition will undoubtedly come under scrutiny. You have asked us to comment on the proposed list of PMRI transactions to be considered in the HIPAA standardization recommendations. Given the committee's approach to message format standards, this list is reasonable. It reflects those systems most affected by the current standards. However, we believe that it may be beneficial to consider the data and documents that would be integral to the claims attachment standard.
Messaging standards for claims attachments will be required, but in their current form, they do not address the completeness of the data for either the purposes of the valuation of the claim or the privacy standard relative to minimum necessary information. Until the data sets or document sets can be better defined, significant variability will exist in the data, which is sent in support of claims.
You have asked us to provide additional comments or critiques. The American Health Information Management Association appreciates the opportunity to suggest additional NCVHS activities that will enhance the development of computer-based patient records. You have talked about picking the low hanging fruit. We are going to suggest that you add some fertilizer to the tree.
By way of background, we all know how the health care profession has increased its acceptance and adoption of standards, particularly messaging standards. Dr. McDonald yesterday talked about how the penetration of HL7 standards is affecting his organization.
Certainly, you have seen that in the NCPDP's standards being adopted by 90 percent of community pharmacies. We believe that continued adoption of messaging protocols will contribute to the advancement of electronic patient medical records. The committee has recognized in its report to the Secretary in July messaging protocols are only part of the solution. Additional standards in the areas of medical terminologies, data quality, accountability and integrity should be pursued.
The report identified seven health care terminology domain areas, including message specific codes, nursing codes, diagnosis and procedure codes, clinically specific codes, drug codes, nursing codes, convergent terminologies and others. There is considerable standards work being done in these areas, but its adoption lags behind and development is significantly more complex.
In the committee report to the Secretary, you described the advantages of collecting information once and then making it available for multiple purposes. AHIMA strongly supports that approach, both for its effect on quality of data and gained efficiency.
To advance this premise, more attention should be focused on creating clinically relevant data sets that would improve documentation and enable automation. Work has begun on the definition of data sets. However, it has not been well coordinated. I give a list of examples, such as the ASTM's minimum essential data oasis, HEDuS(?), MDS, DEEDS(?), the list goes on.
However, there is only limited data consistency across these data sets at the data element definition and category levels. In addition to the work begun on data sets, efforts also have been made to standardize what we currently think of as documents or reports. In 2000, ASTM's E30 122 committee worked on a draft standard specification for health care document format, history and physical examinations and similar documents.
Use of the HCFA AMA documentation guidelines for evaluation and management reporting by physicians has created more structure in physician records. In the past, progress notes contained only information physicians needed for continuing patient care to meet the requirements for professional service reporting and to justify billing or reimbursement. HCFA and the AMA have identified seven elements that should appear to support this date of service level.
They include history, examination, medical decision-making, counseling, coordination of care, time and nature of the presenting problem. We also know that efforts have been made by many specialty medical associations to address documentation requirements for their specialty. In addition, the professional review organizations have identified documentation guidelines as part of their PEP(?) quality program.
The examples I have just cited point out only some of the work to date in identifying data standards and documentation guidelines. Terminologies provide the foundation for the language of medicine and, therefore, may be the most dynamic and challenging piece of the computer-based medical record puzzle. Substantial work has been done on terminologies and several groups continue to advance clinically specific terminologies.
Included with this testimony is a table of all such systems and developers of which AHIMA is aware. In your report to the Secretary, you identified a number of health care terminology domains. Of the seven categories reported, only diagnosis and procedure codes and drug codes are currently addressed by the HIPAA regulation.
The work of standards development in data sets, structured documents and terminologies, with a few notable exceptions, is also occurring largely outside of the SDO structures and, therefore, may be overlooked, duplicated or not advanced quickly enough to impact the PMRI development.
Mr. Chairman, I would like to summarize our comments by making some more specific recommendations to address both your current plans to evaluate message standards and our suggestion that there is a challenging and necessary role for NCVHS to play in advancing the always challenging work of standards and terminologies, data sets and structured documents.
We believe that to advance the message standards work, the committee should consider first formalizing the feedback loop between SDOs and get standards implementers to ensure the development and maintenance of high quality standards. Second, adopting a series of activities that promote the use of standards beyond the establishment of regulation.
Third, using the information collected from your questionnaire to identify overlapping standards development and identify a process to facilitate communication between SDOs and reduce unnecessary duplication of effort.
Finally, as information is collected from the SDOs, that collection could include their self-evaluation of compliance with the criteria. This evaluation should ease the burden of report analysis and provide the SDO with an awareness of how compliant their standards are with the criteria you have established and might address the explicit versus implicit discussion you had yesterday.
We believe that to advance the standards and terminologies data sets and structured documents, the committee should also consider convening hearings to invite discussion on how a process could be established to engage these data set document and terminology development activities in the existing SDO infrastructure.
Move first to work with organizations, whose work is most relevant to claims attachments; that is, emergency records, operative reports, discharge summaries, consultations and therapy notes. Finally, establish criteria for organizations that are not currently SDOs, but are developing clinically specific data sets, document structures and terminologies to advance their standards to a level of adoption by the Secretary.
Mr. Chairman, I thank you for the opportunity to comment. AHIMA is committed to working with the NCVHS and the industry, to advance the state of health information. Your work is clearly integral to that goal.
I will be happy to answer your questions.
MR. BLAIR: Okay. Sandy, we are going to save the questions until after Helen has also had a chance to testify. Thank you very much.
Helen.
DR. SHAVER: Good morning. My name is Helen Stevens and I am here representing McKessonHBOC.
Just to give you a little bit of background about McKessonHBOC, we are the world's largest pharmaceutical supply management and health care information technology company. The information technology business, which I am specifically a part of, develops solutions, including enterprise-wide health care, patient care, clinical, financial, strategic management software, Internet base and networking technologies, electronic commerce, outsourcing and other services to health care organizations throughout the world.
It is a very large organization and has a very large commitment to the health care and to standards. I would like to thank you for the opportunity to provide a contribution towards the NCVHS activity. McKesson is a strong supporter of standards development and believes very strongly in the contributions that you will be able to make through this activity.
I have circulated a copy of my testimony. So, I am going to address just some specific highlights from that. We have reviewed the SDO selection criteria and the questionnaire and agree with the approach and with the questions. We have a few suggestions of clarification in the questionnaire and maybe some additional questions.
With regard to the selection criteria, there was one area that we didn't feel was adequately addressed and that is the publication method for the standard. With the increasing complexity of information standards, it is critical to their successful implementation that they be published not only in hard copy document but we have databases or tools or XML renditions of the standard so that they can be much more easily integrated into our products.
MR. BLAIR: Helen, could you talk a little closer to the microphone, please.
DR. SHAVER: Sorry.
In the indicators of interoperability questions there were -- the term "user" was used. It wasn't clearly defined as to whether you were referring to a user being a vendor/developer of software or a user being a health care enterprise implementing software using the standard. So, maybe some clarification in that area.
Also, the questions of customizations were brought up. Customizations are interesting because they can sometimes be considered a strength in that they allow the standard to be implemented more broadly. So, in the questions of customization, it is important to ask if customizations are used, why are they used and most importantly how are they reconciled?
When discussing issues of quality and accountability, McKesson recognizes that there are lots of alternative vocabularies for the same concept, different vocabularies being adopted for different jurisdications or different functional domains.
We, therefore, suggest that a question be added specifically addressing domain or specific vocabularies if the SDO supports these and if so, how are they reconciled? In the questions regarding market acceptance, it is very important to determine the commitment of supporting application vendors for the standard.
This can be done through determining if there are generally available tools designed to help in the standards implementation. Things like integration engines, compliance tools, specification management tools, do they exist and do they support the standards organization.
Timely standards development procedures are critical that they remain current with business requirements. However, the validity and market acceptance of the standard often lies with the participation in the standards development. As such, we suggest adding a question regarding the SDO's specific structure to support and, in fact, encourage the input from all the stakeholders involved in the standard.
Those stakeholders would include vendors, providers, payers, academic institutions and, of course, governmental agencies. The inclusion of all the stakeholders will help ensure that the standard not only meets the criteria set out by the NCVHS, but is, in fact, a standard that can be implemented successfully.
In determining the flexibility to respond to new requirements, we were looking at the question of separating the information model on which the standard is being built, with the actual messaging standard and from the technology actually used to transmit the message. A core stable information model as a basis for all of the standard -- and if there are multiple standards of organizations developing standards, that they all be working from the same core information model.
That model should be stable. That is the model that we are going to build into our products. Built on that model, we have got the actual messaging standard and that is expected to change over time as business needs change. Again, a separate layer and distinctly separate from that is the technology, something like XML that is actually used to transmit the standard and that the standard be developed independently of how it is actually communicated because this week it is XML is the flavor of the month, flavor of the week.
I don't know what it is going to a week from now, but we need to be flexible to be able to adopt new technology. Those are the specific points I had with regard to the questionnaire. We would also recommend that the NCVHS look at creating vendor and provider questionnaires. They could be fairly brief, but to address the question of vendor and provider implementation, what standards are actually being used out there and implemented and what success they had with implementing the standards.
In considering scope, we had a look at the transactions that were proposed and we found them to be very broad. They included patient administration, clinical care, pharmacy, laboratory, radiology, all of the main health care areas. There were two areas that we felt needed to be more explicitly stated. The first is EMR exchange and this is the actual exchange of messages between EMR systems. Examples would include hospital discharge summaries, physician history and physicals and messages of that type.
The second is identification reconciliation, which is sometimes assumed to be part of ADT, but we feel that it should be explicitly stated. What we mean by identification reconciliation is messages to reconcile multiple occurrences of a person with different identities.
For example, a John Doe comes into a health care institution and is given a temporary number. when his identification is confirmed, merging the temporary identification with his permanent identification. Being able to accurately identify a person is critical to creating an PMRI.
The impact of the scope has to be considered with the impact of the industry when the standard is selected and that must include an evaluation of the cost of financial and operational against the expected benefits, as is any case. The impact on the vendor community would be significant if all of these transactions were mandated to be a standard that is not currently widely implemented, requiring prohibitively high cost on vendors and, of course, translating to the receivers of the -- the users of the product.
