Hubert H. Humphrey Building
200 Independence Avenue,
SW
Room 801
Washington, DC
DR. COHN: Why don't we get started. Could we have our first panelists come up and take their chairs, and will group members come up and join us here?
Welcome everyone to the second day of hearings of the Computer-based Patient Records Work Group of the Subcommittee on Standards And Security of the National Committee on Vital and Health Statistics.
Why don't we start out with introductions by the work group and staff, and then we can go on to the audience.
I am Simon Cohn. I am one of the Co-Chairs of the work group, a member of the National Committee with Kaiser Permanente.
MR. BLAIR: I am Jeff Blair. I am the other Co- Chair of this work group, and I am Vice President of Medical Records Institute.
MS. FYFFE: Kathleen Fyffe, Health Insurance Association of America, also, known as HIAA. I am a member of the National Committee and, also, a member of this work group.
DR. LUMPKIN: I am John Lumpkin, Director of the Illinois Department of Public Health, a member of the work group and the Committee.
DR. FITZMAURICE: I am Michael Fitzmaurice, Senior Science Adviser for Information Technology for the Agency for Health Care Policy and Research and Co-Lead staff for the Work Group for Computer-based Patient Records.
MR. GARVIE: Jim Garvie. I am with the Indian Health Service.
DR. FERRANS: Richard Ferrans, Consultant to the VA and Chief of Informatics, Virology Medical Center and member of the work group.
DR. GREBERMAN: I am Mel Greberman, FDA's representative to the work group and Associate Director for Medical Affairs in the Division of Small Manufacturers Assistance.
MR. MAYES: Bob Mayes, Health Care Financing Administration, co-lead staff to the work group.
MR. GELLMAN: I am Bob Gellman. I am a privacy and information policy consultant in Washington, and I am member of the Committee.
DR. COHN: We will go to the panel members and then go to the audience.
MR. TONKS: Jesse Tonks. I represent 3M Corporation.
DR. RUSSLER: I am Dan Russler. I am a physician with HBOC. My title is Vice President of Clinical Strategy.
DR. MIDDLETON: Blackford Middleton. I am a physician with MedicaLogic and I am Vice President for Clinical Informatics.
DR. TANG: Paul Tang, Epic Systems and Palo Alto Medical Foundation.
(Audience Introductions.)
DR. COHN: Thank you. I just want to do a quick review of the agenda and I want to particularly thank Mike Fitzmaurice and Bob Mayes and Bill Braithwaite who is not here right now for their help in terms of organizing the session, both yesterday and today, I think those of us who were here yesterday, all thought it was particularly useful and helpful to begin our discussions around CPR.
Now, in terms of the agenda today, we will be having two panels, this one and the one afterwards. After lunch the group will be in open session basically discussing sort of what we have learned, revisions of the work plan and next steps.
So, all are welcome to sit in on that discussion after lunch. With that, why don't we ask the first panel to start, and, Dan, I think you are up first, as I understand it.
DR. RUSSLER: Very good. Thank you. Again, I am Dan Russler, Vice President of Clinical Strategy at HBOC. Chuck is going to be putting up an overhead. I have one overhead. That is in some ways our central strategy, the central message I want you to take home. Chuck will, also, pass around, I will be reading my presentation, and we have got copies for everybody.
So, if it is easier to read along, I am going to try to keep this to 10 minutes, and I will skip over small parts, but HBOC is a vendor of software for both providers and payors in the health care industry with software products in over 80 percent of US health care organizations.
Both provider and payor health care organizations come to HBOC to seek software and services that will assist them in delivering health care of higher quality and lower cost.
Consequently HBOC takes the pragmatic view, also, held by others that systems and system standards are only valuable as they relate to delivering higher quality and lower cost health care as demonstrated in evidence-based studies.
Some of my comments may, also, be colored by my role as Co-Chair of the HL7 Patient Care Technical Committee. However, today I am not an official representative of HL7.
The first question from NCHS was regarding congressional instructions, and the answer from HBOC is that we interpret the mandate from Congress on health care standards to be analogous to the recent process used by the FCC to create high definition television standards for the broadcast industry.
In health care we are currently in the state where HDTV found itself a few years ago, early in the development of competing proposals. Like the HDTV industry we know that the business case is adequate for investment in communication of guidelines and patient record information by vendors of information systems only when political forces have settled the question of which proposals will become standard.
The second question is with regard to the delay of implementation, and I think there are several factors. Just to summarize them, one is the lack of investment by government, providers and payors, vocabulary vendors and all software and content vendors in standards that incorporate cost and quality data from evidence-based medicine into the vocabulary choices in information models that are used for communication between health care participants.
Remember that our business goal is to support the cost and quality of health care and the standards need to relate to that business goal or we won't be successful, and our customers won't be successful.
Second, the redirection of current vocabulary investments into non-vocabulary administrative functions of organizations sponsoring vocabulary standards. That is a concern, and I will explain later. The splintering of vocabulary standards organizations into competing business entities with legal copyright restrictions is a concern, and I will explain.
On the investment issue, today software vendors and their customers have the problem of translating current vocabulary efforts into tools for improving the cost and quality of health care. Much of the investment in vocabulary up to now has been aimed at capturing the expressiveness of current health care language rather than focusing on the vocabulary shown to be useful in evidence- based medicine in the improvement of cost and quality.
Efforts of many of the reference vocabulary standards organizations are similar to the efforts of an HBOC customer who created 41 terms for urine, for the color of urine. Unfortunately, the implementation in their institution of 41 colors of urine slowed every nurse's documentation of urine.
What that meant, in addition there were no physicians or nurses who used 41 colors to actually make decisions in medical care. Consequently, the expressiveness, the desire to include watermelon-colored urine and pink lemonade-colored urine didn't translate into cost and quality improvement because what happened is the cost of documentation went up, and the ability to spend that time on patient care went down.
So, that issue was a concern. Despite the future research and the future that might indicate that there is some medical, clinical relevance to watermelon versus pink lemonade, you know, that is the concern, and expressiveness as a future research tool is a concern for the practical customer in the trenches.
Our hope at HBOC is that government payors, providers and all vendors will share the cost of mapping future reference terminologies to the level of expressiveness shown to be useful in evidence-based medicine rather than leaving that mapping to the software vendors and their customers as occurs today.
On regulation, vocabulary standards development belongs in the private sector in non-profit organizations. However, the administration of vocabulary standards organizations suffers the same pitfalls as other charity organizations where investment in the charity may not translate into actual value for the intended target of the charity. Some measure of regulation and audit of these organizations by government is necessary in order to protect investments of government payors, providers and other vendors in the resulting vocabularies.
On copyright, the issue that I mentioned, where can copyright interfere with the quality of care? For example, the use of proprietary vocabulary even by non- profits may limit the ability of a health care organization to receive patient care information in emergency situations from other health care organizations. If the receiving institution has not paid the licensing fee to read the information in the codes the receiving institution may not legally be able to turn the coded information into proper care for the emergency patient.
So, that is a concern under the current copyright because we have to provide or that receiving organization has to legally buy that, the interpreter for those codes in their information systems before they are allowed to read them. Does that make sense?
Okay, regarding private sector, as noted earlier, government will need to participate in the development of standards on health care communication to the same degree the FCC has already participated in the recent development of standards for HDTV transmission.
Government may eventually need to choose between competing proposals, but until these competing proposals can be developed government could greatly help in the consensus building process. As steward of the patient care classes in the HL7 reference information model and Co-Chair of the HL7 Patient Care Technical Committee, I am responsible for carrying recommendations on the changes to the classes, including the appropriate values or the attributes in each of those classes to the harmonization committees at HL7.
For example, one of my classes is the service event dot health issues class. That translates for most of you into the class that includes diagnoses and other problems. The values for that class exceed 80,000 terms. That is going to be a lot of work for my committee to come to the consensus of which terms in various reference vocabularies may be appropriate to the attributes in that class.
I could certainly use some government assistance in supporting that consensus work. I have to mention that I cannot make that request for HL7. I am not an official, but would be glad to carry that communication back to HL7.
Regarding which standards, HBOC, as mentioned is the largest vendor of health care software in the world. Accordingly HBOC attempts to participate in all the ANSI- approved health care standards organizations. However the largest investment in standards by HBOC to date is in resources to support the HL7 and the X12 standards. HBOC depends on HL7 transactions to support communication between computer systems in most of its health care provider customer base, and just to remind you we are already in 3000 hospitals. If we actually close with McKesson(?) we will be in 5000 hospitals in the US, as well as the majority of the payor organizations, as well as a large percentage of the physician, and we are also, the largest home care provider.
So, HBOC, also, supports the developing partnership between HL7 and X12 for communication of claims attachments between payors and providers. However, without the coordination of information models and value sets between HL7 and X12 and the information models and value sets of other standards organizations, the business case for increasing investments in standards-based development is weak.
Like other manufacturers HBOC needs to invest in goals that can be realized in the near term. HBOC cannot raise the cost of current products in isolation from the rest of the vendor community in order to fund risky strategies that are not shared by other vendors, payors and providers.
On the other hand, health care standards in place that are supported by government payors lower the risk of investment by HBOC in standards-based development, and then the last question is with regard to the four focus areas and the appropriateness.
HBOC has connected more disparate computer systems than any other vendor or health care organization in the world. Along with others, we have found that the factors that must be considered in unison are work flow in each computer system, specifically trigger events identified by use case analysis, information models of the sending and receiving systems which must be compatible, vocabulary values for each attribute in the information model and finally and probably least important message formats; whether they are HL7 or whatever message format you want to use depends on almost identical work in the first three categories.
In summary, we have found along with other messaging experts that messages as noted in focus area No. 1 are dependent on the information model and vocabulary noted in Focus Area No. 2.
Consequently, combining Focus Area No. 1 and No. 2 may improve efficiency of analysis. A second alternative would be to complete focus area No. 2 prior to beginning which again, the message formats are No. 1, prior to beginning the work of No. 1. So, if you serialize it that would work as well, if you want two focus groups as defined.
A third alternative would be to clump 1 and 2, but identify focus areas in the area of facilities location modeling, health care participant modeling and clinical and administrative event modeling.
We agree with the proposals recently published by the ANSI Vocabulary Standards Board Vocabulary Working Group but recommend that vocabulary standardization not be separated from information model standardization since these are inseparable.
In summary focus areas 3 and 4 are nicely identified. As noted above government regulation should help ensure that investment in non-profit vocabulary organizations will be money well spent and provide an adequate funding model for vocabulary development and maintenance from which all providers, payors, vendors and reporting agencies will benefit.
DR. COHN: Dan, thank you very much.
Paul?
DR. TANG: Thank you. Good morning, and thanks very much for inviting me to testify in this very important NCVHS activity.
I know most of you already but since I last spoke to the full group I have had a change in jobs, and I would like to speak from my new responsibilities today.
I am currently Vice President of Epic Research Institute a Division of Epic Systems Corporation. Epic is a developer of health care information systems including computer-based patient record systems.
I am, also, Medical Director for Clinical Informatics at Palo Alto Medical Foundation, a 200-member multispecialty group practice. Palo Alto Medical is an affiliate of Sutter(?) which is a large integrated health system in California.
My role between these two organizations is really to bridge the need of the health care provider and the tools that are developed by information technology suppliers, and in this role I hope to accelerate the transfer of technology to help docs take care of patients.
I look at your work group as trying to accomplish a similar objective, that is to accelerate the pace with which information tools are applied to help the efficiency and effectiveness of the US health care system.
My interpretation of what HIPAA legislation asks of NCVHS is really to provide a set of recommendations for how government can accelerate the development and use of computer-based patient records to apply to patient care, to research and to public health.
The legislation talks of standards for data, and when I look at the panel of witnesses yesterday and today, it is clear that the individual data standards are well addressed. So, for my contribution I would like to focus on a broader approach to the term "standards" and concentrate on the role of the Federal Government.
I do believe there is a definite role for government in bringing about the routine use of CPR and the data standards necessary for them to interoperate, and let me suggest three potential roles. One is to set the bar aligning the incentives and jump starting the process.
The first approach is for government to set the bar for health care. That is for providers, for people who receive care, for people who pay for care and for people who measure care.
The common denominators for health care are quality and cost, and government can exert its influence and effect a change in all sectors of health care by influencing one of these two factors. So far government has been emphasizing the cost side, mainly by setting reimbursement rates or by conducting billing compliance audits, both of which have had very far reaching effects.
Since I am a firm believer that if we measure and improve quality that costs will be reduced, I am going to concentrate my comments on the quality side. As a public health trustee government could require or cause to be required reporting on quality using standardized measures.
Now, obviously standardized reporting will require standardized data and their exchange, but I think that a mostly private sector effort in collaboration with government can work out the details for the individual standards provided that they are driven by a time line such as by legislation or by reimbursement policies.
As in any human endeavor deadlines are necessary and good. I believe the effect of national standards for quality measurements will cascade down to the providers by giving them a tool to attract and retain consumers in a way that benefits the health of the population by competing on quality.
For example, General Motors demonstrated the market power of providing consumers with data by setting HMO performance expectations and measuring health plans on meeting those expectations. Using a vigorous review program including site visits to the plans GM scores the HMOs based on quality and cost weighted about 50/50. They use the scores to set the employee contributions to the plan. They post the scores on their employee web sites and employees use those scores to choose between plans.
Year by year GM employees have migrated from the lower scoring quality plans to the higher-quality plans. Using standardized data, then consumers, the ultimate beneficiaries can reward provider groups for their efforts in service and in information.
In effect I am proposing that government set precise quality reporting standards and let the needs that those standards drive accelerate the public/private process for creating specific data content standards.
The next way government can effect change is through incentives. Probably one of the biggest impediments to the rapid deployment of CPRs is the fact that the cost of implementing CPRs is not borne by all those who benefit from the data, that is the risks are not well aligned with the rewards.
Let me illustrate with an example. It is well known that giving flu vaccine to patients or consumers 65 and older decreases the admissions, reduces the mortality and reduces the cost by about half in all of those cases. Yet despite our best intents and our perceived exemplary execution by health providers we immunize about half of the eligible population.
We, also, know because of work by Dr. McDonald and others that if we were to remind providers at the point of care we all comply with our own intents, and that makes a difference in all those metrics that I had cited.
What is wrong with this picture? Basically primarily the primary entities that pay for the system, say for example, the capital-enabled hospitals or the capital- challenged provider groups don't necessarily directly benefit from the savings achieved by this simple intervention.
In the hybrid health care reimbursement system we have neither totally fee for service nor totally capitated hospitals largely live under a discounted per diem rate which is not a system that rewards fewer admissions.
We, also, do not have a well-defined mechanism for some of the major beneficiaries such as payors to share in the cost of implementing or maintaining CPR systems which house the very data that we all depend on for all of our various reasons.
As the largest payor in health care as Bob mentioned yesterday, government could provide incentives for providers to gather and report on their compliance with quality standards. So, sharing the cost of implementing and maintaining CPR systems among the broader segment of beneficiaries I think would accelerate tremendously the dissemination of this pool and hence the collection of the data that we desire.
A third way government can accelerate the pace of change is just to jump in and do it. Perhaps nothing works better to get real work done real fast than a real project with real hard deadlines.
For example, I look at the GCPR project really as a standards acceleration project. It was established in response to a federal mandate, in this case a presidential directive to develop a comprehensive long-life, long-record for those who serve in the military or who have served in the military. To execute this directive the VA, DOD and Indian Health Service have established projects to accelerate the development of standards that would allow their systems to exchange data.
I think it is this project-driven nature of the standards development effort that helps focus its scope on the needs of the application and hopefully speeds the delivery of useful standards.
Having funding for the project doesn't hurt either. The GCPR project is a good example of a goal- oriented project that is driving the creation of standards in collaboration with the private sector and again, deadlines and funding are enablers.
In summary I believe that the sooner we put credible, meaningful and yes, standardized data in the hands of all beneficiaries of health data, including consumers, the better we can deliver appropriate care to patients, and I see an important role for government in catalyzing this flight to quality.
Government could set the bar for quality with precise standardized reporting requirements, could implement mechanisms to reward collection of quality-driven data and initiate and fund projects that require the use of data standards.
I applaud your efforts to accelerate the deployment and use of computer-based patient records. Like you I believe that routine use of CPRs can fundamentally change health care in this country by facilitating the delivery of patient care, creation of new knowledge and the development of rational public health policies, and I thank you very much for the opportunity to provide input to your process.
DR. COHN: Paul, thank you. I guess Thomas is not here. Jesse Tonks?
MR. TONKS: I appreciate the opportunity to participate in the presentation to this panel. My responsibilities have been over the last many years as a research and developer. I represent 3M Health information systems. We develop software used in the coding and classification of clinical data and for the outcomes analysis and storage and clinical data repository.
As to the first question, our interpretation of the congressional directive to this Committee is to narrowly focus on the need for data standards required in the patient medical record and in the electronic exchange of patient medical data, where there are other areas of focus that should have a higher priority than these and probably could provide a greater return on investment, but we believe that the assigned problem is nevertheless a real and deserving one.
Winston Churchill once said that Americans eventually do the right thing after they have exhausted all other alternatives.
(Laughter.)
MR. TONKS: I believe the instruction from Congress to the health care information systems industry is to quit fooling around and do the right thing.
Most other industries such as manufacturing and financing have long been able to automatically exchange information between different computer systems and do it throughout the world. In our industry we spend hundreds of millions of dollars that could be better spent elsewhere writing custom interface software and programs to exchange information between computer systems, in many cases even systems written by the same vendor. If you detect a note of frustration in my voice, then you are listening carefully.
As to question No. 2, there are several inhibiting factors in the development of uniform standards to achieve a plug and play exchange of electronic information. Three layers of communication must be compatible between the sending and the receiving systems. The lower levels of topology and protocols have standards that are widely adopted across most industries, and this is not where the problem lies.
The middle layer of defining what fields are in specific types of message and the structure of the message is a health care industry issue, but progress is being made in this area through the combined efforts of HL7, the Andover working group, Corbin(?) Med, Dycom(?), X12 and others.
The higher level of achieving a common understanding of the data term or code that is placed in each data field of a message is the most significant lacking factor preventing plug and play interoperability.
Historically it has been up to the analysts and sometimes the programmers on each application project to define the meaning of codes used in a field or the form that a term takes when it is placed in that field.
The definition of content is not a simple problem in the medical domain. There are issues of granularity, synonyms, context, relationships, as well as meanings.
Contents specialists can only resolve these issues. However, there is no single organization in the private sector that has all the required expertise in every discipline of medicine to create a comprehensive lexical dictionary with the motivation or incentive to develop it for the common good. I don't believe a standard terminology reference model will occur if left to the private sector alone. It will take some government support and action.
It has been said that good judgment comes from experience and a lot of that comes from bad judgment. In the process of implementing CPRs with some of our larger health are customers we needed to have a super clinical dictionary that contained all of the clinical concepts that would be stored in that clinical database.
There was no single source that we could turn to for this information. For problems, diseases, microbiology results and body parts we went to SNOMED(?). For lab results and vital signs we went to LOINK(?). For medication compounds we went to ULAMES(?). For clinical drugs and generic drugs we went to First Data Bank and Micromedics.
For procedure codes we went to the AMA. DRGs, fortunately, were in public domain. Some of the ADT and demographic codes came from HL7.
In every case there were, also, many concepts that our customers needed that were not available from any reliable source and we had to work with our customers to define them.
This effort must be duplicated at some level by every implementer of a CPR. As a vendor we do not want to be in the content definition business nor should we be.
For question 3, it will take a concerted effort by many organizations to resolve the interoperability issues associated with the CPR in the exchange of medical information. The government needs to play a leadership role in this process by taking the following steps: No. 1, define and mandate standards needed for government programs like DRGs, CPTs, ICDs. However, it is time to move beyond the analysis of encounters to episodic groupers and other standards that will aid in the management of patient care over the lifetime.
These standards should be made available as public domain. When these are given solely to any organization they become a monopoly.
No. 2, contribute input to industry standard organizations such as HL7 and the NCITS Technical Committee, FL8 that is working on medi data repositories across all industries.
The input needs to go through the consensus process with other inputs to achieve a common standard. The last thing we need is a commercial standard and a government standard that are different.
No. 3, fund research efforts in support of standards organizations and provide a forum for all of the major health care standards organizations to collaborate with each other and to coordinate efforts.
No. 4, implement industry accepted standards and mandate them and all government projects and contracts and where possible at least encourage if not mandate them in the public sector.