At this time, the NPRI message format, in fact, the primary message format supported by McKessonHBOC is the HL7 2.X standard. I call it 2.X because we have a range of implementations between 2.1 and 2.4. Most of our implementations at the moment at live sites are in the 2.2 range, which is quite old, not something I am happy about.
The primary barrier to upgrading the interfaces has been the operational impact on the organizations during the upgrade. With any software upgrade, it is necessary to go through extensive testing, even more so in the health care industry and when we are testing interfaces, we are very often impacting at least two systems, sometimes between two vendors and if it isn't broke, don't fix it, unfortunately, is very prevalent among our customers and it requires some justification before changing an interface to a new release.
However, the work that the NCVHS is doing in determining a PMRI standard is going to be an extremely beneficial goal and it will help us have the impetus to move some customers from older interfaces to more updated, more robust standards.
We suggest that the scope be defined in incremental phases, starting with messages that are most compatible with the HIPAA, that most complement the existing HIPAA standards. These would be administrative, financial and patient, as well as demographic management standards as a first round, following medication management, including treatment medication orders and prescription information.
As a primary concern, we feel that these are the areas where there can be most impact and most benefit, again, going back to the cost benefit analysis. We recommend that the NCVHS focus first on the transactions related to patient demographics, identification and that is including admission, discharge, transfer, the identification reconciliational merge transactions identified, as well as the EMR exchanges.
These will build a solid foundation of the PMRIs and then additional transactions can be added in phased approach.
MR. BLAIR: Very helpful.
MS. STEVENS: As you can probably tell from my accent, I am not from the United States. I am actually from Canada, originally from England. I have traveled a lot and I have been working in the United States and I now live in the United States, been working here for six years. But this has given me a recognition of the impact and contributions that the USA can make internationally in this area. For this reason, I would recommend that the NCVHS consider the international implications of defining an American national standard, working in this area.
I was talking last week with some contacts at the Canadian Institute for Health Informatics and they were very interested in what you are working on. They are struggling with the same questions, working on the same kind of projects and they will be looking at you when you provide leadership in this area.
Ensuring that the standards selected and the work that you do is synchronized with existing and future international standards will help not only the international community but also the United States by supporting our organization, such as the CDC, reporting requirements for the World Health Organization and contributing towards global disease management strategies, rules that help reduce cost in standards development as you can leverage the contributions made in the international community.
As a company, McKessonHBOC works throughout the world. We are an international organization and we actively promote the internationalization of standards. Quite frankly, it is much easier for us to develop to one standard and produce a product that can be sold internationally.
So, whatever activity is done, consider the international strength of the standards that you select and consider working with organizations like ISO in achieving international recognition.
In conclusion, I would like to emphasize that the development of the PMRI standard depends not only on the development of a messaging transaction -- a form of a comment information model. Even more fundamental than common vocabularies is the requirement for comment information model and the adoption of a single information model that all standard development organizations work towards would be key to ensure that those standards, even multiple standards organizations that they are working from the common understanding of the information underlying.
That concludes my testimony. I thank you for your time and attention and am open to any questions.
MR. BLAIR: Helen, thank you.
George Arges, could you introduce yourself and may we have your testimony.
MR. ARGES: Sure. Thank you.
My name is George Arges. I am the senior director of the American Hospital Association's Health Data Management Group, located in Chicago.
Again, thank you for allowing me to be here today to testify on the process for adopting uniform data standards for patient medical record information in the eventual electronic interchange of such information. Let me begin by saying that the AHA generally supports efforts towards the development and eventual adoption of such standards.
However, at this time, we are very concerned about any additional regulatory efforts in this area. As you know, during the past few years, hospitals have had to confront a number of massive information system changes. These changes include Y2K, the balanced budget act of 1997 regulations and the recent HIPAA 1996 administrative simplification provisions that include the transaction standards, privacy and security.
All of these changes are massive, complex and costly initiatives that are challenging the viability of hospitals. Finding the necessary funds to adopt such changes would be crucial. Even more troublesome is that many of these changes offer little in the way of immediate administrative relief.
It is important to note that while these changes may ultimately yield improvements to the entire process, they are also in direct competition with the day-to-day expenditures for basic patient care services. While the NCVHS report for uniform standards for patient medical record information identified a series of recommendations that it intends to forward to the Secretary for possible action, it is vitally important that the recommendations also include a corresponding study that examines the financing options. Such a study should also look at the ongoing costs and challenges.
For instance, recent regulations, such as the BBA and HIPAA have ongoing program components that require ongoing funding to accommodate yearly changes. These information system changes are vastly different than the Y2K activities that had a clear endpoint.
Last year's report brought forward a series of recommendations toward the establishment of the PMRI standards. One area that I would have liked the report to explore more fully involves the data gathering and development process. How are caregivers and others involved in the review of gathering, developing and posting the information processes? Which data standards are they familiar with and how well trained are they in their use in applications?
Which technologies do providers readily have available and are these tools capable of complementing the way they deliver care? While it is important to select the data content standards and design national coordinate standards that would allow interoperability of information systems, the development of medical record standards must also take a revolutionary approach not only in the assembly of the medical record but also in terms of functionality.
An important functional goal for the adoption of new approaches is that it must free up the caregivers' administrative work load and ultimately yield complete documented and standardized medical records. The report should place additional emphasis on this aspect of revamping the existing flow of administrative tasks, rather than simply introducing new electronic standards to today's medical record development process.
While improvements can be gleaned from your recommendations, hospitals are weary of new regulations that add new financial and administrative burdens without a compelling return on these types of investments. Streamlining the administrative tasks should be a top priority. For example, today, hospitals develop the appropriate diagnosis or procedure codes at the end of the patient's episode of care. This process requires someone in the medical records department to abstract numerous pieces of information collected throughout the patient's day.
This process must improve the way information flows and should be designed to work with the caregiver. This requires a complete examination of the process. Which pieces at the front end require development of standard templates that can aid the caregiver in their evaluation and management of the patient?
Can caregivers agree on the key components they typically would need to record during the evaluation of the patient, the complaints, symptoms and other existing conditions or manifestations? Can typical treatment option and tests ordered be included as part of that template?
The report's major findings and recommendations are an important starting point towards the development of PMRI standards. Nevertheless, the solution for their adoption does not lay in regulations, but rather available documentation that contain specific performance elements needed to support the goals of each finding.
Whenever hospitals are in the market to upgrade or to acquire new medical record information systems, what performance elements should guide their evaluation? Are there reviews of vendor systems? Is there a certifying body that can demonstrate which vendors incorporate much of the NPRI performance standards within their products.
Can case studies show how some hospitals approach the adoption of these changes, the context of doing so, the method of financing, the challenges they face and the benefits gained for undertaking these efforts.
It is also important to launch demonstration programs that test the validity and can quantify the costs associated with the adoption of these standards. Finally, hospitals today need regulatory relief and financial support, especially for unfunded mandates, like HIPAA's administrative simplification provisions.
In the meantime, help guide the market in their understanding of the standard components that hospitals and vendors should follow for incorporation to their information systems. Let the marketplace determine when it is the right time to replace their information systems.
Again, thank you for allowing me this opportunity to testify and I will be happy to answer any questions you may have.
MR. BLAIR: Thank you, George.
Do we have questions from the committee?
DR. COHN: Mike is raising his hand.
DR. FITZMAURICE: I would like to start with Helen. At the end, Helen, you advised us to build standards around a single, comprehensive information model. Models are difficult to develop and it is hard to get consensus agreement on them, a condition that dump trucks pouring sand down a large hole and the hole never seemed to be filling up.
It is a very simple model, but which I can explain what we are doing to my bosses, to the Department, to congressional staff and others who have seen it -- who need to see the need for putting resources into this effort. It also helped guide us in recommending patient medical records standards to have a picture and say it fits under this box. Now do you want me to drill down on the box?
Do you have some models in mind? There is an Australian model. There is a GEHR, a good electronic health record. There is an HL7 reference information model. There are some models around. Is there any one particularly that HBOC uses or that you would recommend for our use?
MS. STEVENS: Thank you.
Trying to model health care, even within the United States, is an astronomical task. Health care is so unique. It is a unique industry in this. Trying to create a model that is internationally recognized and accepted just makes that an almost impossible task.
As I mentioned before, we are very active with HL7 and I am very familiar with the information model that HL7 has developed and has invested hugely in developing. That is the model that McKesson is now looking at adopting and pulling into all of its products, but not only for messaging standards.
The information model is a foundation piece and upon it you can build information messaging standards. You can also build a number of other things. McKesson is specifically building a web technology based on that information model and using it for other areas. Therefore, integrating it into our products derives a number of benefits.
I am not familiar with the other models specifically. So, in answer to your question, the HL7 model is --
MR. BLAIR: Michael, did you have a follow-up question?
DR. FITZMAURICE: I will hold off with some for awhile.
MR. BLAIR: One of the things that I have found very helpful about this particular panel, with Sandy, Helen and George, is that it is to me representing almost the full spectrum of views because Sandy wound up pointing out that while this questionnaire is taking the first phase on message format standards and we -- and that may not even be all message format standards. Sandy was pointing out the need and the value of looking beyond just message format standards that we should consider medical terminologies and data content and information models.
Helen as a vendor was looking at specific next steps. I am going to have a question for Helen with respect to that because I think that that is kind of where we were sort of focused, on a specific, small, next step.
I think George's testimony was extremely helpful because I think George's testimony represented a great deal of the concern in the provider community, not only with respect to, you know, the past couple of years with Y2K, but with how to accommodate to the other HIPAA regulations and the idea that we would come out with additional mandates for standards or mandates for upgrading of systems would be really very inappropriate and many people would feel threatened and uncomfortable with that.
I think, George, that most of us on the committee don't want to do that. So, let me go ahead and focus a little bit on Helen with your recommendations because if we look at what could be a positive catalyst to help vendors and providers to be able to coordinate data, to be able to share data, without being too disruptive, in other words, take a look at what is being accepted in the marketplace and just encourage the next step maybe in a few targeted areas.
You mentioned the fact that McKessonHBOC for most of your implementations are on HL7 2.2, not even 2.3. Of course, 2.4 doesn't even have all of the transactions. So, I want to hone in a little bit. You may want to think about this. If we wind up having a subset of that list, that laundry list of transactions -- and I really like the way you wound up saying, you know, look at those in terms of different stages and steps.