The role of standards organizations should be to define specifications for standards through open participation and consensus, and they should collaborate with each other to eliminate duplication of effort and conflicting standards, and that is where the government should facilitate a forum to accomplish that.
The role of the lexical content organizations should be to, also, collaborate with each other to eliminate duplication of effort and conflicting standards, again, supported by government sponsored forums. They should make publicly available for a fee the content that has a high quality assurance standard and is maintained for completeness and accuracy.
The role of the vendors should be to create software that implements the specifications defined by the standards organizations and develop applications that add value to the lexical content provided by the content organizations.
As to question 4, many of the required standards are in the process of being developed. HL7 is defining the reference information modeler, the RIM, that should become a standard for industry and exchange of information between government and commercial enterprises.
The government should provide input to the definition of the RIM to ensure that it meets its needs. However, the government should play a more significant role in defining a reference terminology model that would be used by the health care industry.
Keep in mind that we are talking about standards for exchange of electronic messages. This would not be a requirement for the RIM or the reference terminology model to be internally used in every system. To be compliant any system would only need to translate its internal information model and terminology models to the standard to be able to send or receive messages to or from any other compliant system.
Reducing the interoperability problem to a single interface for each system would save the industry millions of dollars.
Question 5, the four focus areas are pertinent to the charge directed to this Committee; however, the government has an opportunity to establish the reference terminology model and provide the services needed in its support along the lines that have been proposed to the HL7 committees.
The government should do the following: Fund the establishment of a terminology service bureau to create a library or reference terminology model of the medical concepts each containing an identifying number that is mapped to all of the HL7 approved content, coding and nomenclature provider terms in code sets such as SNOMED and LOIN.
This library should be closely linked to and support the HL7 specifications. The service bureau should have a VAR(?) contract with the content provider organizations to pay them for their required fees, for ongoing support and maintenance of the content within the vocabulary model.
Any vendor or provider organization could then subscribe to the terminology service bureau for updates from the reference terminology model and library as needed to meet their needs.
This would create a single standard lexical source requiring developers to sign only one contract that would provide the plug and play content for all of the HL7 messages. An ideal pilot for the creation and testing of such a service bureau would be the GCPR project. If it became successful in support of this component-based architecture serving the GCPR partners, it could then be expanded in its services to other customers in the private and other sectors.
Thank you for the opportunity to present to you.
DR. COHN: Jesse, thank you.
DR. MIDDLETON: Good morning. My name is Blackford Middleton. It is a pleasure to be here. Thank you for the invitation to participate and testify.
I have copied a couple of materials that should be before you. One is a set of handout slides, and the other is a set of appendices which will support the slides.
I am currently Vice President for Clinical Informatics at MedicaLogic, and I am, also, a clinical assistant professor of medical informatics and outcomes research at OHSU, and I am an attending physician in part- time practice at Providence St. Vincent's Medical Center in Portland, Oregon, and I am Chairperson-Elect for the Computer Based Patient Record Institute.
I am speaking to you today on before of MedicaLogic as a representative of the vendor community. Just allow me to give you a brief background on MedicaLogic. What we do is implement electronic medical records or computer-based patient record systems, and we will be one of the companies like the others on the panel with me here today most affected, perhaps by the standards and recommendations made by NCVHS to the Secretary and Congress.
MedicaLogic has been vending electronic medical records since 1985. The first record was produced in 1983, by the founder of the company, Mark Leavitt. A second generation product was produced in 1995, and that is known as Logician.
The point here is that Logician is an open architecture, a standards-based electronic medical record based upon the Oracle relational database management system, TCIP protocol communications, 2R3T architecture and browser based or C-post(?) desk top.
In addition, we currently use medical informatics standards such as the HL7 messaging system in representation of clinical concepts using ICD, CPT4, SNOMED, ASTM, E31 standards, LOINK and the NDCGPI.
We have experienced extraordinary growth in this company and the point is that the growth, I think, has been largely attributable to the use of these standards in enabling forced standards in creating content in software that is usable by a wide variety of customers. The informatics standards we use are detailed further on the following slide.
You can see for the various data types maintained within the record diagnoses, historical elements, clinical findings and symptoms and signs, allergies, advanced directives, etc. For all of those we use what standards we can as they exist now from the standards community, and we participate in the development of maintenance of these standards in each case.
Let me point out some desiderata, I think for our work globally. I think the first thing that a computer- based patient record has to do is support or enhance the doctor-patient relationship. This is the cornerstone still of health care in this country. Despite the fact that one- third of Americans are seeking alternative care or care from alternative sources, the doctor-patient relationship still accounts for the majority of care.
The systems we create should improve health care delivery efficiency, quality and service and at appropriate cost, not necessarily reduced cost but at appropriate cost for services delivered and value received.
I think we should empower the clinicians with state-of-the-art tools to support clinical information access in clinical decision making. It is well known that my own CPU is very limited in many ways in remembering all the facts and figures, details, protocols and guidelines that I should remember when I see patients in my part-time practice one-half day a week, by the way, just enough practice to feel incompetent all the time.
So, I am particularly dependent on these kinds of systems. I think, also, we should have as a desiderata the ability to exchange longitudinal complex records between two disparity MRs. The state of the world today is that there are likely to be many different EMRs or clinical systems within the context of a CPR that need to exchange information or in fact entire records.
Lastly, of course, we need to assure the secure access and confidentiality of these data and to allow access moreover to them remotely as patients oftentimes choose not to become ill conveniently in your office and may do so remotely.
I would like to point out just what is happening now at a macroscopic level and then dive into a micro- economic analysis of the potential benefits with CPR and then turn to the questions put forth in your letter.
First, on the next slide you will see a description of opportunities for annual cost savings with health information technology as estimated by a venture capitalist firm, Funk and Zeigel(?) nevertheless perhaps at least within an order of magnitude on the money.
The different categories along the bottom are electronic data interchange, better resource management, redundant test and excess paper reduction, medical adverse event avoidance, reduced diagnostic testing and decrease in length of stay in hospital.
You can see in the bars the net savings for each item on a cumulative basis and then the cumulative line indicated above. So, you can see that with EDI we might expect to save on the order of 20 billion and with reduced diagnostic testing and decreased length of stay perhaps we might cumulatively save up to 250 billion.
The point here is that I would like to propose a macro-economic cost/benefit analysis of this technology in health care. If you agree that US health care is currently spending approximately one trillion of the net for health care delivery services rendered approximately 15 percent of GDP and agree within an order of magnitude the potential savings are 250 billion or thereabouts, and if you assume that industries in the health care delivery marketplaces are spending approximately 8 percent of their operating budget on HIT which by the way is very high, therefore, we might find that a benefit cost/ratio would be approximately 250 billion benefit for 80 billion spent or approximately a benefit/cost ratio of 3.13. That is even with these gross numbers and even with a large allowance made for sensitivity analysis for every dollar spent we might expect approximately $3 in savings.
I don't believe this is too far off the mark. In fact, the academic literature would support comparable estimates of benefit/cost, and I have some data here I would like to share with you from a micro-economic analysis of an individual clinic, one of our customers.
The Family Care Practice of Concord, New Hampshire is part of Capital Region Health Care. Twelve providers are in practice there with providers and staff seeing approximately 1200 patients per month.
They see these patients under a plethora of different health care management plans, that is managed care, seven companies, commercial insurance, Medicare, self- pay and Medicaid reimbursement mechanisms.
In their implementation of EMR they spent approximately $90,000 in the first year for hardware, software implementation, etc., with interfaces, by the way, relying upon those standards like HL7 and representation to allow information to be communicated between systems. They experience annual support costs on the order of $37,000 which cover the hardware-software upgrades and people to make the machines work.
So, their total year one costs were on the order of $124,000. Their savings based upon this implementation using the standards we have been describing were profound. In the first year they saved $43,000 after adjusting for initial dips in provider productivity. Forty-three thousand dollars were saved in transcription. Chart pulls were eliminated in toto saving 24,000. Prescription writing saved another 71,000. Coding research and assignment saved almost 6000. Lab filing went away of course, because the results were reported electronically. Five thousand in paperwork reductions and total benefits of $158,000 approximately.
In addition, there were non-quantitative benefits assessed. Qualitative reporting capabilities improved because they could detect, for example, the true rate of preventive care services and in fact discovered that the practice was eligible for a quality incentive bonus, a common feature now of many physicians' reimbursement packages. If I do the right things at the right rates for all my patients, I will experience a quality incentive bonus.
In fact, the practice was able to go to the insurance company and detail to them their screening rates for various expected preventive care services and the like and whereas the insurance company wanted to reduce the capitation they were able to argue for 110 percent capitation in the new contract.
So, that was very helpful. Drug recalls, of course, patient letters can be sent immediately if the coded information of that medication is on line. Hospital inpatients benefit as well. The EMR is available in hospital. The patients do, in fact, enjoy their doctors using computers. The data here are also, consistent with some of the academic data showing that patients do not fear their physician using a computer. In fact, most patients like it when their physician resolves it certainly at the time of the visit using the computer system.
So, in a gross macro-economic analysis here the total benefits experience in this clinic were approximately 158,000, total expense in the first year 37,000, net benefit 121,000. So, the year one benefit/cost ratio is 1.27, and the ongoing benefit/cost ration is 4.27.
The issue was raised yesterday about ROI, and there are two ways to address the ROI. One is in the manner I have done so, to give the data in some reasonable way, either a macro-economic or micro-economic basis to ally these concerns.
The second is to say that ROI doesn't apply. These are infrastructure improvements to the practice of medicine wherein RO1 cannot be done because it is much akin to an ROI of the telephone system, and it abides by the disruptive economics of Metcalf's law, that is more EMRs that are in place the utility goes up by the square of the number of units on the network.
So, how do we interpret the congressional instructions? I think the primary goal as I understand it reading the legislation is to administratively simplify the health care delivery and reimbursement process in this country, to improve the efficiency and effectiveness of the health care system by standardizing the electronic data interchange of certain administrative and financial transactions, to establish penalties for non-compliance, to protect the security and privacy of health care information by setting standards and through this process hopefully to experience some order of benefit, whether it is the WEDI(?) estimate of 42 billion or other estimates in the literature.
I believe the HIPAA mandate to HHS is here, as listed here to adopt national standards necessary for the efficient electronic administrative and financial health care transactions, that all health plans including government health plans, all clearinghouses and those providers who choose to conduct their transactions electronically are required to implement the standards, to provide privacy legislation and to provide recommendations for an electronic medical record.
Further, I think the mandate includes the provision of standards for the nine EDI administrative transactions which have already been outlined and provide standards for COB, coordination of benefits, to provide standards for unique identifiers, including the allowed uses of these identifiers for individuals, employers, health plans and health care providers and to provide standards for security, confidentiality and electronic signatures.
To the second question, what factors or issues are preventing or delaying the development and widespread implementation of uniform standards for patient medical record information and its electronic transmission, I think we have heard over the past day and one-half many comments about the scope and magnitude of this problem and the inherent complexity of medical narrative and clinical description.
I won't go into that much more. The other thing I would like to point out though, and it has been mentioned this morning is that there are many stakeholders in this industry and the stakeholders as Paul Tang has just alluded to oftentimes have unaligned incentives, that is he who pays for the medical record oftentimes is not the one who experiences the first benefit, and this misalignment of incentives makes coordinated decision making, of course, in the customer base extremely difficult for us vendors.
Secondarily, I think the health care standard setting process is currently grossly uncoordinated. There are many different standard-setting efforts under way, and they are not coordinated. There is not a big picture that is easy for a vendor to grasp onto or abide by.
I think the vendors are guilty of poorly implementing and coordinating the implementation of systems and the home-grown systems suffer the same problem, that is any academic institution building its own system typically does so in a haphazard patchwork kind of way.
Lastly, I think the professional societies have begun data standardization efforts for their own purposes, but here again these are largely uncoordinated, and I would point to the orthopedic societies, the cardiologists, the hematologists and oncologists and the nephrologists as exemplary in this regard.
With respect to question 3, is the private sector able to address these problems, what is the role of government, etc., how might the government help to improve the standards processes, I think there are three counter answers. The first is why can't vendors do it themselves. The second is why can't providers do it themselves and why the government should assume or should the government assume a leadership, an expanded leadership role in this field?
I have alluded to some of the problems the vendors face in addressing the standards issues. Health care standards are different in character and scope than standards developed in other materials industries, manufacturing industries and the like.
The tangled web of interrelated standards setting efforts makes broad-based progress very slow. Vendors are guilty oftentimes of delivering on-off solutions of their products and even when the vendor has multiple solutions within its suite of products, these products oftentimes may not communicate or interoperate one to the other in a seamless fashion.
I think it is possible for the US national standard-setting efforts to establish standards in a reasonable way for those standards I will call hard standards, things that are more or less quantitative in nature, fairly well specified, not overspecified like the Isostac(?) protocols in HL7 and the like, but it is more difficult or even impossible for the, what I will call soft standards, things that deal with terminology, things that deal with the information model wherein one's perspective of course is inherently unique and the multiple perspectives of course make it very difficult to describe a common model in a rational way.
I am reminded of the old joke about the three blind men. Sorry, Jeff, but it is a useful analogy, the three blind men describing the elephant each of which is feeling one part of the animal. They come up with very different animals indeed.
Why can't providers solve these problems? They are not SDOs. They are not software developers despite their grandest claims, and I have been in some of those institutions. We have yet to see the successful commercialization of home-grown systems from those kinds of places.
With rare exceptions, provider organizations have little concern for data sharing issues beyond their home turf, and they are often limited in scope or relevance to their business mission, that is few provider organizations truly cover all aspects of the health care delivery needs.
Why the government should assume an expanded leadership role in this field? I have tried to allude to the bigness of the problem. I would, also, like to point out that I believe health care informatics standards of others have said here are a public good. Not all standards, necessarily are a public good, but some of them are a public good which require federal intervention and leadership and guidance.
There are many stakeholders, unaligned incentives, potentially forever at odds. These informatics standards we are talking about are critical to the health and well-being of the American people and the American economy. Let us not kid ourselves. Approximately $750 of every automobile goes to pay health insurance costs for the members of the employees of that automobile organization.
I think, also, the leadership void first identified in the 1993 GAO report remains unfilled in this area.
Question 4, which standards related to patient medical record information and its electronic transmission would add the most value for improving the quality and efficiency of health care for the nation? Why? First I would like to simplify the goals. I would suggest that we simplify the goals to the bare minimum with respect to identifiers, terminology, data sets and models, interoperability, connectivity, security and confidentiality.
I think oftentimes we have been criticized from the outside for trying to solve everything at once, and I have learned as a vendor to not let perfection be the enemy of the good, that is we can do a lot as the case study I alluded to with simply implementing basic standards to allow basic interoperability in systems to support work flow improvements and improve clinical information management today.
Secondly, I believe we should establish a framework for comprehensive health terminology systems and a trusted agency to act as responsible parties for coordinating the central health information technology standards development process. I would like to quote Shute, Cohen and Campbell at al in the recent Jamia(?) who have set forth a nice description of what a coherent framework for interdevelopment of related standards could be and, also, point toward the need for a trusted agency to spearhead and guide some of these efforts, not necessarily to do it, but to spearhead and guide and establish requirements with the standard-setting bodies which are in existence.
The critical issues for standards in informatics that I see are listed on the next few slides. First, let us not forget the educational and training problem that we face in this country, a perception in this country that is working against us, that is patients and oftentimes providers despite their technical sophistication in health care have a great underlying fear sometimes of computing and the sharing or representation of clinical information in computing systems and some of that I think is unfounded. Some of it I think is well founded. Nevertheless, we need to address the concerns and the issues and identify the benefits which oftentimes are poorly articulated by us because we accept them as dogma.
Secondly, I would propose that documentation standards as has been mentioned a couple of times today and yesterday be established for paper and electronic records. Minimal documentation requirements by specialty or demand and by location of care if put into place would be a strong motivating force for both improvement of the paper-based documentation process and support of that process by electronic means.
DNM requirements from HCFA despite their pain and anguish which they caused were a great boon for electronic medical records vendors because we implemented of course a system which in an expert system manner advised the user as to what code level he or she sufficed with their level of documentation.
I know as a provider I never wanted to remember all those rules. I couldn't possibly remember them, and I ignored every advisory that came across my desk from my practice environment with respect to how to use them in my own clinical practice. Yet, with the system I could be compliant and perhaps even avoid jail time or penalty.
Thirdly, I would like to propose that we work towards establishing minimal data sets. The essential terminology standards for clinical or -- these would be essential terminology standards for clinical findings pertinent to No. 2, and to clinical decision support and outcomes assessment. I think there are many grassroots efforts going on across the country now in the conquest of database efforts, NCQA, HEDIS, JCHO, OREX(?) and others too numerous to mention that are very poorly coordinated, yet will all point towards similar data elements and overlapping data elements very, very frequently.
No. 4, interoperability standards for EMRs I think is the largest technical hurdle we should face and try to overcome in the short term, that is to provide identifiers for person, provider, place, place, plan and employer. Enterprises are already implementing so-called "master patient indices" of course because they face this problem in a local universe if you will.
Within my old academic environment we had anywhere from 30 to 60 systems that had to be interfaced one to another immediately and that was in a relatively small environment. In some settings there are 150 or more systems which have to be interfaced one to the other in a local universe called an academic medical center. This problem is then an order of magnitude more difficult when you think about enterprises trying potentially to share data one across or one to the other.
Secondly, reference terminology and information model has been alluded to. We need something which would support consistency of usage of terminology, semantics and syntax, support the interoperability of clinical information exchange and machine interpretation of these data for clinical decision support and outcomes analysis.
I think we have made extraordinary progress in messaging already with HL7, the X12 and NCPDP pharmacy standards and the like. Nevertheless, as all of you who know these in detail recognize these are only messaging standards and still are often used in an extremely idiosyncratic fashion in interfaces between comparable systems in different settings. I will explain what that means later if anyone is really interested.
I think, also, we do need to work towards a national health care information infrastructure specifically a secure national network for communication of electronic information pertaining to health, well-being, reimbursement analysis and the like all the while protecting the security and confidentiality of individual patient identifiable information.
I would like to propose two new ideas though. I thought they were going to be new, but Paul mentioned one of them already, that is the one about mandating quality standards. I think if the government said that it is a requirement for an institution, a provider organization or a group of providers to report quality standards not only in the quasi-voluntary fashion to JCHO and NCQA and the like but as a requirement for practice with more teeth behind it this would immediately create a great demand for these kinds of systems.
The one I think that has not been mentioned is that we may wish to mandate in some fashion that drug interaction assessments be done at the point of care. This would move knowledge bases from the pharmacist and downstream assessment of drug interaction upstream to the point of care, and we have data showing that this is very beneficial to the provider, the customer, the provider- patient relationship and smooth delivery of at least prescriptions dramatically.
Question 4B, which standards related to patient medical record information and electronic transmission would be most important to the business goals of my organization? Documentation requirements were a great fear, that is they pertained to the pocketbook of the provider, that is the money issues, though documentation requirements would support demand for products which I sell.
Interoperability standards address the implementation requirements we have as a single component in the overarching definition of computer-based patient record. We are one system in many in a health care environment which has to communicate one to the other. So, interoperability standards would help us do more implementation. Outcomes reporting or report cards, I think this addresses the fear quotient or element, that is as providers are becoming obliged to report upon their own practice outcomes and satisfaction metrics and the like this addresses the fear issue and could help again increase demand, and finally, and most importantly I would like to say that standards with respect to clinical decision support terminology, representation and the representation of knowledge could help us address quality issues.
What is the business case for more rapid standards development and implementation? As John Madison alluded to yesterday, I think it depends upon your perspective. Are you a provider or are you an at-risk provider organization; are you a health plan; are you an employer or are you a consumer? I won't go through this in much detail now except to point out that the business case is clear for those organizations who are at risk for their health care.
When a provider organization is at risk for formulary, referral or cost and services of hospitalization or even cost and services of ambulatory care, they all of a sudden want to know very quickly and in great detail what is the nature of the services they are delivering, how much, how often and a clinical level of detail, too, that goes beyond the claims database level of detail.
The case study alludes to how these systems can address cost issues immediately. I think we need further pilot studies and projects to address longer term quality effects of these kinds of systems on the health and well- being of the nation at large.
In closing let me just borrow from William Farr via Chris Shute because this is one of his favorite quotes. I cannot help but say it here. William Farr said in 1848, "Nomenclature is of as much importance in this inquiry as weights and measures in the physical sciences and should be settled without delay." Simon Cohn, Chris Shute, Jim Campbell et al in the recent Jamie article said, "Failure to achieve common goals and standards will preclude efficient comparison of practice patterns and outcomes, and we will therefore have great difficulty identifying best practices and sharing among institutions or nations nor would we be able to consistently implement identified best practices across institutions as health are deliverers begin to develop precision support resources."
I think the point I draw from that quote is as follows: We need to protect and maintain the art of medicine and allow the physician to focus on health care delivery and his interaction with the patient. We need to continue to develop the science of medicine, and these systems are integral to the furthering of science of medicine exploration.
Question 5, the four focus areas, I agree with the focus areas with the additional premise or request that additional focus on public/private mechanisms for the production, support and maintenance of informatic standards be created. I think as Jesse alluded to the notion of a terminology service bureau maintained by a trusted agency is an interesting idea.
Thank you.
DR. COHN: Thank you very much. I, also, always appreciate being quoted.
Questions from the panel?
Jeff?
MR. BLAIR: Dr. Middleton wound up giving us a generalized business case for nationwide benefits of having an information infrastructure that included electronic health records, computer based patient records and in it he pointed out that while the administrative savings is very nice, that that infrastructure would give us the information that would facilitate continuous quality improvements in clinical processes which would yield dramatic savings, if I remember correctly, approximately $250 billion a year out of $1 trillion that we spend on health care, and I would like to hear from the other witnesses if they feel as if these estimates are generally correct or not.
MR. TONKS: I am certainly not an expert in regard to those statistics, but I can quote an expert, and that is Dr. Brent James, and I believe that the numbers he uses are between 20 and 25 percent could probably be wrung out if we were more careful in terms of the highest quality protocols that were implemented and followed, and I suspect that those numbers would correlate to the ones that Dr. Middleton mentioned.
DR. COHN: Dr. Russler?
DR. RUSSLER: We have some evidence. I am going to cite a study done in Portland, Oregon regarding a hospital where paper pathways were implemented in one part of the hospital for the same condition that automated pathways were implemented at another part of the hospital. The study showed that there was an 8 percent reduction in variable costs in the automated pathway area.
So, this is the best data that we have that there will be significant improvement with automated systems and of course, standards are required in order to extend those across multiple hospitals.
DR. COHN: Dr. Tang?
DR. TANG: I think the range is about right. I have seen figures as conservative as 15 percent and maybe as broad as 25 percent.
DR. COHN: I was going to ask a question here. I think all of you have commented one way or another about this issue of alignment of incentives, and one cannot help but observe that there are sort of massively unaligned incentives right now in this whole area, and I tend to think of them starting probably at the terminology level and then at the standards area and then at the area of users and beyond.
Perhaps you could all comment a little bit about where you see the big issue is and maybe what the government could do to help encourage a better alignment of incentives in these areas?
DR. TANG: I go back to the quality reporting standards. I mean just like Bob Mayes was talking yesterday about DRGs, with the stroke of a pen HCFA basically defined and introduced prospective payment into the US health care system. Perhaps at the stroke of a pen we could have QRGs, you know, quality-related groups that would essentially define the need for data and that would in turn define the need for data standards and that would align all of our decision makers that we respond to, for example.
In contrast let us say ENMs certainly aligned a lot of incentives, more disincentives but my fear there is it is almost like the old HIS meaning hospital information systems where what we do is go capture charges, and we would be, you know, hospital systems turned into a charge capture system. I would hate to see the broader continuum of care information systems go back into the billing world.
So far out of the hospital has really been focused on clinical care and quality. I would love to have it motivated even more specifically and deliberately by a QRG and get away from going towards backwards filling and compliance and those kinds of things.
DR. MIDDLETON: I would like to add to Paul's comments I think perhaps that task is beyond the scope of the NCVHS Committee, and I would be reticent to ask for you to solve the restructuring problems of the US health care delivery systems and reimbursement mechanisms in this country given the country's recent experience in this area.
However, I think some things are happening in the marketplace nevertheless that perhaps bode well for us. First, I think the trend from fee for service to managed care is continuing. We see some marketplaces which appear to be in a post managed care marketplace, one that is where consumerism may be raising its head for personal choice and accountability.
I think as the consumer enters the equation in a more empowered way looking at physician report cards, looking at hospital report cards, looking at satisfaction indices and metrics on line, for example, as can be done in some states now that the consumer may become the great motivating force that has been missing from this equation, that is if the consumer expects the physician to use a clinical information system to manage his or her health care or further if the consumer even presents to the physician with his record in his pocket on a disk or device or maintained somehow in cyberspace, then the tables will be turned.
It will no longer be an option or a luxury for an institution or provider to use these kinds of tools. It will be a requirement for practice.
DR. RUSSLER: I would like to bring up the example of the CPT coding, and you are right. I mean you can either require certain standards for reimbursement or you can require certain standards to stay alive in practice independent of reimbursement. I mean those are your two ways to do it.
I consider the CPT service level standards created in the CPT documentation as quality documentation, quality standards. Now, they may not relate to outcome quality standards, but they were certainly documentation quality standards.
We noticed, and we developed like MedicaLogic a medical office application that with structured data was able to use the algorithms to go straight from the documentation to a service level setting. So, that allowed organizations, potential customers who hire large numbers of physicians to be pretty sure that the documentation that was being reported out by the physicians actually was not going to get the managers in trouble, you know with fraud.
As soon as the government pulled back on the implementation or those standards much of the market pressure to buy systems went away. So, we went ahead, made the investment, you know, of literally millions and millions of dollars to develop this solution to help people provide high quality documentation and meet the standards of the government and when the rug was pulled out from us that investment has gone out the window for now. So, that is a major concern, and so that is a disincentive, I mean that is an example of a disincentive for us to do standards-based development until the practice is actually out there.
DR. COHN: Jesse?
MR. TONKS: I think that the shift from looking at just episodes independent of each other moved towards the overall view of the delivery of patient care and the managed care as Dr. Stead pointed out yesterday I think those guys, those DRGs instead of being encountered focused across the delivery of care for our patients for a disease or longer term as an episodic kind of evaluation for quality and delivery of care is where we need to move to as opposed to the independent individuals and encounters with the hospital or clinic.
DR. COHN: Yes, Bob?
MR. MAYES: I would like to make sort of a response but also ask a couple of questions. One, I think you will be happy to know that HCFA actually is very seriously moving into the area of quality care information. Obviously with out long-term care minimum data set requirements, OASIS, we will be publishing our Medicare HEDIS results on the Web.
We are, also, going to be focusing over the next several years on six major clinical areas in requiring all of our peer review organizations to be doing quality improvement projects in those.
One of the things that we have been moving towards within HCFA particularly on the quality information side, and I would like your sort of opinion as to whether this would be a useful type of incentivizing activity is to begin to shift from a paper-based specification for our requirements to an electronic specification, an object type of specification if you will where we pre-interpret what we want, wrap some APIs around it and make it available freely to the industry.
So, I would like just to hear if you think that would be useful. One other question that I wanted to ask, there has been over the last couple of days I think three broad areas of standards spoken about. The first you have obviously addressed the need for and that is in the area of semantic management of data with terminologies and the models and such. I think the second area in terms of architecture and some of the technical infrastructure you have, also, spoken to.
A third area though that came up yesterday and I would like to find out whether you have experienced problems with is in the areas of the social political issues, primarily the legal issues. Certainly privacy is one, but something else that was brought up a couple of times was at least a perception that differences at state level in terms of legal requirements adds a non-value added burden to your efforts to market across states.
None of you actually brought that up but I was just curious. As vendors do you see the difference between state laws, for instance, as having any real impact on your abilities?
DR. TANG: I am sure that is a rhetorical question because of course vendors are very affected by the state laws and so, anything from privacy through data standards should definitely be at the federal level.
As far as your proposal for electronic submission of quality-related data, I think that is an outstanding idea. I would seize the moment and be very specific and precise about your definitions of those data elements and the methods to gather them because that is where you open the biggest hole for interpretation and by being precise about the methods and the measurements you start setting a standard that you can compare across groups, plans and states and then that then trickles down and starts forcing people to establish the data standards which is what you want to motivate.
DR. MIDDLETON: I would second Paul's comments. I think what I heard was the notion of a quality reporting plug in if you will. The software architectures now being created of course are becoming increasingly amenable to the notion of plug ins using modern technology that will do a whole host of things. I believe, in fact that a third market will quickly emerge. Perhaps the Federal Government will be the first major player i this market where knowledge constructs, objects and tools will be developed to be used upon the platforms of EMRs or CPRs all vendors on the table here included and that is very exciting because I think it was mentioned already that the software vendors don't see themselves typically, at least we don't as being knowledge authors. We want to give tools to offer knowledge or use knowledge. We don't want to be the author.
So, with respect to a quality reporting plug in, please move ahead quickly because it would be another compelling event if you will for the electronic maintenance, storage and management of those information relevant to your report.
With respect to the state regulations and issues I don't pretend to have expertise in this area because it is too complicated but it does impact upon the business and we do pay attention to the state-by-state requirements.
I think medicine philosophically I would like to ask the question if medicine will be practiced forever on a statewide or state-regulated basis or if we need to think about federal models for the regulation and management of health care delivery systems like the VA has used for years.
Medicine of course has been pressing this issue for several years in its practice and implementation across the country, and internationally, and I don't think it has been completely adequately addressed.
If the Federal Government could go above the state regulations and provide perhaps some superseding legislation that establishes these minimal requirements that we could all abide by, that would be helpful.
DR. RUSSLER: HBOC is an international vendor. So, not only do we have problems with states, we have problems with countries. The hope would be from HBOC that whatever standards organizations were chosen for various parts of the standards that they be international, that they have international connections so that there are communication lines that are created between the US naturally, the states and even getting down into communities with individual laws. It has giant impact on vendors like us.
So, the more uniformity, the lower cost for us to deliver products to the customer because if we have to customize things by state or by community or even by country those costs have to be passed along to our customers and will raise the total cost of health care.
No. 2 on the issue of objects what you described is almost the exact same description as what is called the template proposal in HL7, and so we certainly would like to encourage that proposals in those sorts of templates be coordinated through the HL7 so that using the HL7 and X12 partnership where X12 wraps HL7, and HL7 essentially wraps those objects as templates so that we can have communication between payors and providers and providers and regulators, etc. That is our recommendation.
MR. TONKS: I think the biggest restriction we have encountered is the encouragement of community information systems where all of the participants in a delivery system are not owned by a single organization as to who owns the data, what data can legitimately be shared and local laws and politics have prevented those from being the blossom to the extent that they certainly should be if there were better coordination and more leveling and definition of ownership of data that can be shared.
DR. COHN: Let us finish up that. Are any of these follow-ups on this one? No? I just wanted to say that this is of particular concern to the work group, and if we could ask you to have your lawyers or whoever it is in your organizations that are really knowledgeable about the state issues or the international issues to send us some additional information from your viewpoint, that would be very helpful.
Follow-up questions?
Okay, we are done with that one.
MS. FYFFE: I have three questions for everyone on the panel. Let me go through each one. They could be short answers. I don't know. Given that the vendors, a couple of you have a presence in many, many health organizations in the country, what I get worried about is the average B flat community hospital that has 250 or 300 beds. There are lots of them out there. What would it take to incentivize those hospitals to make the necessary expenditures or have the appropriate amount of interest in wanting to pursue electronic medical records?
What I envision, and this is oversimplying this is that the administrator of the hospital doesn't have as his five top objectives spending any more money on computer systems, and if they hear that the docs want this, they go, you know, another toy or something like that. So, what do you think it would take not just for the big university teaching hospitals who might be very well enlightened about all of this but the average hospital in this country which really makes up the majority of the beds in this country to buy into this, and I would like to hear from everybody if you have an opinion. If you don't have an opinion, you don't have to give one.
Thank you.
DR. TANG: Kathleen, you are absolutely correct, and I sure that everybody is going to say that that is one of the big challenges in the industry is that the decision makers, usually the administrators have not bought into the fact that this is on their top five list. After the Y2K issue especially it is going to be everybody else's turn.
The short answer is really to make this a business imperative, and the way you do that is through the funding mechanism. A condition on payment is really the demonstration of let us say these quality measures. So, the way that HCFA is going is one of those things that can make it a business imperative. Another one that was a little stuttering was the ENM made it all of a sudden that this was imperative for a while, and I am sure it will come back, but I would much rather have quality drive the business imperative than the coding.
DR. MIDDLETON: I guess we will continue to move down the table for you. I think the incentive that comes to my mind first is to change the date to 2002. If we could magically get beyond Y2K that would be one of the most powerful things, but that won't happen. So, I think as to the reporting requirements, as a business imperative we will do it. If you are required to report and data need to be aggregated from disparate sources in electronic form, and it may impact the provider to provide some clinical data from clinical encounters and the like that would do it.
Secondly, I think though an important thing to keep in mind is that across this country every medical marketplace is in a different stage of evolution. Those marketplaces which are already into pre-managed care, managed care, post-managed care environments, they don't see this as an if. For them it is a matter of when. Those sites to understand how the clinical data melded with the financial data help them manage their business processes don't need further business imperatives. They are already facing them and many of them are at risk for not surviving because they cannot manage their clinical risk or the financial risk, and they are already using systems, and I think the demand is clear in those environments.
In fact, we cannot keep up with it. That is it.
DR. RUSSLER: As one of those hospital vendors your comments were an interesting perspective. Let me tell you our experience. We have our hospital electronic medical record system called Care Manager. It is now contracted for in well over 200 hospitals and implemented in just under 200. That is where nurses carry computers to the bedside or have computers at the bedside to do all their charting, all the OT, PT, respiratory therapy, do all their charting. That is a pretty significant penetration, you know, and I think belies the interest that you mentioned. I think hospital administration is very, very interested in these systems, and they are a top priority.
Now, failing hospitals do not pick up these systems, you know, and so you cannot look at the lower 50 percent of the hospitals and use your judgment. What you have to look at are the hospitals that are successful and what they are doing, and so it is not academics. These are community hospitals, not academic hospitals.
Home care, we have close to 15 to 20 thousand nurses carrying laptops into homes, and these systems, we can hardly install them fast enough. So, I don't think that is the issue.
Now, the issue is that none of these hospitals can communicate with each other because they are all setting up these systems without standards.
MR. TONKS: I think as vendors we are somewhat remiss in pointing out the return on investment for that kind of a hospital and that kind of investment that certainly the longitudinal lifetime when the family practitioner sends a patient to a specialist and the specialist doesn't have to redo all of the same tests and see the details again, there are significant savings that can be pointed out, and we don't always have a good bottom line for the financial managers that have to sign off on those kinds of contracts to appreciate what can be done.
DR. LUMPKIN: I am intrigued by your comments and perhaps if you could provide to the staff of the Committee some of the installations, I would like to go out and see some of those myself. So, if you have some in Illinois, and I get around the states. So, if you don't have them in the Chicago area where I live, then I would like to see them from any of the vendors here. I think it may be helpful for us to actually go out and look at those and maybe talk to some of the hospital administrators to get a feel on a one- to-one basis. I don't think we need to do it as a Committee, but I would like to get some sites so I can do that.
MR. TONKS: We would be happy to help, I am sure, all of us in any way.
DR. COHN: Mike, you are next.
DR. FITZMAURICE: Yes, the Agency for Health Care Policy and Research is virtually an agency for health care quality research. We undertake research into quality measurements and methods for quality improvement, and so we are concerned about where the data are going to come from and how costly it is to get that data. So, you can see our interest in long-term and data standards.
I wonder if you are familiar with the Health Informatics Standards Board. I know it was mentioned, the need for coordination among standard developing organizations, vendors and users, mentioned both by Jesse Tonks and by Blackford Middleton.
Since 1991, with the help of Milt Greberman, Clem McDonald, myself HISB was formed under the auspices of ANSI, and it calls together the standards developing organizations. It has been lauded for what it has achieved, but there is still a lot that it hasn't achieved. It is a voluntary organization. It is under ANSI and so they choose to come to the table or not. They have always chosen to come to the table.
Vendors have, also, come to the table and other users. HBOC has been a stalwart member; 3M has participated strongly as well. Is there something you see that can be done voluntarily with standards development or do you think that something additional needs to be done?
MR. BLAIR: When you introduced it you referred to it as the Clinical Information Standards Board. Did you mean Clinical Information Standards or --
DR. FITZMAURICE: No, I meant Health Care Informatics Standards Board. I beg your pardon. It is the ANSI Health Care Informatics Standards Board. So, let me address the question to Jesse and Blackford. Do you see more that needs to be done, and does it need to be more than voluntary?
MR. TONKS: I think there is more to be done, and it needs to be more than voluntary. There are certainly misconceptions among the organizations, rumors that they are competing with each other where they really aren't. There are certainly rumors that they are trying to outdo each other, and I don't think that is the case. So, I think there needs to be some coordinated public relations in one event, but I believe it should be required that they attend and that there be facilitated meetings that would discuss what domain each -- what is it best qualified for and what is their core competency, those areas, and what they should do to eliminate duplication and to ensure that they are working together. If there are components as there certainly are that need to be defined like PIDS(?) and COAS(?) and so on, those same components should be done by MS HUG(?) FORACTIVE(?) X as an example, that those should be done by one organization, but there should be standards defined in such a way that they can be implemented, and that is where the private enterprise industry comes in. The free marketplace should be able to implement it in any standard whether it is the CORBRA(?) INACTIVE(?) X or DECOM(?) whatever it might be. That should be left to the implementation, but the standards should be consistent across all those organizations.
DR. MIDDLETON: I am aware of the HISBE(?) efforts and have been tracking them. My organization has been tracking them from afar. I think that the issues that we see are first of all that as a newcomer in the field, if you will a 15-year-old newcomer now growing to a midsize company the overhead and costs involved in participating in all of the standards-setting efforts which pertain to the implementation of our software is extraordinarily burdensome.
I mean it is a voluntary effort as you described, but when you translate that into the real costs for a vendor to send an engineer or somebody to participate in and fiddle with these kinds of things is going to cost hundreds of thousands or millions of dollars a year literally to all the different things. So, I think the cost in voluntary organizations for small to midsize vendors perhaps even for the larger vendors is non-trivial and needs to be considered by those entities, and secondarily, I think the rate of progress in those bodies while laudable I think as you say in specific areas is still disappointing in a broad-based sense.
We cannot afford to wait for the standards setting development process. Given our business model, our financial requirements and desire to continue to provide services to a broad-based sector of the health care market and what not we cannot wait.
So, we are proceeding with to the best of our ability adopting the standards which are robust enough now to adopt and pushing in the standards development efforts that have fairly immediate return to the software development implementation process.
Things which we have been involved in include the Andover Working Group sponsored by HP initially, and that had immediate appeal because it translated standards into technology or even software components that could be used very readily for supporting interoperability between applications and just integration types of ideas which had immediate utility.
Some of the longer range issues, for example, in information modeling, etc., are first of all beyond our reach in terms of our own internal development; so, we do what we can to survive today and meet the needs of the customer today, and we participate in those organizations from afar sometimes.
I would encourage perhaps some distinction to be drawn between those efforts which are, again, the two distinctions between the soft standards development efforts perhaps and the hard standards development efforts. The hard ones are more tractable and perhaps more amenable to the voluntary process because they are more tractable. I think the soft standards development processes though are intractable because of their very nature, and perhaps that is the area where we need some government guidance or stimulus.
I wouldn't disrupt the soft standards development process, but I might infuse it with a sense of urgency from government requirements, perhaps or even infuse it with some support to translate some of the good work being done on a voluntary basis into usable technology or usable standards.
DR. RUSSLER: I wanted to comment because I made some implications in my testimony that I think directly relate, and I want to point that out. One is the analogy I made to the high definition television standards. When it comes down to it somebody had to make a decision.
I mean there is a requirement. There are not two standards or three standards out there today. Parts of it were voluntary, but the end of the process was FCC said,
"This is the standard."
So, I think health care flounders because we like to be nice to everybody, but the issue is this is a communication standard. Why it is not under FCC is a good question because it is a communication standard, and I think you guys need to be the same way as the FCC is. You know, you use a carrot and a stick, and the FCC created carrots and sticks for the broadcast industry around HDTV, and we will all be enjoying the results of that in a few years.