We may not want to do all of them at first. Do the ones that complement the existing HIPAA regulations. That was very helpful. But even more sharply if we wind up picking two or three or four, that we encourage or wind up doing a Good Housekeeping Seal of Approval -- I am not saying that is what we would do, but, you know, if it was something like that, not a mandate, but just a catalyst, something to help move things forward with respect to coordination.
Is 2.3 a good common ground to move forward to, even though it doesn't have XML, what 2.4 does? Can you give us a little bit more guidance. I don't want to get into a selection stage right now, but, you know, in terms of how we move forward, should we wind up -- how specific should we be on the versions in the questionnaire?
DR. SHAVER: Absolutely. Just a couple of points. 2.4 does, in fact, include the full impact more transactions than 2.3 does and 2.4 is being introduced with an updated XML rendition. It is possible to transmit even 2.1 interface using XML. I am not sure I would recommend it.
What McKesson has done is we are trying to balance moving forward, taking advantage of new versions of HL7 against minimizing disruption for customers if they have got a very simple interface and it is on 2.1 and it works, they are not interested in changing that. Our customers range from, obviously, huge hospitals to very small enterprises.
Most of our interfaces are on 2.2, but there is heavy peppering of 2.3 and 2.1 in those interfaces. In fact, I would say that within the next year, I would be able to say most of our interfaces are on 2.3, 2.31.
If you were to look at a 2.X series for a recommendation, I would strongly recommend that you only look at 2.4, don't even think of going back below 2.4 as a standard. We are now -- anything that you recommend or put forward, we would start working on development and that would be a great impetus for us to get our 2.4 interfaces coded and out there. At the moment, we have several products that are already coded at 2.4 interfaces.
There is a delay -- obviously, a standard has to be released. 2.4 was released this spring. It has to be coded, developed and then implemented and adopted by a customer. So, there is a delay there. So, definitely look at 2.4 with a recommendation that maybe within a year of 2.5, if 2.5 --
MR. BLAIR: If there is a 2.5.
MS. STEVENS: If there is a 2.5, that that is the direction. Always look forward. We are going to be lagging at least a year to two years before you see these interfaces actually operating in hospitals because that is how long it takes to upgrade and implement them and get them out there.
MR. BLAIR: Other questions from subcommittee members? Maybe, Kepa -- we should make sure we give him a chance.
DR. COHN: Can I ask a question? I was going to ask some questions.
I actually wanted just to sort of get the testifers' views on sort of -- you were talking about information models, which is -- once again, there are so many of them it makes you a little nuts from time to time, but data sets are another wonderful topic in all of this stuff. I was sort of impressed, George, with your testimony.
You were talking about the requirements on hospitals. Certainly, one would observe that they probably have less to do with standards than they do with various groups requiring data sets of one sort or another. I mean, for example, if you have to do a claim effectively, you are submitting a data set to another entity.
It gets to be more of a question of requiring that a data set be submitted -- the standard for that data set. Now, Sandy, you missed her testimony, was sort of commenting that there are a zillion different sets on a zillion different issues, sort of around and about. Is there a need for this committee to wander into the area of data sets?
I mean, I worry about it a little bit. There are lots of standards, but there are almost more data sets. It seems to me that they are much more industry driven. ACHO has data sets that requires -- HCFA has data sets, on and on and on. Is that an area that we should in this work consider?
George and then maybe Sandy or vice-versa? Do you have a comment?
MS. FULLER: Well, actually, I was hoping that might be what Jim was talking about yesterday when he said that you had circulated the committee's report to departments within HHS and that they were looking at ways they could respond. We actually took all those data sets that are listed in my testimony and laid them out side by side and looked at the data definitions and have a huge file.
I think we were working at that before ANSI HSBE(?) started their what is now -- I am going to get the acronym wrong, the USHIK project. Those data sets not only in hospitals but across health care and if you are looking at how information moves about a patient across health care and that was the function we were looking at.
Even in identifier information, there is very little consistency. Certainly in outcomes information there is very little consistency, down to name and age and sex, that almost all of those data sets are driven in this building and I think the one next door.
So, that would be very helpful if we could get a level of consistency there, as well as kind of saying that there should be a standard that when we are looking at the development of those data sets, that the consideration not only for its applicability and utility within that scope of care; so, a HEDUS data set or an OASIS data set or an MDS in long term care. That is, patients now move through a much more fluid health care delivery system and a more fluid information system, we would hope, that that data can be generated, replicated, transmitted reasonably from the same systems, which just doesn't exist today.
MR. ARGES: I guess I wanted to -- to get back a little bit on what Sandy was also saying, too. To the extent that data sets basically are composed also of data elements, it would be nice to basically make sure that many of the data elements that support the data set are common throughout or that the initial building blocks that make up a data set, but the data set itself should have a specific application, whether it is the MBS or whether it is a claim or whether it is for public health reporting.
Its application, its use should be well-defined and it would help if the NCVHS could pick the data set, the maximum elements, data elements, that define a particular scope of activity. That would help providers at least work or build around to support those type of information systems. Also, it would help if they are shopping for vendors or whoever to support their activities, to make sure that they will work within those existing data sets and they would be capable of exchanging information from one facility to another.
DR. COHN: Can I ask one follow-up question, George, on this one? And then we will let the others ask questions. That just had to do with sort of the nature of the standards that we are talking about.
As I think Jeff had commented, there are many ways to consider these standards as we move into these clinical areas. Would the American hospitals and I am not sure if the association, but say that the hospitals themselves e happy if there were a pronouncement that sort of is on the level of you don't have to do it, but if you do it, do it this way.
MR. ARGES: I think we would be happy with the idea that vendors and others would have guidance in this so that when they are in the market to update or upgrade, they have something to evaluate. They have something to evaluate the existing system to what the market is offering. Whether that market itself would support the eventual pathway down towards greater standardization. I think that would be the best of all worlds.
But I think the marketplace needs guidance and we need to be able to be assured that when we are capable of making that leap towards a new system, which vendors are out there that will support us.
DR. COHN: Okay. Thank you.
I think Mike and then Ted.
DR. FITZMAURICE: Good question, Simon.
I was going to ask a variation of it. First of all, because I don't have to ask the whole question, for Sandy and then one for George.
Sandy, you mentioned limited data consistency in your testimony and so we naturally think where should the government help out. What is the role of the government?
One of the projects that you mentioned was USHIK, U.S. Health Information Knowledge Base that started as a pilot between HCFA and DOD, expanded to NCHISB and asking the standard developing organizations to submit their data dictionaries to supply information to this meta-data registry and some can envision it going to other organizations in the private sector, going to government agencies, having databases submitted so that eventually we get to see maybe all of HCFA's databases, all of the databases that AHRQ would use for patient safety.
How do those definitions match with definitions called upon by others? So, one of the things the government could do is collect data dictionaries, collect these things and try to match them up and not force anybody to do anything, but simply to say look at all these definitions. Maybe just by putting it out there, we can get more movement toward some central definitions or some common definitions. Is that a real small effort in your opinion? Is that what you were talking about as some way to address this? Because you listed a bunch of data sets.
MS. FULLER: That is exactly what I was saying and I think you -- the piece that I hadn't seen, at least the last time I was checking in with the meta-data registry is the piece that HCFA could do. A number of those data sets that are completed across health care are HCFA mandates because they are filling their outcomes data sets and if there was this common -- of course, they are not going to have the same data because they address different sectors of the industry, but they do share a fairly common -- they share a lot of data in common.
If those could be common, you would be saying two things. One, I think, you would be making it easier for vendors and providers to provide information. Two, I can't believe you wouldn't be making it easier for you guys to analyze information and, three, you would be saying there ought to be standard definitions and you would be encouraging standard definitions.
So, the sooner you could get to that level at USHIK, I think that would be helpful.
DR. FITZMAURICE: And a question for George, if I may.
George, when I looked through your testimony and listened to it, two words bounce right out at me. One of them is performance elements and the other one is unfunded mandates. In everything we do sometimes in the regulatory process, there is a tendency not to look at the source of the funding, even though we look at, you know, how much will things cost.
One thing that the government could sponsor, could propose, as part of a research program would be studying medical record information systems, study the contracting process, the specifications and then related to some satisfaction with the outcome or the results. Did you get what you asked for? Are you happy you got what you asked for? What would have helped you be more specific and relate that to what you end up wanting?
I think there is an awful lot of uncertainty. It is not sufficient to put in, well, this system must be compatible with all HL7 standards because all HL7 standards uniformly fit together and we are trying to move forward to make them fit more and more together.
I see, too, that some analysis of the use of these systems that hospitals might obtain, can they be used for patient safety reporting if that becomes an issue down the road? Can they be used for obtaining accreditation information or performance measures, HEDUS measures, as well as other clinical performance measures, from the system rather than having to have somebody go through chart reviews and audits?
It gets into not just the realm of unfunded mandates but what are the benefits of the system? What are the costs? And what tools or information could be available? I think many of just don't know what tools would be helpful, what information would be helpful in getting people to make wise decisions about the systems they by in hospitals, in group practices and others and to maximize or improve their satisfaction with the system once it is implemented.
Do you have any suggestions, any tips, any desires for what information you would like to see about that process?
MR. ARGES: Well, to me the process ultimately has to yield better patient care. That is first and foremost.
MR. BLAIR: Slightly closer to the microphone.
MR. ARGES: Better patient care is the ultimate end game, I think, throughout all this. The ability for the caregiver to basically improve on the delivery of care, I think, is another dimension. It would be great if outcomes were identified in a more specific way so that when facilities are looking again to upgrade, they can get a sense of what they can anticipate from that acquisition of new system changes and also so that they can put into motion the necessary other resources of whether it is staff training or other components to help support that.
Clearly, the standardization of data elements, a data dictionary would be helpful. Having systems that are interoperable are a key component. But I think fundamentally you almost need to -- there are elements, I think, of electronic patient medical records that are really somewhat unique and fabulous out in the marketplace. But I think it would be nice to get a sense of why those worked well, under what circumstances do they work well and those circumstances be reported over to a broader audience.
So that, again, the providers have -- hospital providers have a better sense of when they should engage in updating and what they want to achieve as part of their outcome. But safety as part of their process is a key element and, again, that is all, I think, driven back to patient care.
DR. COHN: Ted.
DR. SHORTLIFFE: Yes, I would like to comment first on George's testimony and then ask a question.
The comment comes from the general theme of the pressure that new standards in this area would place on hospitals is one I am very sensitive to. I think the committee is sensitive to it. In fact, I very much remember Clem's warning when we first embarked on this a couple of meetings ago about the need for these not to be viewed as the kind of mandatory standards that would create that kind of sort of insurmountable pressure on institutions and organizations.
On the other hand, as someone who works daily worrying about these kinds of issues within a hospital environment, on the one hand, I am very aware that requirements for new standards in this area would be stressful to say the least, both financially and organizationally.
But on the other hand as I look at the day-to-day complexities that we face now, it is precisely the absence of such standards that accounts for a great deal of the turmoil that exists in the current hospital environment. So, educating people about that and creating sort of guidelines rather than standards, if you will, would sort of help promote the movement in the right direction. Guidance to industry, rationalizing the interactions among the various standards organizations so that there is some kind of consistency across and these all seem to ultimately really benefit the hospitals in precisely some of the areas where they are feeling the greatest stress right now.
Having said that, I was interested in your closing remarks. In a way this relates to a point that Mike was making a moment ago about hospitals needing regulatory relief and financial support. I don't want to get too far astray from our primary charge here, but I do think these things are all intertwined. So, let me try to pursue that a bit.
We see then a hospital industry and, in general, a health care industry with remarkably small investment in information technology solutions as a percentage of total outlay, compared to other information, intensive industries, of which the favorite comparisons are with, say, airline industry or financial institutions. Right? You may know the number better than I across today's hospitals, but the frequently quoted figures are down around 1 to 2 percent of total budgets in information technology, which always looks like a big number to a CEO of a hospital, but on a comparative scale is sort of small compared to other industries, but the whole cost and the financial structures of hospitals are now so based upon sort of existing traditional reimbursement schemes and cost models that it would be, of course, absurd to imagine hospitals increasing to 18 to 20 percent overnight their investment in this area.
So, given that, we are beginning, I believe, to hear discussions about what other mechanisms there might be for actually helping to encourage hospitals to make the kind of investments in computing and information technology to support the kinds of medical record systems that are the motivation behind today's discussions. So, this little phrase about hospitals need financial support peaked my curiosity and so in the context of that sort of long introduction, I would be interested in if you have specific thoughts, suggestions, recommendations we should be considering as part of this.
DR. COHN: Would you hold just a second? I think Kathleen has a question, clarification anyway.
MS. FYFFE: Ed, what was the percentage that you referred to, 1 to 2 percent?
DR. SHORTLIFFE: That is the figure I keep hearing quoted, although I have never -- of operating expenses.
MS. FYFFE: Of operating expenses, 1 to 2
percent --
DR. SHORTLIFFE: In information technology.
MR. BLAIR: Can I elaborate a little bit more and my information is a little bit old, but there was data provided for many years and is probably still being provided now by a particular consulting firm that would sample 3,000 of the hospital's over 100 beds, Shelton Dorn(?) of North Lake, Chicago, and what you would wind up seeing as he wound up giving the data, he would reflect the data specifically by the size -- the number of beds in a hospital and the largest hospitals might be up at 2 or 4 percent, but there weren't very many of them.
As soon as you get below 300 beds, it typically would drop below 2 percent. I don't know that that data has changed all that much in the last few years.
DR. COHN: George, you probably have more information --
MR. ARGES: I don't know that I have any more. I have heard the figure 1 to 2 percent, as well, in the past. But what strikes me about that is, clearly, I mean, there is a gap between that and let's say other economic sectors that spend money on information systems.
But, you know, health care and hospital health care, in particular, is one that is a little bit different model. Many are not-for-profit organizations. Many, basically, do community service and are operating on a different level.
Their margins, you know, are very weak. The operating margin, in many cases, is in the negative area and it is the non-operating margins that -- that pulls in any excess. The important thing here, I guess, what I would like to see is Congress really take a look at the environment and where it wants to go, what it wants to do.
To me, and I am sure an economist can probably make the argument much better than I can, is that if health care is 14 percent or 15 percent of the gross domestic product, it is one of the largest sectors of the economy and, yet, it needs an infusion of resources to basically jump start it into a new information age. Congress needs to recognize that and Congress needs to step up to the plate.
They have got, you know, the right direction going, but they need to understand the funding. If you look at the 1930s and 1940s, when Hill-Burton was created and it basically funded the creation of bricks and mortars for facilities, are we not at that juncture today where that same sort of activity needs to take place and Congress needs to recognize it.
If, you know, this whole issue of health care reform, you know, that was talked about eight years ago was a real issue, it still is a real issue today, and information systems can hold a key element in reducing the spiraling rate of health care cost increases. To that extent, it is a wise investment. It is just that hospitals and other providers need something to jump start to make this investment.
DR. SHORTLIFFE: Can I follow up then because -- I mean, you said the magic words that I keep hearing these days and that is this analogy to the need for a Hill-Burton equivalent in this area. I think this idea is suddenly getting quite a bit of momentum and I guess -- at least I am hearing it discussed a lot in hospital and health care system circles.
So, the follow-up question is if that were to happen, I mean, if this were to be understood by the Congress, if there were mechanisms by which hospitals seeking to sort of do the right thing in this area could apply for either as individual organizations or as consortia the kind of funding support that it would take to more, shall we say, energetically attack some of the problems that we have been discussing.
Wouldn't that make some of the kind of standards recommendations that are likely to come out of NCVHS more palatable if they occurred in the context of the kind of assistance that would --
MR. ARGES: Definitely. I mean, what we are hearing today with respect to implementation of the transaction standards on HIPAA is people need guidance. How do I know I am dealing with a vendor who will send me the right solutions in this environment and the greater the specificity in this type of area, the better off providers will be. They will have a clearer sense of who they are dealing with, what they are acquiring and the benefits that they can expect.
MR. BLAIR: Could I jump in on this because I am trying to -- and I think that Kepa may want to make a comment on this as well.
In terms of guidance, the area of guidance -- let me step back just a little bit to sort of frame this slightly because what we have been doing here at NCVHS is we have not attempted to wind up saying we are going to standardize electronic patient record systems. It is that we are looking to simply facilitate communications, interoperability, comparable data and data quality so that different vendors can continue to wind up producing the best systems they can without limitation, but that they can share data and to be able to share data, that requires standardization.
As you can see, you know, when you look through the report, you could wind up seeing that there is quite a diversity of standards. There is a lot of dynamic involvement in there. So, it is not obvious to wind up saying here is a specific standard. Health care is multifaceted. So, we are winding up trying to proceed in a stair step manner.
Now, getting back to your point about -- because I have the feeling that you agree with the direction but because of the financial conditions of most hospitals, you don't want a poorly thought out direction by NCVHS to cause additional financial burdens or disruption. One of the areas that I would like to ask you about that has come up is the idea of conformance testing and some certification that vendors, in fact, really meet the standards, really comply with the standards. Is that the kind of specific targeted suggestion or recommendation that you feel would be appropriate to move the ball forward, but not in a coercive manner, where hospitals can wind up saying "yes," this particular vendor has gone through the conformance testing and has certified that, in fact, their data is interoperable?
Is that what you are looking for?
MR. ARGES: Yes. In fact, that would be very helpful. If we could have a demonstration testing to test the validity of some of these assumptions and then have a certifying body that could basically evaluate different systems, vendor systems, out there and determine they conform to certain basic functions, I think that would be helpful.
MR. BLAIR: Since you said that it would be helpful, could I invite Kepa, because he is especially focused on this particular area -- Kepa, are you able to hear?
DR. ZUBELDIA: Yes.
MR. BLAIR: Do you have a comment or question with respect to this?
DR. ZUBELDIA: My comment is, yes, I think that is extremely helpful not only in these standards but the certification standards, as you would expect me to say. I think it is very important to have a yardstick to measure the compliance with the standards.
But I have a question for George. I assume, and correct me if I am wrong -- I assume that the patient medical record information standards, like HL7 and so on, are actually a savings for the hospital. They are not a cost -- there may be some cost in implementing, but actually they realize savings in the long term. If the NCVHS was to recommend that there be PMRI standards and the providers, that is, the hospitals implement them -- if they want to implement them on a strictly voluntary basis for implementation, but if they do implement PMRI, they should use the specific PMRI standards for the exchange of the information.
Now, they can implement with any vendor and the vendor may have additional capabilities, but for the exchange of information, they should implement specific PMRI standards. I don't know that anybody has done a cost analysis on that, but how could be detrimental over the situation that we have today, where everyone can implement whatever they want? Is it in their benefit to do standardization???
MR. BLAIR: Were you able to hear the question?
MR. ARGES: I think I missed part of it, but I think I understand where he is going, but let me try and answer it and you can follow up, Kepa, I guess, with another question if I am off base.
Obviously, pointing to a certain set of standards is -- messaging standards and other types of activities and allowing the provider the flexibility to move in that direction when it is right for their organization to do so, I think, is important. It just may take longer for that organization to move in that direction.
But they have to judge the other factors in their delivery of care and how their performance is being helped or hurt, but they have to make a decision as to when they are going to make that investment with the information systems.
I suppose that if grant monies were made available it would be sooner rather than later. The other factor is there are many small facilities that are somewhat disenfranchised as currently. This is not the same type of competitive business model that exists for other industries. They provide a service.
Does that answer your question?
DR. ZUBELDIA: Yes and no. The question that I have is a benefit for knowing what the target is when they make that leap and having anybody jump onto the same standard. When they do choose to make that jump, is there a benefit to that?
MR. ARGES: There is a benefit to that. Having the target, number one, puts market forces in play, at least for the vendor community, so vendors can build systems that would bring to the provider community solutions that would benefit them and perhaps do it a more cost effective way because there is a target out there. So, they would, obviously, compete with one another. Knowing that would allow them to spend money wisely in terms of these systems.
DR. ZUBELDIA: So, what I am hearing is there is the benefit in having a common standard for PMRI to use as a target and what you are also asking for is some way of measuring whether they are on target or off target and make sure that everybody knows who is on target.