Second is the investment issue and the volunteer nature. Not only is it hard on small vendors but it is hard on large vendors where we are competing with vendors who don't volunteer because all of our investments, our volunteer time has to be translated into costs of our products, and when somebody, another parasite leeches off our resource investment and delivers product at a lower price that is a problem.
So, the issue of voluntary versus required if it were required it would level the playing field.
DR. FITZMAURICE: So, what you are kind of describing is either all vendors should be taxed to contribute to standards development or it is broader than that and it is a public good and the funds ought to come from somewhere from every consumer's pocket.
DR. MIDDLETON: We would support any proposal for funding that is equal.
DR. FITZMAURICE: In other words the private market is not able to do this in the area of standards development?
DR. MIDDLETON: That is right.
DR. FITZMAURICE: As fast as you would like it?
DR. MIDDLETON: In the voluntary area.
DR. COHN: Paul?
DR. TANG: You asked what could we do to accelerate this process. I think what I would do is add a deadline, and government could add that deadline. I am not as afraid of the volunteer process because I think I have been part of volunteer organizations where we have gotten something done like the summit that we organized, as long as you put a deadline to it, and I think that is the role of government.
DR. FITZMAURICE: I think I did forget to mention that since 1991 AHCPR has funded the meetings of HISB and partners with the other standards developing organizations in funding meetings in different parts of the country. The next one is going to be here in Washington at the beginning of March or the middle of March, excuse me, right after St. Patrick's Day.
DR. COHN: Thank you.
Any follow ons regarding this one specifically before we move on to Richard?
Mel?
DR. GREBERMAN: I have a follow on and a question for Dr. Russler based on the comments you just made about the FCC and the HVTV standards. The related issue is the allocation of bandwidth for HDTV and the concerns and the issues that were raised in the health care community as a result of the early decisions by FCC in terms of specifically the interference with medical telemetry systems and number of markets.
The FDA became very much involved with this issue, and the bottom line is we ended up forming a collaborative relationship, working relationship with the American Hospital Association, AHA, others. We recently discussed this at the AMA meeting. How do we avoid some of those early on, those kinds of analogous problems by not dealing with related industries? We talked about yesterday the issue. We are different. We are not learning from each other. How do we avoid some of those problems in a constructive manner?
DR. RUSSLER: I think one is first of all looking at successful standard setting programs, and obviously having delivered the HDTV standard so that it is not real means that whatever process they used it was successful.
So, I would study that process and use that process to justify actions that this group takes because it is a best practice. It might not be optimum, but it works, and it hasn't worked yet in health care.
DR. COHN: Okay, Dan, thank you.
Richard?
DR. FERRANS: I wanted to comment on a couple of things. First of all I found it sort of extraordinary that here we have representatives of four large companies involved in this, and people are generally favoring regulation by the government, and I hear a consistent message that industry alone cannot do it itself.
I did want to ask a couple of questions. First of all, I think it was Dr. Middleton who commented that the utility of this would go up by a square to the number of users. One of the issues is the cost of CPRs. I think that certainly a business case for return on investment has been made here. Perhaps it is not widely known. I am curious as to all of the panelists' opinions as to will the development of standards, if that really facilitates software development in the sense that now you don't have to be writing all these custom interfaces, dealing with lexical issues, if that is going to accelerate development of software, would that, also, have some downward price pressure on CPRs making them more affordable, therefore more widely available. As one alternative to sort of if, for example, HCFA mandates that everyone has to report this information, then that would essentially be less of an incentive and more of a mandate type situation, I am curious what you all feel about the effect of standards in that area, and the second sort of follow-on question is Jess and Dr. Gardner yesterday mentioned about some body that would basically administer some government or quasi-governmental body that would somehow administer vocabulary standards as to the practicality of that and just sort of general opinions on that.
DR. TANG: There is a huge activation energy to overcome in implementing these systems, and part of it is cost, but a big part is this changed management that both John Quinn and Bill Stead talked about yesterday.
I think one of the goals for us proposing, let us say these quality metrics is to turn that into a business imperative for the decision makers to decide to get into this very arduous and time consuming and costly process of implementing change and that change is to capture clinical data in the process of care.
So, I think we are not saying that industry cannot supply the products. We are saying that we need help in motivating the buyers to involve themselves in this process of change and getting over that activation energy is the hard part and then once we do disseminate this and get a critical mass, then we start getting back the $250 billion that Blackford talked about.
DR. MIDDLETON: First of all, I agree with Paul's comments with respect to implementation. I think the standards development process is relevant to implementation because implementation is dependent upon intraoperability. So, following that chain of logic I actually think the demand is there. We have not had problems with demand. We have had problems with overcoming this energy of activation as Paul described which is both financial although that is lessening because the business case is becoming more apparent and sociocultural, that is changing the way we do things in the practice of medicine and delivery of medicine.
I propose that the standard setting effort continue, perhaps with the stimuli we have talked about and recognize that the standards which we would choose to receive from a process are really analogous to the DNA, if you will of these systems.
I really don't want to have someone tell me how to code the system, what software language to use to code the system, how to implement the human-computer interface or what modality with which to interact with the system, etc.
I would like to have, however, standard tools with which to build the system as my customers have requested it from me. Thus, I think if you think of vocabulary as really the DNA of these systems, and I don't want to stretch the analogy too far, we will choose to be in control of its phenotypic expression and we can compete that way actually very successfully I am sure, but we don't want the requirements of the government to say, "Well, you know, the screen has to be pink. It has to have this button here, and you have to document it like this." That is not going to work.
In fact, what we do as a vendor is we control the DNA for our customers. Because there is not yet a standard information model or a standard reference terminology we have crafted one that will meet the needs of 90 percent of our users today.
So, when a user has a standards requirement for a new vocabulary we receive that request, look at the standards available to us, standard nomenclatures available to us, take the appropriate terminology, code it, implement it and send it back, and that works very well.
I would propose we extend the analogy to this terminology service bureau or something like that where that request might even be passed through the vendor from the customer to the terminology service bureau or even directly from the customer and then delivered to us in a public domain way for all to use as we express phenotypically the DNA in our products.
DR. RUSSLER: The first question, just a little bit of education, we actually have a business unit which is responsible for return on investment evaluation and one of the reasons we are able to sell successfully is that we are able to go out and demonstrate that the current costs of our products do give a return of investment of better than two.
Return on investment language just to review for the Committee, a return of investment below two which means $2 returned for $1 invested, below $2 means that the system is cost prohibitive.
So, any CFO is looking for an ROI of two. If you can show that you can sell the system. That is no problem. The competitive forces as more and more people deliver these systems may require that you may drop the price down, but the price cannot go lower than the cost of creation of the product. Without standards the cost of creation of the product is much higher.
So, we work in between the cost of the product and the ROI of two, and we can be more competitive if you can help us lower the cost of building the product, and that is the problem with our office medical record and Blackford's. When the standards went away all that investment you know, was lost. So, now what you have done is increase the cost of all of our products. That is a problem. So, health care suffers from that generally.
Does that answer your question?
DR. COHN: Jesse, comments?
MR. TONKS: I think it is important to understand that there is a clear line between what should be standard organizations or government standards defined and where the free marketplace needs to operate. The reference information model and the reference terminology model are meta data. There are reference models and reference libraries and that is where we need the help with the standards and government assistance in putting that together, the model.
The instantiation of that model, the implementation of that model should be up to the free marketplace to develop, and that is where we can add our value to those standards where we can solve the problems of variability between different sites and different customers and that should not be regulated. That should be left to the free marketplace.
So, there is a clear line between meta data and the instantiation of that meta data with real patient data, and that should be the line between the standards organizations and the help from the government organizations and where free enterprise should be free to respond to market demands and market needs.
DR. COHN: Thank you. We are already running over, and I apologize to everyone. We have three final questions. Kathleen is first and then Jeff and then John. I would ask everyone to be as brief as possible both with the questions and the answers.
DR. FYFFE: Several of you are physicians, practicing physicians, vendors and former medical school students, and I am wondering if there would be any value in studying the idea of possibly having required medical school curriculum dealing with standards for documentation of medical records and let us say fundamentals of electronic medical records 101 for the medical schools in the US. Give me your quick opinion on that because that will tell me how you really feel.
Thank you.
DR. TANG: That is a fabulous idea. In fact a lot of medical schools already have some introductory courses, but it is not pitched the way you described it which is what should be documentation for care's sake and how do you complete that within a CPR. That would be a fabulous idea.
DR. MIDDLETON: I will try to be as brief. You know, it is challenging. I would say, yes. Absolutely it is a necessary requirement. Many medical schools are pursuing these kinds of initiatives. In my own practice environment, it is a teaching environment, we have established an informatics curriculum. I believe that medical decision making as a skill and the science of medical decision making is one of the most important sciences that physicians have to learn, and it is the most perhaps least well addressed in the biomedical curriculum as an undergraduate. It pertains to the understanding of sensitivity, specificity, test performance characteristics, biases and errors in judgment which plague us all and will not only be solved by systems. They require attention to medical decision making as a science.
DR. RUSSLER: I really support medical informatics 101 for all medical students. However, medical students aren't going to be the decision makers and system purchasers for many years, and I would rather see a requirement in medical informatics 101 for chief medical officers of health care organizations.
DR. COHN: Jesse?
No comment? Okay.
Jeff?
MR. BLAIR: I think Kathleen had one more.
DR. FYFFE: No, I am done.
MR. BLAIR: Okay, help me clarify something because I heard from several of you today that from a terminology standpoint, and I think, Blackford, you will refer to it as soft standards, and I like that term, that a reference terminology does not exist. Yesterday we had a witness from the College of American Pathologists who was describing to us the SNOMED reference terminology. Of course, we are, also, aware that the National Library of Medicine has the unified medical language system which does mapping between code sets and I know several of you are active in HL7 and you know the work that has been done to try to integrate terminologies and code sets into new message format standards and the HL7 vocab sig(?). So, there are several different initiatives that coordinate in one form or another, but I want to get back to in particular your -- help me understand specifically the statements that several of you made that a reference terminology for vocabularies and code sets doesn't exist yet.
DR. MIDDLETON: I will start if you would like, Jeff. I think what Kent Spackman was alluding to in terms of the SNOMED RT is absolutely a step in the right direction, and we will be, in fact, a partner with him in evaluating that technology, given our proximity to his location in Portland.
By Kent's own description, however, the reference terminology is really not an interface terminology. I believe that this can get very complicated very quickly, but the reference terminology does not necessarily imply an interface terminology that a user may use or customize to his own purposes for interfacing to the reference terminology or the canonical terms as you like. The canonical term set needs to have even above it then the reference information model of HL7, but unfortunately the HL7 RAM effort and the SNOMED RT effort, I am not familiar with how well they are dovetailing one to the other, but the notion of dovetailing the information model and its meta data to the SNOMED RT has not been accomplished in my humble opinion. I may not be aware of all the details behind the scenes, but I see that as being a gap.
MR. BLAIR: Thank you for that clarification.
Any others?
DR. RUSSLER: I will second a little different explanation of that. I had my conversations with Kent as well and encouraged SNOMED to go beyond the capturing of large amounts of expressiveness, and translating because what we really need is what is similar to the microglossary for DICOM(?) that is included in SNOMED.
We need attribute or value sets for specific attributes and for us to take the 140 SNOMED terms and say, "This is the subset of SNOMED that fits in this attribute," is an incredible amount of work, and so, in some ways I tell Kent, "You have done the easy work. Now, who is going to do the hard work for reference terminology?" and SNOMED hasn't taken that on themselves yet.
DR. MIDDLETON: I would like to just add to Dan's comments because I think they are so useful, in fact, although we have the ideal of a reference terminology fully specify what objects, attributes and value lists, in fact, what the marketplace will bear today is hardly anything; it is nowhere near that. In fact, the user will barely tolerate touching the computer. Of course, as a clinician or a physician to write a note, I am more accustomed to either scrawling three or four words on a piece of paper or maybe 20 if it is a good day or dictating, and while we may work towards dictating to structured fields and content we are not there really yet.
The notion of structured data entry is perhaps most germane to this Committee's efforts, that is I would hope that the Committee doesn't overspecify a solution with respect to documentation requirements, reference terminology or model because the marketplace won't bear that and in fact that would be counterproductive. I would argue that we incrementally approach that end by first approaching the development of the reference information model. That is a context in which we can develop from the reference information model specific reference terminologies and interface terminologies and then user paradigms for structured data entry.
I am afraid it is somewhat top down, but on the other hand, I think that is going to be functional in the long run. The challenge is how much can we do from the bottom up to make these systems work today and interoperate versus coming top down in a modeling sense to achieve our long-range objective.
DR. COHN: I don't know if it is my ability to recognize voices, but, Jesse I don't think I have heard your voice.
MR. TONKS: I certainly applaud the SNOMED RT project and effort, and we, also, will be beta testing some of the progress in that area. The SNOMED even by their own definition is not going to be an all encompassing terminology system. There are another 150,000 concepts for just medications that aren't there that need to be there at a detailed level to do order entry or alerting. There are another 45,000 in LOINK that aren't there although they are working together to make sure that they operate together.
There are all the REED(?) codes that are certainly involved with international activities. So, SNOMED isn't -- there is no single source. SNOMED is a great step in the right direction, the RT is and will be a great improvement over the SNOMED international, but it is still not the total picture that we need to have, and it needs some organization to pull that together, I believe.
DR. COHN: Jesse, thank you.
Paul?
DR. TANG: I don't have a whole lot more to add. I mean we are in support of the SNOMED RT effort as well and acting as beta partners as well, and the point that Blackford made and was made yesterday is anything that starts impacting on the work flow certainly has a big impact on the acceptability and the use of these things, and that is the bottom line.
DR. COHN: John?
DR. LUMPKIN: I think I have an important question to ask, but I am not sure if we have time for the answer. So, let me perhaps frame it to you and ask for your input in written form.
We heard a lot yesterday. We heard from a couple of your customers or potential customers, the hospital association, the AMA who are very leery of imposed standards, imposed regulations. We are hearing from you that there should be imposed regulations, and I don't know if the dichotomy is because of what you are talking about and what they are talking about are two different things; they don't want to be told what system to buy or which they perceive a being standards.
So, my question is, and again, I am not sure this is the right format and I would invite you to maybe spend a little time and think about it. We have to write a letter as a Committee to Congress telling them what should be done about this issue.
If you were writing the letter, what letter would you send?
DR. TANG: I think we are talking about two different levels. You know in the HP way, the management by objectives, and I think we are asking to set standard objectives for us all but not tell us either what systems or as Blackford said what should appear on the screen, and the standards by which health care is measured are basically quality and cost, and I am trying to push the quality side as if we could have standardized measures for quality that would motivate and mobilize both the buyers of these systems and the users of these systems and the people who write standards to get them to work, and that is the deadline we need in order to get this voluntary effort further along.
DR. LUMPKIN: The point is that HIPAA does not give this Committee nor most importantly HHS authority to institute standards for computerized patient records or medical records on this issue. They only direct the Committee to tell Congress what authority should be given.
DR. TANG: Right. So, you are supposed to make recommendations on standards regarding computer-based patient record systems.
DR. LUMPKIN: No, actually we are to make recommendations on what directions and what authority Congress should give the Department in order to do that. So, the Department right now couldn't write a recommendation, couldn't issue an NPRM as we have with the administrative standards.
DR. COHN: John, I actually want to clarify that for you because that is not quite right.
DR. LUMPKIN: Okay, but the point is that we need to tell Congress something, and if we just spend a little time, and I am not asking because this is a little bit of a complex question, and it is right before a break, but if you could send us either by e-mail or letter, if you were in our shoes what would you tell Congress.
DR. COHN: I guess I would amend that a little bit. I think the question is a very good one. We would ask for written comments from you on that though be aware, also, that we are sending this to the Secretary of HHS, and some of this may be telling Congress, some of it may be recommending to the Secretary things that the Secretary can do for herself.
DR. LUMPKIN: Could I just clarify then that the point to be addressed is the juxtaposition of our desire perhaps as developers for standards to enable technology development and its implementation, distribution and maintenance and support versus the occasionally visceral objection to the use of this technology in clinical practice by some practitioners or help me if --
DR. COHN: No, I don't think the issue is mandating, I don't think anyone perceives the issue as mandating how people keep records which I think gets to the visceral issue. It gets to how people who use or will potentially use the system perceive -- you perceive it one way. I think they must perceive it in another because they have come to a different conclusion what these standards would do and what these standards would be that would be adopted.
DR. LUMPKIN: If I were making the recommendation I would say that what the government should do is put out a request for a proposal to ANSI organizations regarding a standards and standards process, that what I expect would happen in that situation is each ANSI organization or collaborations between ANSI organizations would generate a proposal and then the government would be in a position of selecting the nicest proposal. You know, I think we can guarantee there are probably not going to be more than three major proposals that come out of that kind of RFP, major. I am talking about from ANSI organizations because there are not that many ANSI organizations, and then we would have some justification for selecting one of those, blessing it and then funding that with government assistance so that they can do a really good job of consensus work.
DR. COHN: A parting shot, and then we must get ready to break here.
DR. MIDDLETON: I appreciate the privilege. You know, one of the things John that, or James, sorry, the notion comes to mind as to what the intent of the Committee's effort is. I mean the analogy might be drawn historically between what happened with the Flexner report and what this Committee is all about. I mean the Flexner report was not received well by as many as one-third or one- half the hospitals in the country, and half of them thought it was bad, and of course, half of them went out of business.
So, I believe that there is going to be pain with these recommendations, but if you think about the public good and utility overarching, then that is the goal. That is the focus and in that regard the Committee can have very powerful recommendations.
DR. COHN: Okay, I am going to suggest that we at this point break for 15 minutes.
Thank you very much. It has been a wonderful panel, as you can tell. We have had a hard time breaking. We will come back in 15 minutes.
(Brief recess.)
MR. BLAIR: Are you seated and ready to introduce yourselves? Okay, then you start from left to right, please?
DR. MC NAMARA: I am Tim McNamara from Multima(?) Information Services which is a wholly owned subsidiary of Cerner Corporation. I am representing both of those organizations today.
MR. RADTKE: I am Gary Radtke. I am with Ford Motor Company in their health care management area.
DR. PETERS: I am Rick Peters, President and CEO of iTRUST. We are an Internet start-up out of California.
DR. MORRIS: I am Jon Morris. I am Vice President and General Manager of Industry Technology Products at Oceania, a clinical information system company based in California.
MR. BLAIR: Okay, Jon, Morris, would you start for us and give us your testimony?
DR. MORRIS: I would be happy to. It is a pleasure to be here and to see some familiar faces. I have a handout prepared, Jeff. I will walk you through some of the pieces that are in that, and there are other pieces that we will be, I am sure touching on during the discussion and the question period.
I think if you look at the congressional instructions what is not clear or what is clear is the desire to adopt as uniform data standards for patient medical record information. However, the types of data, the parts of the patient medical record information that need to be automated are not as clear, and one of the things I think you have probably heard over the last day and one-half, and you will continue to hear over the next hour and one-half is the blurring in the boundaries between administrative data code sets, reference terminology, meta data information models and all the good things that make the informatics world run.
The code sets that are currently being proposed for the administrative and looking at what makes sense for clinical data, I think there is plenty of opportunity for discussion there. Looking at the utility of ICD-9, CPT4, the HICPICS(?) the dental terminology and NDC codes, I am sure Tim McNamara will dive into NDC a fair amount over the course of his discussions.
I want to give you a framework, and actually there is a picture on the second page of the handout, and I will describe it as well, but there is this concept that the information axioms that pervade businesses and this concept of source of data and there is a capture of data, and then there is a series of data processing activities, processing, storing, creating output and then output going to the use of that data and if you look at what has been done in many other industries I think things that we are sort of dancing around here it is to talk about how efficiency is gained if you move the capture out to the source, and then if you actually move the output to the user, the end user of that data.
So, a lot of what many clinical information system companies are in the business of trying to do is to get clinicians at the point of care involved with the capture of data and then essentially have a series of processes of storing, output and then pushing the use of that data back to other clinicians at the point of care.
That model, if you think about, this revolution has hit a number of other industries. If you look at telecommunications and the fact that you don't need to use operators to make phone calls anymore; you do the source capture directly at the time you touch your phone; if you look at banking and using ATMs and tellers or you look at going to the gas pump, energy, word processing, paying your bills, all these industries have focused on data capture, and if you look at how that is done you don't have to go to the ATM and be able to provide the routing instructions, the information, the specific bank processing numbers and how to transfer money.