MR. ARGES: Correct.
DR. SHORTLIFFE: Well, it is just sort of the logical follow-on to this, I think. This model of kind of a certification process for the vendors, if you will, and compliance to inform the hospitals and the other provider organizations, it would be only logical to ask Helen what she thinks of this idea because it is bound to sound good to the hospital.
MS. STEVENS: Thank you.
Let me put a little perspective on certification. We have probably got about 50 primary products. Each of those products puts out a significant release once a year. Many of the products put out four releases a year, are doing a certification step to the development, in addition to existing certification would have an impact on delivery cycles. That translates, of course, into an impact in the cost of the product, which is something to take into consideration. Within our organization, we have in the past and do now perform certification testing on our products, compliance testing, to measure their compliance to different standards, including messaging standards. We have had mixed success with the outcome of that compliance measuring because compliance measuring can very accurately measure whether you support a required field in a message or not.
The question then becomes is it necessary for you to support that field, given the specific application? Some of the applications are very specific and have a very narrow need. Where a product might fail a compliance test, it might be completely adequate and satisfy the customers and provide, you know, valuable information and have a very valuable role.
So, any implementation of compliance of products, one has to recognize the scope of what that would include. This would be a massive undertaking, but also be designed in such a way that it does not measure compliance to a standard that it measures the successful implementation of the standard and the successful integration of the information, which is after all what we are trying to achieve. There is no point in me being a hundred percent HL7 2.4 compliant if I am not giving you the information that you need.
DR. ZUBELDIA: Information is extremely important. We have found the same situation in -- transactions compliance testing that Claredi does. A general practitioner may be perfectly compliant and satisfied with a product that stores office visits and a little more. Specialists may have -- an ambulance company or a -- supplier will have different needs. They may all be completely compliant and completely incompatible with each other.
DR. COHN: Kepa, maybe I should ask you this question before we move on to Bill. Is this a feature of the standards or is this a problem with the standards?
DR. ZUBELDIA: Well, it is a feature of the standards. The standards have enough flexibility for this to happen. So, when you are talking about compliance, you have to be very flexible in that if they are compliant, they don't necessarily implement -- features of the standard. So, you have to identify the subset or compliance profiles, as we call them.
Every -- needs a certain compliance profile and the trick is when a provider is -- make sure that the compliance profile of the provider meets the requirements of the payer. Those compliance profiles are flexible and they can change, depending on the circumstances. So, it is a very complex problem.
MR. BLAIR: Could I suggest that it is complex. It apparently is a very important issue. I almost think that we ought to consider a schedule later on this year having a full day or more to understand this very well so that we could give some good guidance on the issues of compliance.
DR. COHN: Well, yes. I think we can talk about it, but I think it is also an issue that has to do with the current HIPAA standards. That is the concept of the maximum data set and all that. Kepa, am I mistaken about that one?
DR. ZUBELDIA: No, you are exactly on target.
I think what we need to do is when we hear from the DSMOs in May, early June, we need to reflect that what they are describing as the issues we have been going through is not just an administrative standard issue, but is also a clinical standard issue. Maybe we can learn from that.
Bill Yasnoff.
DR. YASNOFF: I have a follow up related to the same topic.
Sandy, I believe, you recommended that we encourage formal return on investment studies and as you know, the committee has been -- has given a great deal of attention to the issue of the business case and so on and also to this issue of relatively low investment in information technology in the health care industry relative to other industries.
Are you aware of any such studies that exist already that could be -- that the committee could look at or that could be models for this type of work?
MS. FULLER: That is a really good question. None are springing to the top of my mind. I think what you would find if you went back to look is either application specific or vendor specific. I know as I have talked with vendors that they would all love to be able to point to a good ROI story. I think you would see that in application.
I think the other thing that -- the other way to look at return on investment, certainly health care needs that financial return on investment incentive to increase the investment in technology. I think there are other forces in health care; the medical errors issue. And there are studies about the efficacy of information systems and reducing medical errors.
So, I don't -- I guess I would caution you not to look only at financial models of return on investment but also those things that touch to the very core of what our business is, which is to take care of patients. I wish I could say, oh, sure. Go here. There probably are some out there. Maybe Helen or George know of some.
DR. YASNOFF: Yes, particularly George. Do you have any comment on that or can you give any guidance to the committee on that issue?
MR. ARGES: I am not sure that I would have any hard core examples in terms of what that return on investment -- how that approach should proceed. In many ways, I guess you just have to look at some best practices that are out there where people have implemented it and really document the areas that change within that facility, what change in terms of outcome. Some of it is financial. Some of it may be quality. Some of it may be patient satisfaction.
All of those need to be looked at as part of their process. But I am not sure if there is a standard model that I could point to in terms of a return on investment.
DR. SHORTLIFFE: Bill, could I comment on this? I mean, Stan Huff is here. It seems to me that the most studies that I have seen have been from Utah and from InterMountain Health, where we actually have seen dollar return on investment for certain kinds of information system interventions, maybe not broadly, but sort of single interventions that have been studied and published in the literature.
Do you want to comment on it, Stan?
DR. HUFF: That is true. It is certainly not my work, but the work of Scott Evans and Stan Pistotnic(?) and Brent James, many others, who have demonstrated that and that is certainly our belief. We believe we provided a higher quality of care because of the information systems that we implement. That is one of the reasons we try and work hard on them.
DR. SHORTLIFFE: But you also data on dollar savings. Correct?
DR. HUFF: We do in specific areas. I mean, we keep wishing we had some way to measure the overall cost of the system and the overall benefit. The best we have been able to do is measure or actually probably estimate costs in a given area and measure and estimate benefits. So, for instance, we can -- we have been able to show if you use systems to make sure that people get prophylactic antibiotics before surgery that we can reduce the risk of post op infection by from 4 percent to 2 percent or that sort of thing, that in the face of other increasing costs, we actually reduce the cost of antibiotic use by using the computer system to indicate which drugs would be the most effective drugs in a given area for presumptive treatment of infections, all of those sorts of things.
So, in focused areas we have shown that. We have fallen short of being able to show globally that the system, you know, provides more back to us than what we put in, but that is certainly our belief, based on everything that we have done.
DR. COHN: Can I just jump in here for a second because I don't know that we need -- I believe your data, I mean, first of all. I don't think there is a person in the subcommittee who is not a believer in the cost/benefit of clinical information systems. Obviously, I think what we thought of this is that you need standards to make reasonable ability to implement in a cost effective way those information systems, which I think is sort of how we have all thought about this typically.
Now, the thing that I am hearing George bring up and I guess I just would reflect on for the committee to think about, recognizing that in our original report we did talk about the financing issues. But the problem is is that in the modern day world, there is a disconnect between long term return on investment and short term costs. I suspect that George -- maybe I should ask you if I am right about this one -- in the world where every quarter matters and every six months matters and every year matters and you are talking about information systems that take three, four, five years to be implemented and show an R&I, we are dealing with sort of a major disconnect and so we can talk about, you know, benefits of patient care or we can talk about ROIs, but the American economy as it is structured is not looking at seven year ROIs. They are looking at three month ROIs.
George, would you like to comment on this?
MR. ARGES: Yes. I can put on my actual hospital hat when I worked at a hospital. It wasn't quarter to quarter. It was actually month to month battles. It is always something you have to be vigilant with and follow. There is so much that changes within the operation that you just need to follow all aspects of it and the impact that it has to the organization. It is difficult for the hospital to look at these as a return on investment five years down the road. It is a hard sell. It is a very hard sell when you have got others who are also competing for the same dollars within a different department, who may want a new MRI and other technologies that they believe would have a more immediate impact.
MR. BLAIR: I suggest that we just take one more question and then have a break.
DR. YASNOFF: Can I follow up on that because I was leading somewhere with that.
We have discussed one potential mechanism to increase investment in information technology; namely, essentially making grant funds available. Is another potentially viable mechanism -- and these are not necessarily mutually exclusive -- to have increased reimbursements for increased quality of care? Is that another way that this -- that might increase hospital investments and information technology and other provider -- other elements of the health care system as well.
Whoever would like to comment on that.
MR. ARGES: Sure, that is another alternative and we tried that path recently with MedPac with respect to Medicare payments and in many respects it is hard to basically quantify the additional costs, especially in today's environment and then present it to health plans and other third party payers, that they need to step up the additional investment. But at the same time, it is an alternative, but I think what providers are probably looking for really is immediate grant money to be made available to basically jump start it.
Many of the reimbursement formulas have always been ratcheted down to the point where it is very hard to basically tweak the revenue side anymore.
MR. BLAIR: Simon, would it be okay if we take a break and then we --
DR. COHN: Yes, I think people need about ten minutes. I think we probably ought to just reconvene. Does anybody have any final comments? Okay.
MR. BLAIR: We want to thank the testifiers because we really had a wonderful spectrum, small but informative group. Very useful.
DR. COHN: We will take a ten minute break.
[Brief recess.]
DR. COHN: I think people should assume that we will be going, I presume, probably no longer than 12 noon, just on the basis that we are in sort of the last round up at this point.
Let's talk a little bit about what we have learned and sort of talk about next steps. I am not sure that I have the sort of view on this one. I know that Mike and Jeff, obviously, have strong feelings, but we probably just ought to go around.
I do want to sort of start out by sort of commenting that in our original recommendations for clinical data standards, we had made at least one very specific recommendation that had to do with more general financing support, et cetera, in relationship to trying to get things moving around this, recognizing that standards are one piece, but if you have no way to implementing the systems, standards become sort of an interesting but someone useless discussion.
I don't remember the exact wording of it but I do think that whatever we are doing, we may need to be looking a little more at this around the cost issues and the burden on the providers, et cetera, et cetera. I think I sort of agree with where Ted was going with some of the discussions, that we need to somehow as part of all of this, take note of the financial situation and if we can make recommendations to cope with that, I think it would be a good thing.
Those may be recommendations that, of course, are primarily targeted at Congress, but that is certainly within our charge to make recommendations.
MR. BLAIR: We could just pass the hat.
DR. COHN: That was sort of my general comment to start out with, but maybe -- Kathleen, do you want to make any comments about what we heard?