Yet what we are doing in health care today is you essentially have to do that if you are a provider and you want to interface with the data standards and work within the terminologies that are being used today.
So, if we think of this model and begin to look at where do data standards and where do coding standards make sense it helps at least in my mind to clarify some of the thinking.
If you look at the concept of data capture and you look at a process of a clinician using a system, it is very, very beneficial to have that be able to be created and used in a manner that the clinician is familiar with and comfortable with, but the processing and then how that information is used internally in the system may be a very different set of standards or communication for exchange protocols, and we tend today to blur those two. We tend to bring the internal processes and the data standards that we used internally and we push them out to clinicians. We push them out to the users of systems. I think one of the things that the Committee needs to look at is where we draw the line, so that if we talk about data standards, and we talk about standards development organizations, we need to clearly define which space it is that they are playing in; are they actually doing messaging and transactions between systems or are they actually looking at content and capture at the point of care, because what we have seen in a number of sort of amalgamations and offshoots of various developments now is that these things blur together and in the blurring there is confusion, and as Blackford and Dan and Paul mentioned in the previous session our ultimate customer is the clinical end user, and we look at obviously providing better care, better access, higher quality, lower costs.
We believe you have to get out to the point of care to be able to do that, and if you are taking system- specific or messaging-specific or technology-specific interfaces and making those be pushed out to the end user that is a losing battle.
I will give you a few thoughts under a series of issues that prevent or delay the development, implementation of these standards. I think that we need to look at the dissociation of the content and the messaging formats that we use and essentially looking at which standards focus on which areas and how we do that.
It is, also, clear that when you look at the data standards for content and for messaging you have to focus on who is the customer; who is using those. Some of the HL7 work is very, very different than the SNOMED RT work, and if you begin to say that one size fits all you are going to end up having a mess that nobody will be able to use.
I think that in many cases today the ability to create these coding schemas and the electronic standardization has exceeded the pace of our ability to build systems that make sense of that for clinicians to use, and I think if we look at some of the questions about clinical data and where clinical data stands there is basic disagreement on what is the electronic medical record; what is the granularity of that; how small do we get down to; how far are we going to push this out to a single element?
Are we going to describe what a chart looks like, what are the components of a chart? Are we going to talk about the messages within that? I think that one of the areas that we need to focus on is the construct of the medical record and then be able to fill in pieces in terms of how the content and how the messages are passed back and forth within that.
I think a key thing to remember is that we are going through this transformation in health care of going largely from human readable processes to machine-readable processes, and that that is not an easy transition, so that when we look at what has been done at least to date much of it has been administrative billing and back office automation which are machine processes but are not based on clinical data at the point of care, and now as we begin to get into taking what is human readable, the encounter documents, transcription and be able to turn them into machine readable processes we are going to have some friction. So, it is natural to be where we are right now, and it is good that we are talking about it.
I think if you look at Page 5 and 6 there is a fair amount of detail in the handout, but I wanted to lay out for the Committee a few thoughts on what would be not the specific standards to use but what framework or what type of guidelines should be supported. So, if we were to bless or talk about standards organizations, how would we like them to function?
There are about 18 or so points that I put out, and I will walk you through a few of them, but the concept is that we want to have this obviously have a code assigning authority or have this centralized in a way that we have the ability to have a global coding or coding structure get out. We would like it to be available in the public domain at as low a cost as possible, a nominal cost, if possible.
It is important in this process that the code assignment process and in fact the work that is done is open for review, comment and participation by the stakeholders. The "Thou shalt obtain and use this type of coding scheme" doesn't fly. We have to have participation from all the significant stakeholders.
It is important to have it be frequently and continuously updated. Ideally there are some structures within it that would support classification and different types of hierarchies that make retrieval and using the coding schema valuable.
There should be essentially a series of user friendly tools that let you be able to look at synonyms, equivalent meanings, looking at concepts, being able to determine what is equivalence as we look at these different types of coding schemas, and again I am just highlighting some of the issues that we point out in the written format.
Updates need to be available. They have to be able to be accessed. Updates need to be done frequently. They need to be forward and backward compatible so that we are not locked into a single version of a coding schema that changes and then everything becomes obsolete. Don't delete identifiers. These are sort of behaviors that if we look at many of the current potential standards or coding schemas we look at don't obey all of these, and one of the things that might be worth considering from this group is putting out the framework by which these standards organizations have to behave, that if they are going to be putting things open to the public or getting feedback that they begin to look at some of these behaviors as ways of behaving.
I think the final thing that is important is to look at the domains that they cover. There has been a lot of stretching or extensions of coding sets beyond where they initially were useful, and again, I will let Tim get into the whole NDC world and what we see with using drugs and drug codes at the point of care.
I will give you a model that I think is worth looking at and talking about in the pursuit of this, and actually not knowing that Blackford was a budding geneticist in the previous presentation, one method to think about is the human genome model of government, industry and academia collaboration, and if you look at that and you look at processes that are put in place and infrastructure common sort of baseline foundation and then a series of goals and targets then you can begin to have a framework where you have collaboration between government, academia and industry.
I don't necessarily believe that it should all go to the government or all to industry, but if you look at the processes and we sort of agree on what the playing field is then we have much better chance of being able to drive this, and I mean the genome is an example where you have centers of excellence and people that are focused on different areas, but they are sort of a common model and you are driving towards a common sort of synthesis.
I don't believe in a creationist top-down design. I believe in evolution and that it is going to evolve, and I think we would not be wise to invest a tremendous amount of time or effort right now and say that here are the specific details, and we are looking at where it is going to go.
I think the things that are important are business capturing clinical data is extremely important and the best thing that we see out there today is SNOMED, and you heard about SNOMED RT. Kent was talking I believe with you all yesterday.
Clearly it is not a one size fits all, but it is a beginning point, and some of the processes that we believe are important that I highlighted a few minutes ago are processes that Dave adopted in terms of how they are looking to migrate and move their standards or their terminology forward, and I think that other organizations could be well suited to follow that lead.
In summary I think there is a tremendous amount to cover. I am less concerned today with the specifics of which coding sets are out there, but I am concerned with the process and the infrastructure that we put in place.
I think that the Committee has the opportunity to be able to set the framework or the guidelines for that, and I would like to see an industry-specific partnership along with government, along with academia in a way where we can actually begin to then evolve this and migrate it. We look at the dollars that we spend on health care if we were to take a percentage of those dollars and put those into information technology and optimizing our processes. As Blackford said, we could put an informatrician(?) on the moon in less than 10 years I am sure if we really tried to.
So, there is plenty there to go through. I look forward to the opportunity to answer your questions and thank you for the opportunity to speak.
MR. BLAIR: Thank you, Jon Morris.
Dr. Peters?
DR. PETERS: Thanks, Jeff, and thanks to all of you for listening to me particularly after hearing me several times on security and confidentiality. I appreciate your inviting me back on a different platform and a different discussion.
I wish that those battles had been won and all that had been taken care of, but we will continue to do that at a different agenda.
Not to belabor the importance of security and confidentiality, but in addition to promoting that I have, also, shifted some of my professional efforts now more towards both within ASTM towards XML and SGML and, also, at the other side of things, and I will explain some of that as I go through, away from just an orientation towards clinical information but, also, an orientations towards what we need to do administratively.
As I have changed my company alliance having formerly been with Oceania and now founding a new company, as we go out I switch markets from the integrated delivery systems out to the so-called "real world" which is practicing physicians in independent medical practice and IPAs. All of a sudden clinical information has validity but only where it ties back into administrative issues and those directly tie into other HIPAA-specific issues particularly in the areas of financing relating to getting bills paid, doing coding, tying up specific reasons of why people have done specific diagnoses, also, being involved in CalLinks(?) I am a CMA or California Medical Association representative to CalLinks. CalLinks is as I mentioned before on security and confidentiality, they are a consortium of payors, employers and providers in California directly addressing issues not just under HIPAA but really the data interchange issues that are required to provide a cooperative environment between employers, payors and health care providers, also, tying in the vendors as well to say that at least from a state level even if nothing is done at the national level we need the kind of coordination that you are all talking about here; also, recently have joined the AMA CPT5 subcommittee looking at restructuring of that among the formats and structured work group, find that very interesting to hear the physicians talking of what they need and what they want versus what we look at in terms of the current structure and the way the current structure is put together.
In addressing the questions that you put before us and particularly whether we should give you our interpretation of the congressional instructions, I certainly hear and have read the testimony in front of your Committee from September and have an understanding certainly that this is advisory in nature and probably has less teeth than the HIPAA initiative that related to financial transactions. It is one of the issues I think we need to address and certainly many other speakers have and that is a more proactive role may be of benefit and that is one of the things I think I will push because I think if we look beyond the way the wording of all of this legislation and the various instructions come down we need to look at intent, and I truly believe that the intent is to try to foster some exchange of information that is going to be to provider and to patient benefit, and as far as we need to go may or may not be mandated by law or legislation but I think we need to make some professional decisions about what the best way to obtain the benefits of standardization would be.
What do I think are the issues preventing or delaying deployment and development or widespread implementation, the word you used to this thing? I think there are three primary entities. I think this has been addressed in some respect by the other speakers, but I would like to address them, also.
One is the health care information systems industry. I think we need to take a significant amount of blame, and I will go on to some of the reasons for that. Also, health care providers, I think there was some discussion from the prior panel that there is some level of resistance. I think Dan Russler put another point forward though that that resistance is in some areas, but certainly there are people, and the 200 clients they have using inpatient electronic medical records are an example.
There are people interested in doing this, but they need things put together so that there is an economic and an administrative and clinical benefit to it.
Also, I think the payors, and in that I include the Federal Government, not to criticize but Medicare and the issues that revolve around the concept of how you pay for health care, I think this has caused a problem, not necessarily one of explicit blame that people are going to stop the implementation of electronic records and the use of clinical data, but I think there are lots of factors we need to look at as health care reimbursement changes that will affect the way that we can implement and deploy these systems.
If we look at the health care IT industry and particularly the EMR industry I think it is important and I have been asked to sort of bring a reality check on that.
Being a start-up we have been out there raising money, and I can tell you for a fact that this is a moribund industry. You are talking here to the medical directors of most of these companies and not their CEOs and CEOs are having significant trouble raising money either from industrial sources or from the investment community. It is a sector that has basically soured because they haven't seen and what investors expect is a rapid turnaround and rapid profit. They are seeing nothing in the health care industry like they have seen in the Internet industry, for general commerce and other things like that where they can see rapid return on their investment.
Therefore, they are souring on putting money directly into the electronic medical record industry as a whole and as my talk progresses I will talk about some of the newer vendors that they are funding which are basically taking this to the next level and I think will have a profound effect although none of them are represented at the table at this time.
I think we need to as vendors, and I have been very involved in this industry for a long time in electronic records, we need to take the blame. We are not providing what physicians and health care institutions need in the way of software. We are very involved in looking at the needs of integrated delivery systems. This is not the majority of delivery of health care in the United States and it is not going to be if we look at trends in reimbursement.
We are looking at managed care. Managed care is on its death throes in some of its payment models, particularly in capitation and particularly in the primary care gatekeeper model. This is something that many of us as vendors have concentrated on, and we think that managed care capitation risk and primary care gatekeeper will be the wave of the future. Therefore we built systems and systems to manage referrals, systems to manage capitation, systems to manage the concept of controlling access to services that actually is not proving out as we return to a sort of mixed fee for service model.
The other is that 85 percent of medicine is becoming outpatient. We know that for a fact. If you look at what we admit to our hospitals now particularly in the State of California where you have got to basically be dead before you can get into a hospital, it is important to understand the vast majority of care is shifting to outpatient.
I recently spent a few days in Miami with someone who was critically ill and that patient had
been in the hospital for a week before being admitted to the ICU. Regrettably in California that patient would have been an outpatient and would have been admitted directly to the ICU through one of our ERs because of the way that we are changing the way that we manage care, not necessarily good but not necessarily bad because we are preventing some things that we know that hospitalization has caused in the past, but the other issue is that we know still that 80 percent of decisions on all of the expenditures that we make in health care are still essentially made by physicians regardless of the control of physician behavior.
Physicians are still the ones who are prescribing. They are still the ones who are ordering tests. They are the still the ones who are admitting patients and discharging them from the hospital.
Also, what we know for a fact over the last year and one-half from the employer's, the payor's and the provider's perspective is that cost and cost on a daily basis, not cost projected 2 or 3 years down the road, not return on investment on investments made now but the cost of providing care and the profit margin related to daily visits has become the critical driving factor.
This has had a profound impact on the ability to sell systems in there and to make people make that, if you will, quantum leap and assumption that there is something different that they need to do.
It is difficult to do that when you are looking at your shoelaces trying to deal with Y2K and specifically trying to look at the cost of providing health care.
Health care providers, again, I think that both Paul Tang and Dr. Middleton mentioned it in detail, there is a conservatism and a difficulty in getting large organizations. I think Jon Morris in some of their work both with the Federal Government and with Kaiser could reiterate this.
There is a difficulty in getting large organizations to move quickly. Now, that is understandable and there is no blame placed there because these are expensive decisions, and they take a lot of sea change to make them happen, but it has, also, been difficult to get a lot of results. If you look at the results from the electronic medical record industry, we are still looking at the work that was done by Reed Gardner and his group and Clem McDonald and his group to really prove that there is a lot of clinical benefit to these things, and that was work that was beautifully done, but it was done a while ago.
We have not had the output that we should have had from the use of these systems and the benefits to the provision of clinical medicine to patients that if we had collected this data and had electronic medical records in place would have been of benefit.
Again, there are too many failed projects, and this is painful. Vendors have had disasters and every one of the vendors here can speak to one disaster. For example, the core site for Pegasus which is HOB's product is a failured project from Oceania that I was directly responsible for historically.
That is a hospital that had one of our products and actually switched to HBOC and has now had a successful project with that. Paul Tang's job with Palo Alto Medical Clinic, that was a MedicaLogic project that was won fair and square that was replaced by Epic when Setterer(?) acquired Palo Alto Medical Clinic, and if you look at Epic's success which has been in Oregon for Kaiser Permanente that will eventually be replaced as will IBM's successful project in Colorado with products from Oceania. So, it all comes full circle, but the issue is that failures have not helped us because there are, and I think this is fair to the people who are purchasers, there are lots of jaded people out there, and I think they have a right to be because we haven't served them well by really being successful out of the starting gate with our implementations of electronic records.
Y2K is certainly an issue for the larger organizations. We don't see that in the outpatient ambulatory market. We see a lot of frustration with the practice of management vendors because they are having difficulty with Y2K, but these are systems literally that the practice will throw away. They are very much ready I think in that market for a change. The trouble is they have a dearth of products to say the least.
The other is wildly expensive infrastructures. It is very, very cost prohibitive to install a lot of these systems, and I think the question of early on return on investment is key because too many of the systems currently would cost too much to implement.
Health care payors, I think in fairness they are a critical element here if we are going to tie in valid use of clinical information when we look at the bottom line and that is the providers are looking at getting paid. There is a critical impetus for them to implement systems. One, most of the health are payors are putting their energy into HIPAA legislation for financial transactions, the transition to X12837 and also, dealing with Y2K with their legacy systems.
They made their decision a long time ago not to invest in new technology. They have a lot of main frame systems they need to preserve. So, they put a lot of time and effort into that.
Two, they are suffering significant heartburn over the failure of managed care. They have had a number of down years. Most of the major indemnity players have pulled out of health care. It is mostly the newer organizations, the managed care organizations that will be the payors of the future. They have had negative revenues over the last 2 years as everyone knows. Although they will have a turnaround this year with everyone getting significant increases in their monthly benefits, therefore, this will be probably a record year for profits in the managed are industry according to analysts on Wall Street. So, we will see a turnaround from 2 disastrous years to record profits, all because they have been able to raise premiums, but that turns their perspective away from the issues of implementation of systems and more towards the bottom line. How can they become profitable and profitable for their shareholders, which I think is regrettable to all of us in the systems industry, certainly regrettable from a patient or physician perspective personally.
The question is the role of the private sector versus government in regard to implementing this, and I have a relatively succinct answer. I think for the private sector the answer is no, and I would like to reiterate that for a number of my colleagues. One is vendors are too protective of their proprietary advantage, and I think Dan Russler's perspective that there are parasites out there who would like to leverage off the good money spent by the other vendors I think is a fair perspective.
Everybody in the vendor community feels at some level that the preservation of their proprietary perspective is of some benefit. That is a detriment I think all of us know in coming up with standards, and it is something we need to address.
Two, and I think that Dr. Russler addressed this as well, and he did it in very coached terms, and I will try to use the same. There are too many people still trying to make careers and money off of the standards process, and I think that is sad because I think the perspective that the standards organizations are fighting each other or trying to one up each other where their organizations particularly in terminology are trying to outdo each other or somebody is trying to make money off of terminology and somebody else isn't has been very destructive for the cooperative efforts that I think as Michael Fitzmaurice mentioned we are trying to do under HISB and now under the new ISO efforts internationally.
We need to get away from people making an independent career out of the standards process to make a name for themselves. We need to get out of the notion that there is money to be made in this or that there is a proprietary or corporate advantage to doing it and get back to a position where what we are doing is saying that standards are of benefit to everybody, to lower the bar to entry and to provide interoperability.
I think the role of government therefore, is to influence a process but probably not to set the standards, and I will give an example there. I think that effort needs to be placed and specifically HIPAA is an example of that, particularly HIPAA related to financial transactions where the Federal Government did have the draconian ability to step in and say that this will be the standard if the rest of you don't get together and get your act together.
I think that it is regrettable again that the issue with clinical data is it is advisory and doesn't have the teeth that the financial transactions element in HIPAA did. I think that we can sort of hopefully move in that direction because I think the more that is not mandated but the more that is influenced by the threat of being mandated will have a profound benefit in terms of influencing industry.
The other is in terms of not setting the standards. I have to be honest and not to offend those of you who are involved in that, it is not my intent but I am personally stunned by the government GHCPR effort because what I see is and I will tell you both my taxpayer paranoia and my industry paranoia is that we will see a repetition of the CHCS project because again you are bringing in outside vendors when there are good vendors in the industry who could help solve this, and, also what it does is it is flying in the face of what is really happening in the IT industry which is away from single source data interchange to wide open standards, and I think I would like to use that as a way to sort of lead into the rest of the discussion, and there is a whole slew of vendors not represented here that are doing exactly what the GHCPR is talking about funding. Those are specifically companies like Healthion(?) Cinatics(?) Azosterion(?), Aboton(?) Dotcom(?) Digital Medical, Care Flow Net. All of these are smaller companies, some of them very well funded such as Healthion which is the darling of the industry who are looking at data interchange and being the data intermediaries which is exactly the type of thing that GHCPR is looking towards.
Again, the key I think is that the government needs to push the private sectors and the STS to finish the work, needs to use the threat of mandate as the lever to do that but we have to have private sector involvement, and I think what is going to happen is we are going to have to have private sector involvement outside the health care IT segment to make this a reality.
Again, what standards? I think that interoperability is the key and I think that if you look at content which many of us talk about one of the things that we have missed in many of these discussions is content is best managed by the medical subspecialty societies.
Those of us in the health care IT or in the government or from the payor side mandating what information should be exchanged from the clinical care of patients, be they pediatric or emergency medical cases or be they someone transferred into the ICU for critical care, we need to look to the subspecialty societies and medical knowledge from those groups to start to talk to us about common data sets and other things like that, i.e., what data should we exchange at any particular time.
For syntax I, personally, and many of you know my bias, I think XML SGML is the wave of the future. We need to face up to that, step up to the plate and move as quickly as we can.
I think if you look at the HL7 special interest group and the impact on HL7, HL7 has come out and said that version 3 will be in XML syntax. It is a done deal as far as I can understand.
SIN(?) has stated categorically that they want all of their object models and specifically all of their data transmission to switch to XML. They are actually pushing EDIFACT(?) on the financial side to do the same.
Therefore we can have a uniform syntax. ASTM's recent efforts as you know in transcription are working with JCHO and others to try to get uniform document content. We will also be, if we can push an indirect direction based on XML. Also, we have a new work group under ASTM on DTDs which would define health care documents and their structure and what we did was we tried to do what HSB has been trying to do as well.
The two chairs of that organization within ASTM are the chair from the HL7 special interest group on XML and the chair of SIN's special interest group on XML.