MS. FYFFE: Again, at the risk of sounding like a broken record, my understanding about the financial status of hospitals in this country is generally that a third of them are in the red, a third of them break even and a third of them are not in the red.
It troubles me that in an environment of decreasing reimbursement to hospitals in general, that we are trying to advocate and encourage the use of increased technology when the environment is one of decreased reimbursements. So, I think that if we as an advisory body to the Secretary can do anything to brainstorm and come up with various suggestions for this cost problem, I think that it would be very, very worthwhile.
As you might recall, I am the one that kept talking about the business case for all this and I really do see some -- my instincts tell me that there is a positive cost benefit analysis that can be done over the long run to this end.
So, to the extent that this could be discussed at additional hearings or additional meetings of the subcommittee, I think it would be very worthwhile.
MR. BLAIR: Just as a reference point, I don't think this should limit our discussion at all, but just as a reference point, I believe it was Recommendation 5, where we indicated that HHS should provide support for cost benefit studies for patient medical record information standards and Recommendation No. 9, which indicated that HHS should explore incentives to provide for equitable distribution of costs to help the provider portion of the health care delivery system in implementing PMRI standards.
Those are the two that I recall that related to costs. I think that there is no need for us to limit ourselves to those two, but those are at least, you know, on the record.
Let me know when our cost discussion ends and I think we should summarize, you know, our observations from our testifiers.
DR. COHN: This was really not meant as a cost discussion. This was just sort of what did you learn, what are your points. We just sort of want to make sure we get heard on this. If you have other things, you can certainly mention them. We are just going to go around and ask everybody on the committee.
MR. BLAIR: Okay. My only other thoughts that this morning's testifiers, in my mind, continue to enhance what we learned yesterday from the two panels. Again, I was really happy because the spectrum was testifiers that pointed out there are additional things for us to do and look at in terms of examining standards for PMRI -- selection of PMRI standards and caution on the other end of the spectrum with the fact that the provider community has limited resources and is pretty weary with respect to especially unfunded mandates.
So, you know, the sense that I had is that the approach of being very, very careful and proceeding in a manner where we are looking at incentives and being in line with standards that have pretty good market acceptance and could act as a catalyst and do so in a way that also can mitigate the implementation costs, at least what I heard was that that direction is reasonably on target.
In terms of our going forward over the next several months, I think that we summarized at the end of the day yesterday and I didn't hear anything this morning that indicated our conclusions were not correct.
DR. SHORTLIFFE: The making of -- as you know, I am new to this. I haven't memorized what numbers 4 and 9 were on an earlier report. I haven't passed the test yet. I am sorry. But you know, at its core nobody out there is opposed to the standards. Right? I mean everybody sees the potential benefit of getting together.
The vendors do, the hospitals do, other providers do. If they were informed about, even the patients would, I think. They would like to have their medical records more mobile when they change doctors, for example. That won't happen without better standardization in these areas.
But the resistance that we encounter and that we are hearing pieces of are somewhat related to logistics, not invented here syndromes, where people are feeling ownership of something that they have been working on as a standard and they feel challenged when they hear about an alternate way of thinking about it. In fact, sometimes that alternate way makes sense for whatever the other group wants to do, but not for whatever some other group wants to do, which is why we have so much trouble trying to rationalize the standards that are out there in competing organizations.
But, ultimately, I don't think you can ignore this financial part of this and it really has to, I think, be front and center as an issue and for us to say, well, we made recommendations about it before probably isn't enough. I think it has to be highlighted as a key aspect of anything that we come up with in the way of PMRI standards as some way of identifying the fact that there need to be -- it is not so much financial incentives, although incentives are part of it, but it is basically solutions to the fact that it will cost money to do the right thing.
Some of the people that need to do the right thing don't have any money. We want them to do the right thing and, therefore, fiscal reality has to be part of any recommendations we make. I don't want to lose sight of that. I think for us to just promote some particular standard, a solution, without addressing that would be shortsighted.
DR. COHN: Mike.
DR. FITZMAURICE: I heard a lot of things. I want to take issue with one thing that Ted said. I think there are some people who are resistant to standards or changes in the way that they do things. Try to change DICOM(?). Manufacturers have begun building that in. Try to change other standards where a manufacturer has got it built into equipment and you will get some resistance from those manufacturers, those vendors.
But that is also a strong reason to get to the standards table and to participate in standards that can work out for a broader range of equipment. I think by putting some seriousness into standards, by saying that the government -- or by recommending to the government specific kinds of standards, which ones are the leaders of the pack and keeping in mind that things have to fit together, it gives them a better reason for coming to the table to protect what they have and to make what they have more saleable in the United States and in international markets.
Some of the things I heard were we should start from models. I have an aversion to spending a lot of money on models, but a great likeness for phrasing things in terms of where they fit within a simple model. Data dictionaries, compliance and certification, standards require some of this. I heard a cost issue and the burden on the provider of standards has to be considered.
In our financial recommendations, we may want to consider targeting Congress, the case for some sort of action. I think we are well on our way on the patient medical records standards, a portion of what we do as the National Committee on Vital and Health Statistics. I didn't hear anybody say this shouldn't be done. This was not a good idea. I think current leadership is critical in accomplishing this mission.
I did hear that we need more tools to implement standards, things like Meta Data Registry, conformance testing with qualifications and to help with things that hospitals are already burdened with in their reporting and analytic burdens of patient safety, HEDUS measures, were two that were mentioned and a need for more research into the costs and the benefits.
What is information system satisfaction from the customer's viewpoint and what are the elements at the beginning of that process of putting in an information system or an improvement in an information system? What elements are related or associated with greater success and greater consumer satisfaction at the end of the installation and the actual implementation? There is very little that has been done on that. I can think of only maybe two articles that have been done on the process of installing an information system. If there are more, I would certainly like to see more of them.
So, we need more information but we have enough information to make recommendations on patient medical record information standards. I think having the open process and make some judgments, flying them by the public in open hearings like this is going to get us to where we want to go within a year. And that is our first set of recommendations.
DR. COHN: Do you have a comment on what Mike was saying or some other ideas?
DR. YASNOFF: I did, but let me wait until it goes around.
DR. COHN: Go ahead.
DR. YASNOFF: I did have just two quick things to add. One is that in addition to -- this is yesterday's testimony as well -- people seem to be supportive of standards. They also seem in general to be supportive of the process that we have outlined. I didn't hear anyone say you are on the wrong track. You need to go this other way. Obviously, people had specific suggestions.
The other thing I heard this morning, and I want to thank Helen, is this issue of transferring entire electronic medical records as a transaction. I think this is something that I haven't heard discussed before, but obviously is an important capability given the tremendous movement of patients from one health plan to another. This might turn out to be a very common transaction if it was supported. So, I think we need to give that some consideration, although, obviously, you can say that transferring entire electronic medical records is just a sum of all the individual transactions, but on the other hand, I think there may be other issues involved.
So, I think that is an important item for our consideration.
DR. COHN: There were two that she talked about. One was EMR exchange. The other was identification reconciliation.
DR. YASNOFF: I think the identification issue is something that we have discussed before and I might say has an obvious solution that we can't talk about.
DR. COHN: We can talk about it. We just can't spend money on it.
Kepa, do you have any comments?
DR. ZUBELDIA: I am going back to the unfunded mandate. I think that the -- industry is not used to have that long term view and it is more month to month and to implement one of those PMRI systems takes more than one month. But at the end, the reward that goes back to a provider in terms of more efficiency and so on and the word "grant" was mentioned. I am wondering if we should suggest grants or more a loan type of arrangement that -- and get over the hump of getting it installed and ready to go.
Then as the rewards -- I think that something like that may be more appropriate because I think that the provider is the one that gets the reward in this environment, although the government also gets the rewards. So, I don't know.
DR. COHN: I am sure if we take testimony, I am sure what people would say to us, given the options, but I think it is an interesting thing to put on the table.
DR. SHORTLIFFE: I wonder if we couldn't just have some hospitals declared disaster areas and they could get loans from FEMA.
DR. COHN: Marjorie.
MS. GREENBERG: On another topic that was raised, Simon won't be surprised that my ears perked up on the recommendations about data sets. Data sets are near and dear to my heart and actually the National Committee has at least a 30 year history working in the area of data sets. I do think that given the unlikelihood that the committee is probably going to be in a position to spend a lot of dedicated time trying to rationalize data sets, you know, advance this work too much in the short term future, I do think the suggestion and maybe a recommendation from the committee that some of the data sets that Sandy mentioned that aren't currently being entered into the Meta Data Registry should be.
I mean, I do -- it won't solve the problem, but I do think the Meta Data Registry holds a lot of promise for at least making it very clear where the differences are and maybe, you know, having some incentives moving towards rationalization where things are different, not for a reason but just -- you know, they are just different.
Although some of those data sets, I think, are being entered into the Meta Data Registry, I don't know that there are plans for some of the others, like OASIS and the other ones that are in the department. It is not a small job to enter them into that Meta Data Registry and to map them to the Australian model.
Actually, we can thank Suzie Burke-Bebee, who was sitting here yesterday, for doing that with the NCVHS, core health data elements. She was able to combine that in a project in her doctoral program as well. That has been done, but I do -- I don't know if you are all aware of that, that the core health data elements that have been entered in.
But I think that -- at least that is one thing that the committee could support.
DR. COHN: I had written down the USHIK piece and I was just thinking that perhaps that is a presentation that should go to the full committee with a discussion around that. Certainly, they were talking about standards, but I don't think they had really talked about data sets. Of course, they are all the same thing, but they tend to call them different things.
MR. BLAIR: Could I suggest that in terms of a presentation on USHIK and data sets, that in that same meeting, that we have a presentation also on the good electronic health record and on the reference information model? Because there is a lot of interrelationships with all three.
MS. GREENBERG: Did you say the good electronic health record?
MR. BLAIR: Well, it was called -- it was originally the good European health record and it has become -- they changed "European" to "electronic" when Australia took it over and there is a lot of work that has already been done in coordinating and sharing information at HL7 with the GEHR. So, that is why I am suggesting that I think those three topics together would be instructive.
MS. GREENBERG: You said the U.S. --
MR. BLAIR: USHIK, which is the Meta Data Registry.
MS. GREENBERG: Right. The Meta Data Registry, the GEHR and what was the third one?
MR. BLAIR: The GEHR and HL7's reference information model. It may be in addition -- the reference information model also has a subset called USAM(?), which I think also there should be maybe a separate presentation on that. If Stan is still here -- does Stan feel comfortable with that as a recommendation?
DR. HUFF: I think that is a great recommendation. The reason that I say that is that I think these things all are related and I think the case, for instance, for USHIK has been stated appropriately that it allows for some cross referencing and some open discussion.
The thing that occurred to me, though, is that you won't get convergence until you have -- on a model, until you have the stakeholders all sitting around the same table. In other words, it is easy to look at USHIK and say, oh, I am doing this data element. Here is a value set and that looks good, but as soon as what you look at doesn't -- you have questions and then the person who created that isn't present to represent their view, you just go ahead and do it your own way anyway. It is a lot easier to do that.
So, we won't come to a convergence of models until GEHR people and ASTM people and HL7 people, all of the stakeholders are sitting around the table. Then you can approach coming to a common information model for those things.
So, I very much applaud a session where you could talk about those things together, getting some understanding and then maybe make recommendations about how you can get all the stakeholders together to come to some convergence of modeling and what that might mean.
MS. GREENBERG: I think the idea of having it before the full committee is very good because I think it has a lot of implications for the work of the Subcommittee on Populations, too.
DR. COHN: I am thinking that Marjorie means she can help us put this together.
I think the idea -- I think I want to ask Jeff a little bit and maybe the others -- I worry -- and this is a slippery slope, a slippery slope into oblivion, as you move from data sets into data models and information models and then suddenly, you know --
[Multiple discussions.]
It is the only one I can remember the name of very quickly, but there are others. The question is is there a value to moving from the data sets into the data model issues? I mean, is there a value added to that particular step?
MR. BLAIR: Here is my thought because I don't want to prejudge our actions to this. I just think that it would be good basic education to the subcommittee if we understood the objectives of each of these different information models and data sets and the difficulties they are having and the issues where -- of the difficulties of trying to get them to either have common definitions because they have different use cases that are underlying them or different purposes.
I think it is good education and where it will lead after that education, I am not sure, but it relates to all the other things we are doing.
DR. COHN: Helen may be one of the few people that understands this. I mean, I am -- why don't we let Helen make a comment because she may actually understand some of this stuff. Then, Marjorie, you next.
MS. STEVENS: Thank you. I am listening to this conversation and the previous conversation about cost benefits and, you know, where do you allocate money to help make things better and I am looking at the amount of volunteerism that goes into the standards organizations at the moment. There is a massive amount of volunteering from all the different stakeholders to build all of these different SDOs.
If there is one thing that this committee can do to really make a quantum leap, I think it is to support bringing those stakeholders together, to support inter-SDO activities, like building an information model and some of the core underlying foundation to developing an SDO. As Stan mentioned, you need to bring the different players and the different stakeholders and the different SDO organizations together -- and this is a wonderful forum to be able to do that because it is so independent and actually use that to build the underlying -- the foundations and if you have got a solid foundation, a solid information model, building upon that becomes much easier.
MR. BLAIR: Could I add one other piece here because I think it is important to understand? And it gets to the question that you asked is, you know, what is the purpose of this. The information models, when I say they have different purposes, the reference information model is intended to be a common stable information from which new message format standards can be generated in the future.
So, whatever we do in terms of message format standards, if we have an understanding of the RIM(?) and the USAM, service access model, which is the clinical portion of it and the struggle that they have had to try to wind up getting consensus on this model, just within HL7 alone, yet the richness of what is there, I think it really helps us have a much better handle on medical terminologies, as well as data sets, as well message format standards.
MS. GREENBERG: I have kind of a procedural question here, but as I am thinking about the June meeting and the full committee and everything we currently have proposed, although we -- but one thing is, you know, Denise Koo was scheduled for the February meeting to talk about the public health data conceptual model and NED(?) and she was ill and wasn't able to come. So, our current plan is to reschedule her to June.
I just wonder if it would be good at the same time if she is going to be talking about the public health data conceptual model to have the presentation on the HL7 RIM. I don't know if we can also fit in the knowledge base at that one, how important it is that they be at the same time.
DR. COHN: Well, it is an interesting idea. I mean, the CDC effort has actually been one of the few that I understand, primarily because they go through such efforts that define every single piece and what they mean by it because one of the things about models is that everybody tends to define things a little differently.
This has nothing to do with HL7 because I have actually not reviewed the RIM in terms of all the supporting documentation. So, it might be a very good --
MS. GREENBERG: I know the CDC's goal is to rationalize or synchronize, whatever the right verb is, with the RIM.
DR. COHN: You are right. It might actually be -- assuming we have enough time to put this together, it might actually be a very interesting presentation and discussion. Actually, I think Stan has sort of nodded his head that it makes sense conceptually to do it together.
I think we have sort of put this as a place mark and we have sort of said "yes" here. We have said -- Kepa made a comment earlier about -- and we will talk about the sort of optionality making standards incompatible.
-- reconciliation, which are two things that Helen had commented about as, well, gee, things ought to be expanded and, obviously, we need to at some point figure out whether we add those to the list. I guess one of the questions I would have for Stan is when I look at EMR exchange, I think I see it as other transactions that we have identified in that list of transactions that we were going to send out.
Am I -- I am just going to ask Stan for a little bit of help on this one. Do you see that as something different or is it already included in that list?
DR. HUFF: It is, but the performance part, I think there is a different aspect to it. I mean, thinking of my own system, we would need to change some things to be able to in a couple of minutes move everything in somebody's patient record out to an external system. It would be -- it would ultimately be an accumulation of all these other kinds of transactions, but considering the performance of that and how to really do it has some extra things to it besides just saying, oh, just transfer the lab data and then transfer the radiology, et cetera.
We would have a hard time doing that in a high performance way right now, I think.
DR. COHN: Does that require a reference information model that you map to?
DR. HUFF: Yes. You need to -- well, you do need the information model and you need -- I mean, that needs to be shared between the systems and in a sense that is encapsulated in the fact that you agree on the transactions. It is probably also a reflection if you took the transactions that have been specified and compared it to what is in people's electronic medical records, it would right now probably sum to somewhere between 70 and 90 percent. So, there is this other 10 percent of stuff that is probably not listed as one of the transactions right now that we have. I don't know much it is for sure without doing some analysis.
But there are these other important pieces that don't have transaction names in the current list, I think, too.
DR. COHN: Helen, I guess I should ask you about this one because I think we are sort of struggling. Did you mean -- I mean, as Bill was taking it, sort of ship the MR from one place to another or were you thinking more about some of the more common constituent elements?
MS. STEVENS: The two examples I gave were a discharge summary, for example, patient discharge from the hospital system, sending a discharge, a summary basically of that visit to their primary care provider's office. History and physical in the opposite direction, maybe going from the provider to a hospital. In those cases, it is a subset.
But there is also the need to be able to send all the current relevant information on a person if a person is moving maybe from city to another, one provider to another and be able to transfer electronically that medical record. That doesn't mean to say that you have to resend every single lab result on that patient. Obviously, a lot of lab result values are now irrelevant six months after they were originally drawn.
That becomes more complex than just trying to resend every transaction that was ever sent on that patient. If the patient has been in the sending health care system for decades, that could be rather large. So, what information is relevant for the receiving system to be able to effectively treat the patient.
Certainly, the fact that they are allergic to penicillin, but what other information -- I am not a doctor, so I couldn't possibly speculate.
DR. COHN: Jeff and Mike, this is a survey. Can you ask a question like this? My own thought is that this is getting to be -- this is not a low hanging fruit any longer, as we are describing it, I mean, and I think we are probably -- I am intrigued, but I don't think it should be part of the questionnaire.
DR. FITZMAURICE: I wonder if it would be useful to ask them do you have an information model that you refer to in one question. Answer it in eight words or fewer.
MR. BLAIR: Here is my feeling. Okay? Because we had considered this once before and, you know, Stan is here, so I will be very explicit. Okay? I think we know that there are certain message format standard development organizations that are building reference information models, which will enable them to be able to produce more clinically specific message format standards more quickly, hopefully, in a more stable way with greater continuity.
It is an area of futures in my mind and I almost feel as if we have to separate that out especially if we are going to be talking about Version 2.3 or 2.4, that is not based on a reference information model. So, it is something that is important for us for future directions, but in terms of what we are talking about right today, my inclination is that it almost would, I think, be construed by many SDOs as something that maybe gives an imputed unfair advantage to one SDO over another and I think we really want to try to be fair and equitable and have a level playing field.
Given that I have made that statement, Stan, do you feel like I am saying something that is unfair?
DR. HUFF: No, I think that is fair. I mean, I -- trying to reconcile the information models would be an important step but it is a huge task. Certainly, I think you can accrue benefit by use of the standards without doing that step. So, I guess I am in favor of it, but I can understand completely, you know, holding that to a later date or -- in a sense, it may take care of itself; that is, if the information models do what they hoped we would do, then the people who use them will be successful and those who don't won't.
So, it will come out in the wash anyway.
DR. COHN: I think we are obviously not -- are we going to ask a question about information models or not? It sounds like not. Not at this point. Which is fine. I mean, we will have another session on information models.
Now, let me ask the other one here -- and I think Dr. Yasnoff made an interesting comment when we mentioned identification reconciliation, which does bring back some deja vu for some people who have been around this table. But there are obviously many different ways to handle identification reconciliation. I think that probably, Helen, I am not putting words in your mouth, but obviously a unique patient identifier is one way of information reconciliation, but there is also various mediation approaches. The issue is that there are lots of mediation approaches and I know that there has been a significant amount of work done over the years trying to deal with these sorts of mediation approaches.
Now, is this something that we want to -- I mean, it really actually is sort of a high value if there is, indeed, a standard that ought to be the mediation approach. Now, I don't know myself whether there is a mature standard out there or something that is generally used or something that ought to be generally used that would be something that we should recommend. But I just want to put it on the table.
Comments? Don't all get excited at once now.
DR. FITZMAURICE: I don't know. I can see that any large organizational use, whatever they want, invent their own number, but they will keep a file of all the other identifiers just to make sure they have got the right patient. Maybe it is a discussion that can wait until we really get into the patient identifying business, but if there is a need for it out there and NCVHS has something to say about it, other than we need a unique patient identifier, we can go ahead.