So, now we have a different standard organization doing a different part of XML. We have the chairs of the other two organizations who are running that who will make sure that there is cross communication and that there is commonality to that work.
Terminology and coding, again, many of you know my work as well. I would agree with Dan Russler and some of the others. I support the SNOMED RT effort. I do not think it is practical for those of us who are vendors who are implementing systems. We have been involved with David Roth, one of a variety of others who used to be with SNOMED and looking at a different system which is based on XML and SGML and is really inoperable, an attribute-based system for vendors.
Again, it is a non-profit approach. It is the notion of doing cooperative work and sort of lowering the barriers so that nobody can use a common thing. We are going one step behind an intellectual approach which is the way I look at SNOMED RT or in many respect RAID(?) or in many respects the work of a number of groups in Europe and going into saying what vendors need is really practical terminology, things that they can use in exchange so that they can exchange content.
Modeling and data models, one thing I know and I don't want to be offensive to other people here, too, because I know this is a pet issue to a lot of people, it is something that I think the industry will work out, and again the more we look at XML and SGML and the more we look at data object modeling and using that as a format which is basically what is being done in the general IT industry and in commerce the more we can get away from the concept of modeling and get back to the exchange of data and exchange of content.
So, I, myself, would belittle that as a primary issue and say that the data interchange both in terminology and in terms of structure and syntax will be the primary thing.
Lastly, the business case for rapid standards development, one, providers I think particularly physicians in this country are ready for a solution.
As I have gotten out of the integrated delivery system market and gotten into the private market I am stunned at how much private physicians are willing to implement EMRs and make them run.
Our sales cycle averages 2 to 3 days. Our installation cycle averages 1 to 2 days to get a practice up and running, and we offer dictation, and we offer transcription, and we offer language recognition and all these various things like that. They do structured data entry.
I am stunned because we had a difficult time as a vendor, particularly while I was working with Oceania getting large organizations to face up to that. Private physicians say, "Hey, it generates my bill. It does it straightforward. I don't have to do extra paperwork. It decreases my overhead. I will do it. It is a fractional cost to me in terms of daily experience."
So, I think we need to look at the industry as hospitals, integrated delivery systems and then physicians and physicians still make up the vast majority of the market and still a vast majority of the decisions on these systems.
Also, patients are getting poor quality care, and we know that, and we all know intellectually that electronic medical records exchange of data would be of benefit. We have got to confront the confidentiality issues.
We have, also, got to get interoperability in exchange of data because we, also, know that our patients do not stay with one physician. They don't stay with one physical entity in terms of health care. They don't stay with one payor.
So, it is adamant that we do that. Also, standards decrease the barriers of entry, and I think that Dan Russler mentioned the concept of parasites. I think that is fair. I, also, support parasites. If I can do work in the standards world and make things happen and make it easier for other people to come into this industry and build companies and if that will accelerate the efforts, I think that is a beneficial thing to do, and I think as we move forward that is a core issue that we should address.
So, in closing I think I strongly agree with the four areas of emphasis. Two, I think that the government ought to play a very proactive role and ought to preempt the private sector at least in doing the same efforts that were done under HIPAA for financial transaction which is using the club or the carrot and stick approach which is to say, "Here is the carrot. Do it. If you don't do it we will beat you with the stick." I don't think that is a bad idea, but I would like to do one other thing, too.
Let's be progressive about it. Let's look at the general IT industry. Let's look beyond health care IT and to close I would just like to mention codes and code sets.
That is used very much in the discussion of this work group. I think we need to get away from codes and code sets. In modern architecture there is no real need to take data and match it up to a numeric code or talk about specific code sets. We need to start talking about content. We need to start talking about common forms of using it and exchanging it. There are no real numbers used in Internet commerce anymore. There are no codes. There is no breaking things down into anything other than dollars signs and dollar amounts. It is all context-based data based on data models that are related to either XML SGML or HGML. The standard relationship models are object models in the background. There is really no use of numbers like we use in medicine.
So, what I would like to do personally is see us talk more about content, move away from code sets and start talking about what we need to do to exchange data for interoperability and move forward as rapidly as we can with the Federal Government as a lever.
Thank you very much.
MR. BLAIR: Thank you.
Dr. McNamara?
DR. MC NAMARA: My name is Tim McNamara. I am a physician. I am the Vice President of Research at Multima Information Services which is a wholly owned subsidiary of the Cerner Corporation, and I am representing both those organizations today. The organizations I represent and I very much appreciate this opportunity to testify and to provide our industry-oriented perspective to this Committee on the reporting of the development of uniform data standards for medical record information, electronic transmission.
The Cerner Corporation is a leading supplier of clinical management information knowledge systems for health care organizations in the United States and abroad. We have more than 1000 clients in the US and around the world. These clients include integrated health organizations, integrated delivery systems, community hospitals, ambulatory clinics, physician practices, blood banks, reference labs, etc.
Multima Information Services is ordinarily focused on the creation, maintenance and distribution of clinical information, including in particular drug information. As an organization we have attempted to create systems that reduce the frequency of adverse drug events, a national problem that kills about 60 to 140 thousand Americans each year and costs the nation about $76.6 billion each year.
Much of that expense is shouldered by agencies of the Federal Government, a number of which are represented here today. Multima's clients include Cerner Corporation, of course, and a large number of other prominent health care information suppliers and health care provider organizations, more than 1000 organizations license drug information from us. Private and governmental organizations like HCFA, for example, license our data and again are represented here today.
The focus areas of this hearing cover topics that are of direct business interest to both us at Cerner and Multima. We have created a large number of products and services that rely upon the kinds of clinical messages, patient medical record information and data dictionaries that these focus areas address.
We, like many organizations in our industry have faced significant business challenges in the area of medical vocabularies. The problem is not necessarily the absence of standards but more often the proliferation of numerous disparate, incomplete pseudostandards which in almost all circumstances either lack the clinical depth or wide acceptance in the industry needed to facilitate interoperability.
Because the problem of space regarding medical vocabularies is so broad and because time is short, both in this hearing right before lunch and the implementation of code sets in general needs to be quick under the timetable as outlined by HIPAA we would like to focus our contributions in this area to a particular topic which involves drug vocabularies. We think it is important to focus on drug vocabularies because it is an area where there can be significant achievements I think fairly rapidly made and from a practical perspective we believe that such improvement can be made with relatively modest effort, resources and leadership from the Federal Government.
First, a little bit of background. As you know the NPRM regarding standards for electronic transactions and code sets that was published in May of this year proposed that national drug codes, NDCs become the national standard for transactions requiring drug codes starting in the year 2000.
We believe that the national drug code system has both important design problems and fundamental shortcomings that will in fact both hinder HHS and the industry at large in achieving a number of the important goals of HIPAA and significantly and importantly increase the administrative burdens placed on health care providers.
In June of this year several Cerner colleagues and I commented on this NPRM on this topic in depth. We described 12 important deficiencies of the system and one fundamental problem with the system. We, also, proposed 13 criteria that any drug vocabulary system for national use should meet, and some of those, also, I think appeared in Jonathan Morris' testimony earlier today more broadly talking about other sorts of code sets.
Lastly, at that time we proposed specific actions and language for the Secretary's consideration on the topic of NDCs. Our intention today is not to reiterate all that material in detail because it is available on the HHS administrative simplification web site, and a number of you may have seen it, but we will be glad to submit a copy if those comments might be useful as written testimony.
Instead what we would like to do today is to focus our comments on the NDC system and very quickly summarize some of the problems and then discuss specific approaches to the problem of drug vocabularies and especially how we think government can provide leadership and direction in this critical area of health care activity.
As we mentioned in June and very briefly the NDC system is hindered by a lack of central code assigned authority, the lack of a common interorganization meaning for each NDC, variable numbers of digits and variable structure of the internal components of each code, assignments of NDCs for large numbers of products that clearly are not drugs, absence of NDCs for a large number of products that clearly are drugs, absence of a public open process for code assignment and absence of regular updates to public NDC lists, inappropriate blending of multiple other coding schemes like health-related GPCs and so forth, all of which are 11-digit codes but have different characteristics.
In addition the NDC system is based on one deep shortcoming, and that is that the assignment of codes occurs at the level of the drug package. The NDC system was designed for health care providers who manufacture drug products or pay for drug therapy. In this regard I think the NDC system can be considered to be a marginally or moderately successful code set even given the deficiencies that I have just described.
However, the design of the NDC system is completely inappropriate in our view for the needs of most health care providers, especially those providers like physicians, nurses, physician's assistants and even pharmacists who prescribe drug therapies, dispense drugs or administer medications to patients.
The NDC system is, also, inappropriate for administrative functions that involve data aggregation and analysis even regarding the simplest sorts of queries that people might want to do. For example, NDCs alone without reference to other external systems using those alone it would be impossible to locate all the patients on a plan who might be taking for example, fluoxetine, SSRIs or antidepressants in general.
The central problem with the NDC system is that it identifies drug products at the level of a package and that is much to granular to be of any practical use for health care providers. Health care providers, especially clinicians need access to a drug vocabulary that supports multiple levels of abstraction of drug products. In other words, they need a national code set for drugs that supports descriptions of therapeutic categories like macrolide antibiotics, descriptions of drugs like erythromycin, descriptions of drug products like erythromycin ethyl succinate 250 milligrams oral tablets and descriptions of manufactured products which are much more specific, things like erythromycin ethyl succinate marketed under a general name from Heartland(?) packaged as 2000 unit dose 250 milligram oral tablets with an NDC, for example, of YXY.
The NDC system only identifies drug products at this very last and very granular level of description or detail.
The fact that the NDC system is much too granular for most health care providers has important consequences for the implementation of HIPAA in our view.
If NDCs are required for electronic transactions involving drugs, the administrative burden on typical physicians who might elect to use electronic tools for prescription generation becomes unmanageable. Such physicians would need to specify the drug, dose, route, frequency of administration and duration of therapy on a prescription pad as usual. That is not a problem, but they would, also, need to specify the manufacturer or repackager of the drug, the type of packaging and the specific number of tablets or the dose forms in the bulk container from which the pharmacist might actually fill a prescription, and since physicians have no way of knowing specific manufacturers, repackagers, specific packaging, specific container size stocked by the pharmacy that might fill the prescription we think that the likelihood for the use of an NDC will facilitate physician ordering and administrative communications between physicians, pharmacists and health plans is quite small.
In fact, a requirement for use of NDCs in this regard I think will actually increase confusion, administrative burdens and the need for communications between various types of health care providers.
We propose that the following criteria should be used for evaluation of code sets, and I will review them quite quickly. First the code set must be maintained by a central code assigning authority. That is a theme that I think we have heard well addressed this morning, and we will talk about that in a little bit more detail at the end.
Second, the code set must be available in the public domain at low cost or no cost.
Third, the code assignment process must be open for review.
Fourth, the code set must be continuously updated as new drugs come on the market.
Fifth, updates to the code set must be made available to the public on a regular and very current basis, weekly or daily, perhaps.
Sixth, once a code is used it should never be removed.
Seventh, each code assigned must be unique, a fundamental problem with the current NDC system.
Eighth, the structure of individual codes should not carry meaning.
Ninth, the code set should support multiple levels of abstraction as I have described.
Tenth, the code set should be backwards compatible with the NDC system but not reliant on the NDC system, and I think that this sort of requirement reduces the potential economic burden of implementing potentially a new code set for existing systems while providing much needed flexibility for future systems.
Eleventh, the code system should be designed with a long-range view towards compliance with evolving international standards around drugs.
Twelfth, the code set should include information regarding the regulatory status of drugs and drug products in the US, for example, like controlled substances act schedules, over the counter schedule, therapeutic equivalence and so forth.
Finally, thirteenth, the code set should deal exclusively with drugs not other sorts of medical supplies and devices.
We believe that most individuals and organizations who might participate would therefore like to discuss the who and how of creating a national standard for drug nomenclature that embodies these principles.
In considering the creation of a national standard drug vocabulary that facilitates electronic transactions, at least of the types that are contemplated I think by the authors of HIPAA it seems to us that there are four general ways to proceed.
First, the government could proceed with the creation of final rules that use NDCs exactly as described in the current NPRM. In this scenario private organizations could presumably continue to create a heterogeneous collection of proprietary drug vocabularies, a number of which might claim to be compliant in some ways with the NDC system.
Since, however, there would be no central list of NDCs maintained by an independent assigning authority the system that generated a transaction containing NDC would have no particular way to confirm that the system receiving NDC would interpret that code as being the same drug.
In addition this approach would probably deter the development and use of prescription writing systems by many providers, especially physicians and nurses for the reasons described above.
Because of the significant potential hazard of misidentification posed toward the drug-consuming public and because of the inordinate administrative burdens of this approach on clinicians we strongly recommend that this approach not be followed.
Second approach, the government could encourage industry to develop a standard drug vocabulary, and I think we have heard a lot of discussion about that this morning. To a certain extent the government is already taking this approach. Representatives of the National Library of Medicine, for example, regularly participate in the Vocabulary Technical Committee of HL7 which has focused some efforts in this area. Unfortunately, the scope of HL7's interest and its ability to commit to and sustain the drug vocabulary initiative are very much in doubt particularly since the organization if governed by consensus among its members, and a number of the members have proprietary reasons to object to or limit HL7's activities in this area.
Over the course of the last 16 to 18 months HL7's progress in the area of drug vocabulary has largely been limited to some general agreement on the important attributes of drugs and drug products but meaningful work toward the assembly of specific drug lists has not yet occurred and appears in our view quite unlikely.
To our knowledge no other broad-based industry led initiatives are occurring regarding standards for drug nomenclature.
Third, the government can create and maintain a new robust standard for drugs and drug products and require that it be used for transactions covered by HIPAA.
In fact, current US Code, Title XXI, Section 360 for example, gives the Secretary of Health and Human Services the option to assign listing numbers to certain registered drugs, and the Secretary has in fact, implemented this in some ways over the years.
However, the quality and usefulness of this assignment process at least as implemented by the Food and Drug Administration since 1972, has been very poor for the reasons cited above and in particular FDA has taken no significant steps to group registered drug products in clinically useful ways that support the business needs of clinicians and, also, the timely production of FDA drug lists has been abysmal.
For example, for the 10-year period from 1985 to 1995, the FDA simply did not produce any sort of drug list whatsoever. While it is possible that the government might provide leadership in this area, such leadership appears unlikely.
From informal discussions with the FDA's drug registration staff, for example, there are apparently no plans to broaden the scope of the NDC system to support the needs of clinicians. In addition, other federal organizations that have shown initiative in the area of medical vocabularies, like the National Library of Medicine have specifically elected to work with existing drug vocabularies such as they are but not to create or maintain an independent comprehensive nomenclature in this area.
Given the lack of specific funding for such efforts, the historical lack of demonstrable success, at least in FDA's hands and the government's lack of governmental leadership in this area, we believe that a drug vocabulary initiative sponsored solely by the government is unlikely to occur, but if it did occur, I think it might actually be a very good thing.
The fourth approach and potentially the most successful and practical approach we believe is for government to work with industry to create or select a robust drug vocabulary that addresses the business needs and realities of organizations and individuals who may wish to send drug-related transactions covered by HIPAA. There are several approaches to this scenario. One of the more promising might be for government to select a specific proprietary vocabulary and mandate its use, and there is some precedent for this action.
For example in the NPRM for code sets the Secretary specifically proposed that a proprietary closed and somewhat expensive coding system developed and marketed by the American Medical Association be used for transactions that involve medical procedures, namely, CPT.
Government might, also, subsidize the continuing maintenance of such a vocabulary and/or charge a nominal fee for its use. This model has been used elsewhere in the United Kingdom, for example, with respect to the TOREED(?) codes.
In the case of an American drug nomenclature we suggest that the Federal Government ought to evaluate existing proprietary nomenclatures and then select one as a standard for use in transactions covered by HIPAA. Criteria for evaluation of the code set should include at a minimum t he criteria described earlier in these comments and, also, alluded to by Dr. Morris but from a pragmatic perspective end user cost of such nomenclatures ought to be very carefully considered in the evaluation.
For our part at Cerner or at Multima, we are willing to assist NCVHS, HCFA, HHS and/or any of the federal agencies in this effort. Specifically we are prepared to offer the comprehensive drug vocabulary that we have developed as a by-product of our clinical drug information business to the health care informatics community and the public at large in perpetuity if the Federal Government would like to use it as a basis for a national standard vocabulary system.
Our system, called Multima Lexicon is current available for download and use by any interested party from our web site. We are prepared to offer the lexicon under a no-fee license agreement, and that is no fee for system developers and no fee for end users, and we believe that such an arrangement is an economic win/win for nearly all the parties involved.
We are aware and sympathetic to the fact that the creation of a standard in general in the short term can carry significant development expenses for organizations that are in the area, in the industry that is affected.
The creation of a national standard vocabulary for drugs is certainly no exception regardless of the quality and source of whatever vocabulary may be adopted, but we believe in the long run industry will benefit. If a rich vocabulary standard were created it would no longer be necessary, for example, for health care information systems providers or organizations they serve to create multiple point-to-point locally negotiated inexact mappings between a variety of disparate systems.
More importantly we believe that government selection or creation of a standard drug vocabulary will pave the way for widespread deployment of clinical alerting systems and thereby help to reduce the problem of adverse drug events. We believe that there is no code set problem posed by HIPAA that would benefit more from strong governmental leadership, and we urge this Committee and all federal agencies involved in the implementation of HIPAA to work towards the selection and creation of standard vocabulary to the benefit of all drug-consuming Americans.
MR. BLAIR: One moment here, if I may. We are running kind of late. Is the panel able to return after lunch or do you have flights?
(Administrative discussion.)
MR. BLAIR: Well, then, Gary, could we have your testimony, please and to the best you can please try to keep it within the 10 minutes.
MR. RADTKE: Sure. I would like to thank the co- chairs and the Committee for inviting Ford Motor Company to come and testify before you today. My name is Gary Radtke. I am the Health Systems Manager for Ford Motor Company on a global basis.
Many of you may be surprised to hear that Ford Motor Company has a global health systems entity, but we do, and I think as you hear my testimony today you will see that we have very much of an interest in this topic.
Why is Ford here in the first place? One, in general Ford is interested in standards and has proven that standards in a wide variety of areas are helpful. It has been specifically instrumental in helping develop some standards in the health care industry, and we work closely with some of our so-called "competitors," GM and Chrysler, Daimler-Chrysler to develop standard reporting mechanisms on health plans, on hospitals within Southeastern Michigan, and we are now expanding that out nationwide.
So, we do believe in standards, and we think that they are valuable. I think that you will hear that our testimony today is relevant because we do many of the things that have been talked about today that you may not be aware of, and we support the Committee's involvement in trying to come up with standards that everyone can live with and that would be valuable and useful for the community as a whole.
As we have gone through and we have listened, you will find that most of the testifiers have a very specific focused area. There are medical facilities and individual providers. There are health carriers. There are health care researchers. There are health care software vendors. There are consumers of health information. Interestingly enough Ford Motor Company is all those things.
I will give you some examples. For instance, in medical facilities and individual providers Ford Motor Company runs, operates 150-plus occupational medical clinics in 29 countries around the world. These clinics are designed mainly for occupational medicine, but we, also, tend to the personal needs of people, mostly our employees but in some cases dependents of employees in many countries throughout the world.
We have 150 full-time doctors on staff and we have hundreds of other part-time doctors, especially in countries outside the US. We have 200 full-time nurses, and we have other clinical staff. So, even though we may not have the volume of transactions or the number of people involved we have a significant presence in clinical management of our employees' health care.
In addition, OSHA requires that Ford Motor Company report all injuries and illnesses of a significant nature that occur on our premises. Also, in other places such as in the UK, RITTER(?) their governmental agency requires the same, and worker's comp in Canada and other places. There are regulatory agencies around the world who expect us to report on injuries and illnesses that occur on our premises.
How do we do that? We tend to us an on-line medical record. So, we are no different than many other hospitals, clinics and providers in terms of maintaining on- line medical records and using that information to report to governmental agencies.
We, also, have the same issues mentioned earlier of language translations and other rules in other countries that we have to deal with. As a health carrier we are a self-insured indemnity plan. We run and maintain that. We do that through intermediaries specifically Blue Cross Blue Shield of Michigan and Unicare(?) but we are the health carrier for our employees in the indemnity plans.
We, also, administer over 150 other carriers and 80-plus HMOs. So, in our view we are a health carrier as well as a medical facility.