So, I am wishy-washy on both sides of the fence. I would like to get it done but we can't get it done.
DR. COHN: Which part are you trying to get done?
DR. FITZMAURICE: The patient identifier.
MS. GREENBERG: Are there standards that need to be advanced for the master patient index?
DR. COHN: That is what I am sort of wondering about. This is, I think, sort of where we are going with this --
MS. GREENBERG: Because people are identifying people everyday right now with and without -- without
these --
DR. COHN: They are taking, I think, as Helen has commented, multiple different identifiers or somebody coming in from the outside and is this the same as another person who is already here or different or whatever. You know, it is a fundamental aspect of sort of all the work we do. If Hammond(?) were here, you know, he waxes long and hard about this. Of course, his solution is the unique patient identifier. If you had one, you would still likely have a lot of patients that had two unique patient identifiers. So, it becomes an issue even if you have that solution.
Jeff, comment?
MR. BLAIR: If we included it among the scope for the questionnaire where that would be a transaction that SDOs could respond to, we could just simply see whether HL7 responds, whether KOBAMED(?) responds. In short, we open the door for SDOs to respond with a standard.
DR. COHN: And let's see what we find.
MR. BLAIR: Let's see what we find.
DR. COHN: Comments?
DR. FITZMAURICE: Would we ask them if they have a standard for patient identification or if they use a particular identifier for patient identification? Do they use a social security number?
MR. BLAIR: This would be a transaction, you know, right now because our scope is limited to transactions. So, what kind of transaction do they have for --
DR. FITZMAURICE: I think in all of their ADT information, that is what they use to identify patients.
MR. BLAIR: Well, I think that Sandy -- I think it was Sandy, who was saying that this is more explicit than just ADTs. Yes, you are right. They do, but she was winding up saying we should also separately list --
DR. COHN: Let's ask Helen to clarify what she was sort of talking about.
MS. STEVENS: I think there are two things at question here. There are probably several billion people in this country, who already have some kind of patient identifier. So, if it is a question of creating a standard identifier, let's say the social security number, then the question -- the problem becomes even more acute because you are just introducing another number into the mix.
What I was primarily targeting with my testimony was the need for a transaction to reconcile multiple identifiers. I am assuming that everybody has multiple identifiers. If anyone has been in the health care enterprise, you know, you have got your driver's license and your social security number just to start with. So, it is not so much of assigning an identifier, as in reconciling existing identifiers to uniquely identify a person.
MR. BLAIR: Is there an organization that has developed a standard to do that?
MS. STEVENS: To reconcile identifiers?
MR. BLAIR: Yes.
MS. STEVENS: No, not that I know of. There are several products on the market that use fuzzy logic to identify duplicate persons.
MR. BLAIR: My thought is that we are --
DR. HUFF: The way I saw this is -- I mean, there are transactions within HL7 that are specifically for merging patient identifiers and merging their transactions to support these actions. What I saw -- what I thought Helen was asking for -- and I would appreciate her clarifying again -- was asking for that to be a separate line item from ADT. You know, strictly speaking, ADT is just discharge transfer transactions and this other set are really for managing master member indexes and for merging duplicate records and those sorts of things.
That second class is -- I mean, if there is one place we have trouble within InterMountain Health Care, it is getting support for those transactions and that is the only way that you can keep records consistent across systems is -- that is the one thing that we have started putting in contracts is support for that set of transactions.
So, I would be very supportive in having that be a separate category and recognized as a second kind of transaction set separate from ADT per se, though it is sort of grouped in that group. It would make it more explicit and clear to people what we are talking about.
DR. COHN: Mike, does that help clarify what we are talking about?
DR. FITZMAURICE: It doesn't clarify what we would ask them. I mean, I guess, maybe it does. Maybe we ask them do you have a separate set of things that you would use to identify patients, other than the use of ADT information for admission, discharge and transfer. Is that the question you wanted to ask them?
DR. HUFF: I would say the question this way. Do you have transactions that allow merging of records in master member indexes and for merging encounters within encounter records, which is a different set of transactions that admit, discharge, transfer?
DR. FITZMAURICE: Simply asking do you have an algorithm that goes through and finds duplicate counts.
DR. HUFF: No, not an algorithm. Okay? It assumes that I have run an algorithm or the person presents and by whatever evidence, I think this is a duplicate. Now, that I have recognized the duplicate, I want to send transactions to multiple systems in my network that cause those duplicates to be consolidated.
DR. FITZMAURICE: So, it is changing variables out there so that everybody -- all the information looks the same.
DR. HUFF: I can tell another system. You know, I admitted two different people to you, one in November and one in September and now I recognize those are the same. Could you please merge those records so that they are now only one. That is different than admitting the patient or transferring them or discharging them.
It is asking for a consolidation of records in that other system.
DR. FITZMAURICE: So, do you have a transaction that reconciles duplicate patients?
MR. BLAIR: A standard that helps you to do that transaction.
DR. MEYER: Chuck Meyer.
What Michael points out is that third component, which we all struggle with, too, and that is that mediation part. What do you have internally -- this normally happens because to execute Stan's transactions for merging, somebody has already made a decision based on whatever information, that, in fact, these two are the same person or this account addressed the same person.
There now comes into play -- and at one point, we had a cig(?) in HL7 under NPI that was trying to deal with that mediation issue and it was hard for them to come to closure on what the appropriate data elements are that would allow that and it became even more complex when we got into the structure of privacy because most of the elements that they noted were the elements that were excluded, i.e., how you de-humanized the record by throwing those away.
So, there was the exchange issue and minimum requirements and all that. So, there is really the three components that are involved there. The reconciliation, the merger of the account and then the internal identification of how you determine that those accounts needed to be merged in the first place. I am not sure that lends itself to a transaction per se, although there may be some standard defined to say given these elements, you will have a percentage of probability that this is the same person, relates right back to the fuzzy logic products that Helen was referencing.
To a certain extent that many vendors implement in their own systems to say if I hit on five of these eight elements, then I have a high probability this is the same account.
DR. COHN: Okay. Ted did you have a comment?
DR. SHORTLIFFE: I have been getting a little confused by this. It seems to me that there are two activities. So, I am confused by the difference between the notion of reconciliation and the notion of merging. It seems to me there is a process, which is the process by which you identify two individual records that are, in fact, the same person.
That, it seems to me, is not particularly amenable to the standardization of the sort that we have been describing. But having done so, as Stan was saying, then there is a need to be able to merge records because you have decided that duplicates were created. That requires transactions, which could be suitable for standardization.
So, I would think that if we are going to ask a question, it would ignore the issue of identifying the duplicates, which is a hairy problem, not really likely to be what the SDOs are looking at and instead ask have you considered the transaction standards necessary for doing these mergers.
Now, have I missed something? Is there a difference between reconciliation and merging? You seem to be using those as two separate notions. I don't understand the difference.
MS. STEVENS: I do distinguish them because merging is combining two records. The final requirement might actually be to change an identifier if somebody came in and was assigned the wrong identifier, to change it or to split identifiers. So, merging is a subset of reconciling the identification --
DR. SHORTLIFFE: So, reconciling can be lots of different kinds of processes, of which merging is one.
MS. STEVENS: Right.
DR. SHORTLIFFE: After you have identified some kind of complex interaction between two patient records or one patient record that should have been two, I suppose.
MS. STEVENS: Exactly.
DR. COHN: Now are we clear enough -- I see Stan is nodding his head, which is good. Helen, are you nodding your head, too?
MS. STEVENS: I am nodding my head.
DR. COHN: Mike, how close are we to having your head nod?
DR. FITZMAURICE: So, we ask them if they have a transaction that reconciles patient information for two patients, who happen to be the same person. Is that the answer -- question to which you want an answer?
MR. BLAIR: I would just say --
DR. SHORTLIFFE: Isn't it true that there are only two possibilities? You have either got multiple patients who are the same or you have got one patient that really should have been multiple. Is there any other possibility beside those two that I have --
MR. BLAIR: Maybe it is a transaction that facilitates patient identification. That is a simple, generic, broad --
[Multiple discussions.]
DR. HUFF: I think the definition of "reconciliation" that Helen used is appropriate. As far as I can think, there are only the two cases that Ted said. You have either got something that looks like one record that really refers to two people -- I guess it could be two or more in there or you have two records that really should reference a single person or more. And do you have transactions that will cause either the merging or splitting as appropriate for those records?
DR. COHN: Okay. And actually rather than -- I feel that we are sort of grappling about this a little bit. I am going to suggest that Mike come up with something that -- along with Jeff --
DR. FITZMAURICE: The most confused person you ask to come up with the answer.
DR. COHN: With the idea being to circulate it to all of us to make sure that we agree to it before it goes out. You might want to send a copy to Stan and Helen to make sure we reach agreement before -- if we can't reach agreement on this one, we can't ask for this as part of the questionnaire.
I do think it is actually a valuable piece because in the world where you don't have a unique patient identifier, even when you do, there is still a need for -- I mean, at Kaiser Permanente, we have unique patient identifiers and a number of our patients have multiples of them.
MR. BLAIR: Let me just add one little piece here and I am not intimately familiar with this particular standard, but KORBAMED(?) has a standard called patient identification, the PID standard. I think we want to examine to see, you know, if we define this in a manner that enables them to also submit a response to the questionnaire.
DR. COHN: Sure. I think that is appropriate.
DR. HUFF: One clarification on that. The KORBA PID standard is -- its semantics are completely dependent upon HL7. So, all of the PID's fields and all of the PID's transactions are based exactly on the HL7 standard. In fact, it references the HL7 standard within KORBA.
So, it is just a different technology for implementing the same transaction set.
DR. COHN: Thank you.
Other comments? I am hoping that we have sort of -- as close as we can get to reach a conclusion on this one, I think we have reached a conclusion now. Are there any other comments? I mean, I know that -- I am hoping that we have pretty well exhausted the discussion.
Any final comments before we adjourn the meeting today?
[There was no response.]
Okay. Well, I want to thank you all and thank our experts for joining us and participating in this particular part of the discussion. We will adjourn the meeting.
[Whereupon, at 12:15 p.m., the meeting was concluded.]