We are, also, a significant health care researcher in that we have developed a quality consortium which will review and determine which types of analysis to do on our populations and to effectively manage a rollout of intervention strategies.
We have done things in the area of metal-working fluids, asbestos and cardiovascular disease with our internal researchers and in some cases have published those results.
We are currently involved in a flu study with the CDC whereby we are trying to determine the effectiveness of providing flu shots to people under the age of 65 and see if that is cost effective for employers or the community at large.
So, we are, also, a health care researcher. We are, also, a health care software vendor. We developed our on-line medical record using the most available technology that we had, and that occupational health and safety system is in over 60 medical facilities that we run in the US and Canada.
The technology is somewhat clugy(?) and I think you might find if you looked under the covers that it looks like a medical record that was designed by an auto company, but it is still something that we are quite proud of having delivered, and it has allowed us to meet our mandate from OSHA.
We are, also, a consumer of health care information. We provide mammography information and the benefits of doing mammography to our employees and our hospital and plan profiling efforts have been published on the web so that our employees can use that information in selecting hospitals and plans.
So, we are, also, a consumer of health plan information, and one of the biggest aspects of all of this is we are, also, as a result of being an employer a partial payor of all the systems that you have heard talked about throughout the last 2 days since we pay for that through some of our administrative costs.
We work closely with others in developing standards, but again, these standards are fragmented, and we work closely with the other autos and with other companies within our regions that we have a significant presence.
A question raised yesterday was should we have electronic medical records. I don't think that is the right question. I think that given the fact that companies already have electronic medical records, including somebody like Ford Motor Company, an auto and truck company and many more are being added to those companies every day, is there a role for government in facilitating standards development.
At Ford Motor Company we believe the answer is yes. We believe that we can facilitate the development of generic flexible standards, common outcome measures, common vocabulary, and the government as a payor can require that these things be used for its employees as well which I think has a tremendous ability to move the industry.
In regard to specific questions that you asked I will go quickly through those. How do you interpret the congressional instruction? I believe that Congress is looking for advice on one, whether to use its mandate to provide for specific standardization in the industry and two, should we recommend that it use that mandate, how should we recommend that it use that mandate, and clearly they are interested in creating value for the health care industry through the standardization effort, but I think they are looking for advice. Should they even do anything in this area? Is the time right now for this to happen?
We, also, feel at Ford Motor Company that you don't have to standardize all medical information. We feel that there are many things that the industry can develop that would be useful that using the entrepreneurial spirit will facilitate faster development, but there are many places where intervention is required.
What factors or issues are preventing or delaying the development and widespread implementation of standards? Standards are not difficult to develop. Standards are being developed all the time. The problem is we have so many standards and so many places with so many meanings that it is difficult to pass information or when companies merge to effectively merge that information together into something useful.
Uniform standards are much more difficult to implement. One of my favorite sayings, and I am not sure who I would attribute it to said that uniform standards are clearly a good thing. That is why we have so many of them.
However, you know part of the problem is that doctors and hospitals and clinicians in general practice medicine differently, and as a result of that variation they have built corresponding medical records or have bought packages that help them facilitate the way they do their work, and this provides a very big impediment because the rework involved in going to a new system is not trivial at all, and the vendor market for electronic medical records is relatively immature. So, there has not been a sufficient time for a clear de facto winner to emerge.
That may take time. That may take 10 or 20 years. We think it will eventually happen, but there is a lot of cost that will be involved for no good reason in the meantime.
Is the private sector able to address these problems satisfactorily? Due to the current weak incentive structure for collaboration the private sector will continue to proliferate standards. It may move marginally towards some common standards, but we think the regional standards and standards for specific focus areas will develop and not a generic medical records standard that everybody can say, "Yes, that is the answer."
I believe the government's role is to help facilitate a more rapid and orderly transition to a uniform standard, and similarly if the government can provide the impetus for moving to uniform medical record standards, more attention will be developed in efficiently transmitting electronic data and less time, money and energy will be spent on integrating new medical record standards into old systems.
In regard to what we think you should specifically do, we suggest that the government facilitate the development of the standard that is more generic and extensible than those currently in use. Instead of multiple standards a more inclusive approach to standards should be adopted.
Standards for medical records should focus both on data storage, the medical record itself and on data transmission. If these are not developed jointly, then we will be in the same situation on transmission standards for medical records after we have all agreed upon a standard for the medical record itself.
Also, in regard to do we agree with the emphasis on the four focus areas listed, in general Ford agrees that these four focus areas are important, but we are very committed to confidentiality and security of information, and we think that any medical record standard that is developed has to consider new rules and regulations that may come in this area and it has to effectively accommodate any of these rule changes.
Also, if uniform transition standards are not implemented then a variety of integration issues will arise needlessly on the back end.
In conclusion, Ford encourages the development of uniform standards. One can only look at the Internet to see what a good standard can do. The Department of Defense many years ago adopted a standard protocol for electronic communication known as TCPIP. Throughout the nineties many independent networking standards existed much like health care standards currently exist as every manufacturer had their own standards.
Many years of high cost and incompatible software existed before a single standard, the Internet replaced the cultured set of independent protocols with a single protocol.
All companies not using TCIP have now had significant rework to get into that arena, and some companies, some very large companies I might add are still not in that area and are falling behind because they threw their hat on a standard that now is no longer available for use.
A flexible data model with common definitions would be a great start. An inflexible standard that takes years to modify is a disaster waiting to happen. We work with many groups and coalitions, one of which is the Michigan Health Management Information Systems Coalition which is a subgroup of the Greater Detroit Area Health Council. They are helping us develop electronic standards, but what we found there is that in order to pass data effectively you have to somewhat get around the standard and use reference lines and other lines in ways that were never meant to be used in order to effectively transmit data.
As a result I think even within that standard we have multiple standards. Uniform standards are not easy, but they are worthwhile. Oh, and don't forget about employers such as Ford who have medical facilities and need to have our needs considered whenever you are developing a medical records standard.
I think the same would be true for the Department of Defense and the fact that they run hospitals as well, that they have to be included in that and so those groupings should not be excluded from this process.
So, again, in conclusion I would like to say, "Thank you for inviting us here. We encourage you to continue in this effort. We think that this is a good area to focus on, and we would like to provide our support in this area."
MR. BLAIR: Thank you very much.
The diversity of perspectives that is on this panel and that has been on the panels earlier this morning and yesterday has been very, very helpful in helping the Committee understand the whole issue of data standards for patient medical record information and the exchange of that information.
I want to thank Michael Fitzmaurice for taking the lead in pulling together these panels at this point. We don't have a whole lot of time at this point.
Could each person on the Committee, if you have a question, we will go around on the Committee members first and then the audience, and, Simon, you could probably do this faster than I can. So, maybe you could take the lead on that, please?
DR. COHN: Okay. I have a procedural question both for Tim and Rick, and I noticed that you both had testimony. However, it was not in front of me. Will we have copies of that or is it back there and I just didn't see it? Okay. I appreciate that.
Actually I asked this question to the first panel today, and I will ask this one again to those of you who want to respond, and it has to do with issues of aligning incentives.
One would observe that somehow the incentives to move forward with uniform standards seem to be somehow -- it doesn't seem to be quite there, and it seems to encourage a diversity of activities as well as incomplete standards.
What role do you think the government has to help promote that and what is it that we should do or recommend to Congress and the Secretary to help promote that?
Jon, do you want to start out first?
DR. MORRIS: Sure, I think that the incentives, and I will keep the response relatively brief, the incentives need to be aligned. They need to be things that are tangible for us to be able to work with in the short term.
Everybody wants to go for long term, make the world a better place to live in, and that is an incentive that we all hold holistically, but that is hard to put on the bottom line on your quarterly reporting. So, if we are going to look at incentives or ways to work, I am not necessarily a big bang type supporter, but I think that if you get a momentum, if you get a large group involved you can rapidly sort of set what the de facto standard will be without having to be overly prescriptive, and the incentive there is to not be left behind, and so it would be getting large groups, large organizations, the major employers essentially to get together to drive this and say that we are going to adopt this. We are all going to adopt this, and if you don't work within this standard or work with us to develop the standard you will be left behind.
I think there is a tremendous incentive there for companies of any size to participate. I don't think it has to be a financial but you can use a subtle coercion.
DR. COHN: Okay, Rick, comment?
DR. PETERS: I would concur with Jon. I think subtlety is a strong benefit. On the other hand, I think that if you look at, again, to reflect back on the HIPAA legislation, financial transactions that was a little perhaps less than subtle but it wasn't something that people walked away from and said that they wouldn't do. In fact, we had some payors saying that they would be willing to comply with this before the legislation is even mandated, that compliance; that is a good thing.
What you dealt with there was data interchange, and I think if we can concentrate up front on data interchange between EMR vendors, also, which is to force that issue that patients move between entities and use that as a primary theme we put initial standards on, I think all of us intellectually know that if you standardize that data that goes between that is going to reflect back on our fundamental architecture and back on the fundamental systems themselves, and as Jon says, "That is less restrictive," and I think a lot of people mentioned it is less of an impact than if they have to change their proprietary architecture, their proprietary advantage, their end user interface, their data models, other things like that.
MR. RADTKE: Based on my conversations with a lot of different groups that Ford deals with I would not underestimate the ability of the government to influence opinion by saying, "Here is what we are going to do for our facilities. Here is what we were going to do with people who deal with our employees. We want you to do X for our employees." Amazingly enough in terms of Medicare and other things when the HCFA 1500 form came about people immediately then said, "Okay, we have something that at least we can move towards. It is not perfect. The government says that they are going to use it for their transactions but even though it is not perfect that is what we have to use and let's go for that, and we can all agree upon that."
As a result you have seen the number of forms and billing activity coalesce around the HCFA 1500 form. So, I would say, "Don't underestimate the ability of the government by saying that here is what we are going to do when we have our employees go to certain places, this is what we require." That will have a huge influence over the rest of the industry.
DR. MC NAMARA: I would like to answer the question about how the Federal Government can align industry with incentives narrowly on the area of drugs, and I think that is an interesting area because in general there is a lot of alignment already. Everybody needs a drug list. Everybody wants one, wants to have access to a low-cost one. There is only a handful of organizations in this area that really do need to be aligned and they haven't been able to do so. Those are the organizations that create drug vocabularies like my own.
So, I think the Federal Government could have a direct effect by aligning this very small club, this very small set of organizations that develop drug vocabularies by insisting on certain standards in that particular arena.
DR. COHN: Thank you. Jeff, do you have a question?
MR. BLAIR: Yes. Rick, you wound up mentioning that you felt, and I know from your background you have dealt with clinically specific coding systems and financial administrative coding systems. So, now you have made a major commitment with your new organization, and you are directed more towards documentation standards, and you wound up saying that you thought we should avoid coding systems to some degree.
The piece that I would like you to help us understand is if the payors require that the claims use codes and if those folks that are trying to measure quality and utilization and performance measurement, whether it is HCFA or other institutions and managed care organizations and I know they are looking for codes, how should we deal with that?
DR. PETERS: I will try to keep that very brief because it could be a long discussion, but I think my involvement with CalLink which is really being able to finally sit down with a payor has totally opened my eyes.
I am stunned at the flexibility that they represent. At the same time they take the lion's share of abuse in terms of being the people who demand these things.
They basically understand with no question that the data that they get out of ICD-9 CM and CPT is wholly inadequate for them to do the things that they need from a business perspective. Basically to adjudicate the claim they know it is inadequate but specifically to look at anything to do with actuarial risk from an indemnity perspective they think it is inadequate. It has nothing to do with severity of illness or a variety other things like that.
It really doesn't look at utilization. Particularly when you start to get into capitation and managing risk, i.e., they are not longer just the direct payor and indemnity insurer but they are managing risk because they are lower down the food chain such as in IPA that wants a capitated contract, the numbers mean nothing. They really have to look at their expense and their utilization which goes way beyond what code you assign to it.
They have another financial reality and that is that they have main frame systems that still have restrictive code field links and other things like that, but I am stunned as well at how much they are willing to map, and they have made specific statements, at least in California that say, "Listen, for adjudication of the claim we will push it through this whole main frame system that is going to dispense for X number a specific amount of dollars but to actually decide on whether or not we should pay that claim or not give us more data, and we will build systems and build translators that will allow us to accept that additional data and then eventually shove it through that payment mechanism which exists because it manages money well for us."
So, I think if we look at the payors and say that they are the opposite goal that is unfair because I think they want more data as well.
My perspective was that you just don't need a number. We don't need to take myocardial infarction associated with X, Y and Z and say that it is a 420.000 or I am adding too many zeroes there. There are only two allowed, and they cannot be zeroes. I mean that is the absurdity that we are into.
We really want to know when we can pay on that myocardial infarction and whatever clinical things there are that are complications associated with that. So, my perspective is only that I think we defend codes by using other people's reasons to have them but even those other people are saying, "We don't necessarily need these." Again, help us deal with the expense of our main frame systems and these restrictions that we have, but when you talk about clinical data, and I think that is an important thing that we can look at it here, we have done HIPAA. We have decided on X12 8-7. We decided on ICD-9 CM and CPT-4. Now, we are talking about clinical data. I think the field is wide open.
Now, we can drop those restrictive issues, drop a lot of main frame issues and start talking about what we could really do with data, how we could package it and manage it.
MR. BLAIR: Rick, I would like to ask you for something here because the approach that you are taking is away from that that clearly most folks have been taking. In order for us to better understand this on the Committee, I think it would be helpful if you could pull together a five- or-10-or-15-page paper saying what the alternative is.
In other words how do we do the other things we feel we need to do without the code systems that we are looking at? I think that would be helpful to us.
DR. PETERS: I would be glad to pull that together. It may not be all my work, but I will get the people who have that perspective and I would be glad to provide that as additional information, again, as open to the industry because I think one of the things is to share it. We have got to decrease the barriers to do these things.
DR. COHN: Kathleen?
MS. FYFFE: A quick question for Gary and then a question for Rick. Gary, do the other major automobile companies have information system structures and departments similar to Ford?
MR. RADTKE: They have similar areas, but they have a different focus. GM, for instance, has its focus mainly on the financial end of things. They have the quality groups that they have go out and talk to employers and to plans and to hospitals, but that is a separate group from their IT organization. Their organization has focused mainly on the financial aspects. Ours is mainly focused on the quality end of things. So, that is one of the reasons why we work collaboratively to bring both together so we don't have to reinvent the wheel with each other.
Daimler-Chrysler utilizes a lot of outsource resources in that area and as a result they have a much smaller group of IT engineers.
MS. FYFFE: Do you all have some sort of consortium where you share ideas?
MR. RADTKE: Yes. There is a three-auto initiative. Every month we get together and share ideas. We, also, are all involved in the hospital profiling and the plan profiling.
MS. FYFFE: Okay, thank you, and then Rick, you said something early on, very early on when you started your presentation, and I jotted it down, but I don't know if I got it correct, but I would like you to expand upon it for just a minute or two. You said something about clinical data only begins to have importance when it is related to administrative issues, something along that line. I don't mean to misquote you.
DR. PETERS: That is probably not a misquote, but it is a condensation of various things, even condensation gathering on the glass is an example. It is a regret rather than a positive statement. When you look at what drives people to buy systems and to implement electronic records I think that is the regret, that unless there is that economic incentive which now regrettably is tied to administrative data and is tied to payment and is really tied to all of this incredible interest in claims attachments rather than the concept of what is actually in the record itself, that is the thing that I was referring to, and that is that we cannot from an economic perspective in a business world separate the fact that the real impetus to implement clinical systems will be administrative.
MS. FYFFE: Okay.
DR. PETERS: Does that make sense?
MS. FYFFE: Yes, thank you.
DR. COHN: Dr. Lumpkin?
DR. LUMPKIN: In the prior panel Dr. Russler proposed that what should be done is that an RFP be let for a standard and let groups of standards development organizations get together and sort of respond. Do you think that the process is mature enough that that would be a feasible approach to moving this thing forward?
MR. RADTKE: I will take a shot at that. I don't think I know what would be in the RFP at this point in time, to be quite honest. I think you need to do a lot of pre- work before that to decide exactly what you are going to ask those standards bodies to do.
I heard a lot of discussion over the last 2 days about different alternatives. I think you need a group of people to get together and say, "Okay, if we want a formal RFP where we are going to ask for something, we all have to agree what we are asking for," and I am not sure I think we are at that point where we know what we are asking for.
DR. PETERS: I will take a stab on that basically to back up Tim's perspective, and I think there are specific areas, and I would take two. I would take drugs as one and two, the other thing that I advocated which is the XML SGML as a syntax standard.
I think what you could do if you just took drugs for example, you could without an RFP or anything else, we know what needs to be done. We know the vendors involved, and we know the decision makers, and we know the experts in this area. We could take Tim and some of his colleagues from First Data Bank, some of the people from HL7 and a vocabulary group and a few others. We could actually put them in a group in a room, sponsor them somewhere to sit down for 4 or 5 days and ask them to come up with an approach of how to address this.
They may come up with a process that says that we need to do this and this and this, and we need some funding to do so, and they could come up with then an RFP or they might, and I hate to say this, but they might come up with a solution in that time frame. Then you could turn around and offer that to the Secretary and say, "Listen, we had an industry group get together with government sponsorship that came up with a solution, and this is what we recommend."
DR. MC NAMARA: I would agree. I think that specific domains are more amenable to having an RFP created. I think the area of drugs clearly could occur. I think the area of laboratory nomenclature, and I think when you start talking about diagnostic coding, procedure coding it is a little more up in the air. Those things are a little less firmly understood.
DR. MORRIS: One caveat. Make sure you get the process or the infrastructure set up before you do that. The RFPs will be scattered, disjointed and you may end up having multiple difficult-to-maintain standards that don't actually have the right sort of infrastructure.
DR. COHN: Mike?
DR. FITZMAURICE: A quick question for Gary Radtke. Coming from a large corporation, do you have research going on that says that we have problems? People are out of work. We have got clinical data. We can find out what those problems are and we then try to address those solutions and so. Because we had uniform data or made the data uniform in one area we are able to address the problem and help people with their backs or change the assembly line. Do you have research going on like that that depends upon uniform accurate data? That is one question.
The second question for all and it can be brief is that we have existing standards developing organizations. Do they suit your needs? Can you work with them to get the uniform standards you want? Is the job too big? You find it too cumbersome, too slow moving; what can be done to get you encouraged to work more with standards developing organizations? I guess that is the way I want to phrase it, but first Gary.
MR. RADTKE: First of all, yes, in terms of the data that we have for our occupational medical facilities that is fairly well understood. I would say that it is probably slightly better than the average you would find in say a health care or some of the hospitals. I have, also, worked at some of the health care areas as well. So, I have seen that data. I think in general it is better than what you get there.
The problem we have is our data that we get from our internal sources is all non-confidential. We know exactly who got injured, where because we have to because we have to investigate those injuries and illnesses.
The data we get from the outside is all confidential information. It is scrambled information, and the difficulty is when dealing with 30 or 40 or 100 different HMOs that all code things slightly differently that all get it from different systems from providers who code things differently, it is very difficult in that environment to say exactly what is going on.
So, I would say that it is a mixed bag, on the occupational side probably okay, on the personal side it is so-so.
To finish my response to your second question which is are there standard-setting committees that are out there, they may be out there, but I am not sure of one that addresses both the occupational side as well as the personal side. We have been looking for a group set, are doing both, and so if you know of any groups in that area we would be happy to participate in that activity because we are looking to help generate standards in the occupational and non- occupational side.
DR. MC NAMARA: We are certainly interested in standards beyond drug nomenclature, but specifically in that area I think we have reached a little bit of an impasse. There is only one organization in the health care informatics community that is working on vocabulary standards for drugs and that is the technical committee of HL7, the vocabulary technical committee, and we have kind of run against the wall to the point that we have had multiple discussions now for almost a year and one-half about what the standard vocabulary may look like, and we are getting to the point where organizations are going to need to contribute or expose some of their codes and the meanings of some of their codes in order to actually get the rest of the work done.
Some of the other companies that are in the area of drug vocabulary and drug information have not elected or at least to me have not been able to express that commitment or ability to generate those particular codes. So, I think, unfortunately, we are at an impasse with HL7 and the vocabulary technical committee and that is why I think federal action in this area could potentially gel those efforts and create new efforts in that area.
DR. PETERS: Just to back Tim up, as a vendor we have access to his product and to all of his competitors. We look at all of them. There is very little difference. There is no reason that any of them is standing on a proprietary foot to defend that, but that is just one area.
Now, if you start looking at HL7 engines, you start looking at data interchange, there is a whole industry that exists in health care just for the exchange of information that doesn't need to exist there.
There is no industry like that in the general Internet market, none, zero. Everybody exchanges information in a defined format. Again, it is somewhat limited. It is HGML right now, but as they get into X amount of other things like that there will still be no one who is a data intermediary.
MR. BLAIR: When you say that there is a whole industry, are you referring to integration engines?
DR. PETERS: A permanent integration engine, and each of us has our own integration engine that we have built or that we cobble onto somebody else's. So, it is really layers of abstraction, and if you, you know, Dan Russler or Chuck Myer from SMS or HBFC can give you the number of things that they have done to try to do interoperability which is out of this world. It is regrettable because now people then have another industry to defend. Yes, we will put most of those companies out of business, not the HBOCs or SMSs but many of the data interchange vendors, but they are representative of many of our standards organizations and other things like that, and I cannot blame them for defending their turf, not at all. So, I would agree with Tim that outside activity or even government intervention to say, "Sorry, but this cannot continue," would be of some benefit.
In the vocabulary area in terms of drugs I think again that we could probably do this off line, but it would take a concerted effort and cooperative people like Tim and others to sit down at the table and say, "Yes, we will sort of take the kimono off, and we will compare our stuff and we will come up with a common standard."
I think there is still enough proprietary benefit in all of the other things available in drug management and all the available things in data that you don't need to make money off of doing those things.
DR. MORRIS: Briefly XML I think is going to be a huge piece moving forward. Separating the syntax and content is very important, and I think that we are at an interesting opportunity now because if you look at what is largely automated today it is administrative laboratory systems, a lot of back office systems. We are at a point in the next 5 years where we can begin to automate hopefully if Rick and others in our organization are successful to automate the clinicians at the point of care.
If we don't have these things lined up going in it is going to be a mess because a lot of what HL7 has done in the last couple of years is try to dig out how all these islands were created and then be able to actually communicate back and forth.
So, I would say, "Specifically look at messaging content, syntax and look at XML specifically for an open way of migrating.
DR. COHN: Peter Groen?
MR. GROEN: I represent the VA at the table here. I have a general statement. Maybe it is a question. Maybe it is not. As I have listened to all the various groups over the last 2 days we have had the physicians, the hospitals, the vendors, etc., commenting on these systems, and a lot of the focus is on the computerized patient record with the end user or the ultimate customer constantly referred to as the clinician, and yet I am starting to see more and more the design of systems and a recognition that we may be designing systems shortly for our patients recognizing them as the end user or the ultimate customer, not necessarily the clinician which changes a lot of things when you start looking at the design from that perspective.
You actually start to produce systems that will have a lot of unique applications and capabilities that don't currently exist. So, issues on interoperability and common terminology when looked at from the point of view of the hospital and the clinician and the health care industry are somewhat different when you are looking at it from the point of view of a patient and the systems that we are going to be designing in the future that are going to be patient oriented. The patient is the user. The patient owns the computerized patient record. That is his record. He wants certain things in that. How does all of that somehow relate to the standards area and the emphasis and again, I am sort of rambling on here because I am not quite sure what the question is other than I think we are spending a lot of time focusing on issues and maybe having introduced the patient as the ultimate end user of these systems and what these standards activities ought to be doing with regard to the patients.
DR. PETERS: The largest most competitive market segment at this point is in the patient information, patient data management, patient data entry industry. There are probably 30 venture funded or investment banking funded or privately funded companies that are pouring money into that segment. Many people see that even in drug management as a phenomenally big market, more so than even the traditionally MR market. Most of those tend to be web based. They are all working in HTML with web browsers as a primary format. They are certainly not getting at the content level which I think is important as Jon mentioned, but their syntax is HTML. That is it, and they are moving as rapidly as they can to build a market segment out of that. Many of them are now looking at the patient data entry side of it, i.e., to enter information from a clinical perspective that then could be sent to your doctor through e-mail or other things like that.
So, I think it is something we need to address. It is certainly something that if we do good work in the general EMR arena it would be adaptable to that as well. Virtually every EMR vendor represented here, too, gets requests all the time to build systems to go in the waiting room so that patients can directly enter their data as well. So, I think it is very important to look at that, plus, like Healthion(?) and all the others who are coming outside that aren't from the health care IT industry, these companies are not offshoots of HBOC and SMS and Oceania and all the others.
These companies are offshoots of Internet companies or privately funded concerns that are going to go solve a problem that they think is there with money and technology that they want to use.
DR. MORRIS: I think one of the challenges is what we have tried to do in our industry over the last 10 years is to replicate a paper-based process that we know is flawed but there is a business process that is articulated and oftentimes if you spend enough time with someone you can actually understand what it is they are trying to do. There isn't a paper-based business model for the consumer interaction, and indeed the definition of the consumer medical record. So, we are going to come into this in the next few years. I think we can be somewhat proactive in thinking about it, but the risk is that we will be there sort of ahead of our time.
The technology can give us a lot of opportunities to do these things, and we don't have a social way that this is actually being done right now. You know, just the fact that you can e-mail your provider and ask your provider questions is a new phenomenon that raises whole issues about what is the consumer medical record and we don't have something to fall back on to sort of look at the paper and say, "Here it is. I can photocopy it and give it to you." It is going to be an excellent topic for discussion moving forward, but it is going to be very, very different than what we have done in the past.
DR. MC NAMARA: I think your question is somewhat vague because it is a really complex issue and there is a couple of reasons for it. One is I still think we have to overcome the concerns about confidentiality and security of that information flowing across the network for the general populace before we can move too far in that area and two, consider the fact of a patient.
I will give my father as an example who has had three different doctors in the last year, has had two hospital stays in two different places. How would you combine that data together into one medical record and then who would be responsible for making changes based on the fact that one is inconsistent with the other? That is very difficult especially given the fact that many providers don't have an electronic medical record at this point in time.
So, I think companies like Ford Motor Company and GM and others are very interested in providing that information so that they can manage their health but the infrastructure is just not very mature at this point in time to do that.
MR. RADTKE: I think this is going to come on us a lot faster than we think. The banking industry, for example, I mean we know ATM machines showed up, and the next thing you know you have PC home banking and now I just basically do it all from my house, and I am getting very frustrated living out in West Virginia where my nearest, my physician is about an hour away, and I am extremely disturbed that is not on line that I can interact with him. I have got televideo on my home PC, and maybe I am not quite the normal West Virginian, but on the other hand the times are changing very fast and with web TV and all these other things happening we are going to have to start designing. If we wait too long in terms of thinking about the standards that are needed for the consumer based patient record we may find ourselves facing a lot of interesting problems down the road.
DR. COHN: Thank you.
Comments, questions? We trust you will be brief.
DR. FERRANS: I just wanted to add to the patient entered record that it would, you know, I would envision something of a plug and play architecture where it was merely another interface to the computerized patient record where a patient could under some of the privacy and confidentiality legislation proposed be able to validate the accuracy of the record and also, enter in some information. I think that mark-up language would provide some necessary context for this being the patient-entered information.
I would think at the current time that every American could buy a cable modem and an interface engine to be able to even come close to it because we are way far away.
DR. COHN: Thank you.
Other comments, questions?
Okay, I think that with that in the interests of time and letting everybody have some lunch why don't we adjourn until 2 o'clock, and I want to thank the panel very much for this very interesting session.
Thank you, all.
(Thereupon, at 1:05 p.m., a recess was taken until 2:22 p.m., the same day.)
DR. COHN: We are actually going to be covering a couple of things over the next hour and one-half, hour and 45 minutes, maybe longer if we really get going. One has to do with scheduling some upcoming sets of hearings and we have done some informal paring down of the dates. So, I will announce those in just a second.
Then what we want to do is just to brainstorm a little bit about the thoughts, learnings, things that we think we need to investigate further from this set of hearings which for all of you I think has been tremendously valuable, and we want to thank all those who are left from the panels this morning, but begin to distil, I think a little bit our thoughts about that.
Yes?
MR. MAYES: Could I make a quick suggestion?
DR. COHN: Sure.
MR. MAYES: Given the few people that are left could we just invite everybody to come join us at the table?
DR. COHN: Sure.
MR. MAYES: I am just giving them the opportunity because if they want to talk it makes it -- we have got lots of room here.
MR. BLAIR: I would appreciate as you come forward just tell me who you are because I cannot see you and I probably know half of you.
DR. COHN: Okay, but we will talk a little bit about learning, areas that we think we need to investigate further, reflect back upon our focus areas related to the work plan and how those may need to be changed or otherwise and I think at the end come looping back and come up with some ideas about what it is we want to do in the hearings that we are beginning to describe.
Now, is that okay with everybody in terms of the plans?
MR. BLAIR: Is Gracie here by any chance?
DR. COHN: No.
MR. BLAIR: Any staff person?
PARTICIPANT: Tracy is here.
MR. BLAIR: I only brought four copies of the work plan from September. Do you feel like we will need more copies of that?
DR. COHN: I actually think that pretty nearly everybody has copies of the questions that we asked the panel, and they actually list the four focus areas for the group. I think what we are really trying to do primarily is to validate those focus areas in our discussion, and Jeff correct me if I am wrong on that one?
MR. BLAIR: That is fine. I was thinking afterwards, after we identify the new areas of focus whether we would approach it in a similar way that we identified in the work plan where we would have stages of information gathering, stages of analysis and stages of building to get a consensus on the recommendations and laying that out. We may not even do that. That was what was in the work --
DR. COHN: Sure.
MR. BLAIR: That is for later.
DR. COHN: Okay, now, first of all I guess we should have people around the table who are staff and members of the Committee just verify that the dates we are proposing for hearings are March 29 and 30 and May 17 and 18, and the twenty-ninth would start at 10 a.m., which is a Monday morning. Those of us from the West Coast are sorry, though sorry mostly for ourselves.
Are there major conflicts that people have?
No? Okay, great.
So, once again May 22 and 23, March 22 and 23 and May 17 and 18.
PARTICIPANT: No, no, March 29 and 30.
DR. COHN: Whoops, I am sorry, March 29 and 30 and May 17 and 18?
PARTICIPANT: Correct.
DR. COHN: Thank you very much.
PARTICIPANT: Again, will the May 17, meeting which is on a Monday begin at 10 a.m.?
DR. COHN: Yes.
Thank you for catching me, since I was writing it down wrong myself.
Okay, now we could actually spend some time talking about what we have heard and want to reflect on over the last day and one-half, and I know, Jeff you had some thoughts that you wanted to express.
MR. BLAIR: Let me begin here? I was getting my little Dictaphone here in sequence. By the way do we have somebody who could capture some of these thoughts? I don't know whether Simon you have a notepad or I could do it on the Dictaphone?
DR. COHN: I think this is being taped, and there will be minutes.
MR. BLAIR: I meant in terms of what are the areas that we will focus on?
MS. FYFFE: I will take some notes.
MR. BLAIR: Let me see if I can focus in on all of the areas where I either felt I needed to understand existing focus areas better or people raised new focus areas that we needed to learn about.
I am going to start with that first panel and Peter Waegemann had suggested that we focus on preemptive federal legislation that could address the divergent state laws for retention of records, signatures on records, essentially the fact that most of the states have laws that make it difficult for vendors to have one system that is common across all the states in terms of an electronic health record. So, state legislations that are impediments - -
DR. LUMPKIN: I would just like to point out that Illinois was not listed on that.
MR. BLAIR: Except for Illinois and New Mexico and California. Anyway that was one item.
No. 2 was several folks suggested that we focus, and I don't know what this is going to mean yet, but that we focus more on documentation standards. So, that is No. 2, and I think we are going to have to find out what that is a little bit. I have feelings on that, but a third area that I heard quite a bit about, and this may be from many different perspectives, well, let me hit one perspective.
The third area that I heard was that the Federal Government do things to encourage or facilitate quality measurements or performance measurements. I am sticking them both in, quality measurements and performance measurements.
I think that is to enable and encourage the development of those as standards. Another area, I guess this is the fourth, MEDSTAT mentioned this. They had a great deal of difficulty with data quality and so data quality at the data capture stage is another area that we need to understand better.
DR. COHN: I think that was actually throughout all the stages essentially.
MR. BLAIR: Throughout all the stages?
DR. COHN: We tried to get them --
PARTICIPANT: It is actually the transition of data.
MR. BLAIR: Okay, all the way through, but I do think that there is a distinction between that and the performance measurement which might be NCQA level and --
DR. COHN: I think they are two completely different things.
MR. BLAIR: Yes, okay, good. That is what I thought, too.
Another area was under the broad umbrella of coordination of terminologies, and I think I will just leave it at that although we heard several different recommendations as to what that could be and maybe I ought to just put some sub-bullets underneath that. Among the sub- bullets are the coordination of generic clinical terminologies, but I think we, also, heard a separate need for drug knowledge base terminologies, and they may not be the same problem or the same solution.
The other things that I heard I cannot forget Ralph Korpman. Ralph talked about the fact that we should focus on standards that are infrastructure type standards and the kinds that he listed. Am I too fast?
MS. FYFFE: Just keep talking.
MR. BLAIR: Okay, infrastructure standards and the types of things that he referred to as infrastructure standards were accountability and trust and data integrity. He hit something else. Data security he mentioned in there?
PARTICIPANT: Authentication.
MR. BLAIR: Thank you. Authentication.
PARTICIPANT: Process integrity.
MR. BLAIR: Thank you. Process integrity as opposed to process modeling.
The other piece that I thought I heard was when we had Dorothy articulating the situation with trying to get the view of behavioral health and mental health in place, the message that I heard from her was that even though she wanted to be part of the process of standards development, and I heard this from others, too, I guess, they couldn't afford to send representatives to all of the relevant standards development organizations or to use Blackford's word soft standards organizations, the terminology developers to make sure they were part of the process.
I sort of feel like there must be something we need to do to make it easier or facilitate or enable or do something, but I think that is a problem that a lot of small domain vocabularies have that have to be brought into the system and they cannot afford to pay the fare to come to the meetings.
I think I will stop there.
MS. FYFFE: Okay, let me recap? I have documented seven items here. First was the issue of preemption of state law on the part of federal legislation.
Second was a focus on standards for documentation.
Third, the role of the Federal Government to facilitate quality measures and performance measures to enable standards for these.
Fourth, and this was brought up by MEDSTAT data quality, possibly at the data capture stage but really at all stages.
Fifth, coordination of terminologies, sort of an umbrella term of generic clinical terminologies. There was, also, mention of drug knowledge-based terminologies.
No. 6, and this was brought up by Ralph Korpman, infrastructure standards, under that sub-bullets, accountability, trust, data integrity, security, authentication, process integrity and then the seventh focus area, again, the initiation of this was from Dorothy Webman, issues related to behavioral and mental health, really an extension of that to the smaller organizations or parts of the health industry that have their own languages and needs but simply cannot afford to attend a variety of standards meetings.
Is that right? Is there anything --
MR. BLAIR: The only thing that I would ask, on the very first one which was the preemptive things, I think you might just put in parentheses after that that was especially targeted at medical record retention, signatures -- pardon?
MR. MAYES: Signatures are already covered under HIPAA. There is already a preemptive national electronic signature standard.
PARTICIPANT: Have the states signed onto that?
MS. FYFFE: For all types of --
PARTICIPANT: That is correct. If you look at the HIPAA regulations. Security regs under which that is part of are not confined to the transactions. They cover all.
MR. MEYERS: However, the electronic signature section specifically states that the signatures are only applicable to the transactions and where required or simply designated.
MS. FYFFE: Transaction sets or electronic medical record would cover all that?
DR. COHN: Yes, without trying to solve -- I think we have identified this is an area we want to look at further.
Jeff, do you mind if I amplify or add on to a couple?
MR. BLAIR: It is all yours.
DR. COHN: I just want to comment on a couple of things. I think one of the things we heard very clearly from a wide variety of people was that there are really sort of two separate focuses. We needed to look at sort of low- hanging fruit, the areas that we thought that we could make a big impact on in the next couple of years and then the longer-term issues some of which are infrastructural or whatever else they called it because I think everybody talks about infrastructure, and they all mean different things, but maybe longer term issues.
Now, I would, also, amplify that I heard from a variety of speakers starting with MEDSTAT, Dr. Schutt from MEDSTAT and others that in addition to other types of data that we are collecting there seem to be opportunities in the area of lab data and drug data, pharmacy data in terms of those being relatively high-value items that one would observe especially based on Tim McNamara's comments at that last session that these may be some sort of low-hanging fruit that maybe we should be investigating to see if there is an issue and if so what there is.
I would, also, observe that in some ways the issues around drug classifications if indeed there is a problem there in many ways probably mimic a lot of the other coordination and terminology issues, issues that we see in many other environments where there might be some value to understand that one a little better.
MR. BLAIR: Should that be listed as a separate item or are you just reiterating that?
MS. FYFFE: You were talking about two focuses.
DR. COHN: Around drugs, yes, and one would be the low-hanging fruit, but then, also, if there are conflicts there it would provide sort of almost a case study for the types of conflicts that we should expect to see related to incentives and trying to move forward towards uniform data standards that we would likely see in almost any of the other domains, and that is a hypothesis. That is not a statement of proof. It is more that that may be a useful area for us to look at.
MS. FYFFE: Does that go along with, I think more than one person mentioned the different clinical specialties. They mentioned cardiology and a couple of others.
DR. COHN: I don't know. I mean it is a slightly different slice of the pie, but it probably applies to that. I have a little bit harder time in my own mind thinking of how the subspecialties relate to the terminologies only because their interests cross so many of what I think of as domains like diagnosis, procedures and on and on, but I think it probably is very similar to the sort of questions, and there is one other piece and I will stop talking.
That had to do with a notation that I made around that there needed to be some better business, I mean we are kind of a business case here today in many ways or business issues, and I didn't want to lose the fact that we sort of came out of this feeling that there are business cases to be made or synthesized around capture of more clinical data and EMRs and there may need to be some independent work that this Committee or work group commissions in order to get that stuff pulled together and certainly Blackford Middleton's work or presentation this morning began to show some issues of return on investment and otherwise that would be useful.
Now, I will stop talking. Bob?
MR. MAYES: I would like to make a comment. I guess I sort of grouped these things into three major groups and areas rather than trying to enumerate them all because they were subtly different and I guess the three areas that I sort of saw were kind of sociopolitical types of issues that had to do with the preemptive legislation and privacy.
I know we purposefully did not stress that, but it did come up a couple of times, in other words, the legal issues, the kinds of things we normally think of Congress as doing.
The second area was this whole issue of the semantic management of data, and the reason I sort of broadened it beyond vocabulary is because I kind of got, you know, you hear what you want to hear, and I heard a lot of modeling or conceptual framework that you need, you know, it is the vocabulary, that actually it is the semantics that actually provides the conceptual framework.
There was some discussion as to whether or not you wanted to have a more general information model, and then the third area was these architectural or infrastructural standards, but I think for the Committee, I think an even more fundamental question that was indirectly put to you and to us was what is the appropriate level of the abstraction, if you will and I think we heard it both ways.
I mean the first couple of panels I think we heard very clearly from some groups that they didn't want us in their business, that they were willing to say that there were some very high level kinds of national things, perhaps some of these laws for very broad level but by God they didn't want us anywhere near the implementation specifics.
Now, of course, you, also heard the opposite. People were saying, "Well, they ought to come out with a drug vocabulary which is a pretty specific implementation," but I think that question of what is the appropriate role for the Committee in terms of recommendations is a legitimate one.
MR. BLAIR: It might turn out to be a different answer in each of many of these different areas.
MR. MAYES: It could well be, but I guess as we go forward with these recommendations and discussions that we sort of set the scene as saying that this particular recommendation is sort of a generic one if you will, a very high level one. This one we see the opportunity here for some very focused legislative or regulatory activity or what have you.
DR. COHN: Any more comments?
Good.
DR. LUMPKIN: I am still struggling, and I think the more I hear, I was using the analogy earlier, it is kind of like the projectionist inside the booth next to the el- track. Every time you get in focus another train comes by, and it just sort of jars it, and part of my difficulty is that it is a nomenclature issue I think as I am sitting here thinking about it.
What do we mean by the patient record? There are some fundamental terms that we haven't defined, and by not defining them we have heard of many different solutions that I am not sure are solving the same problem. So one of the things I think we need to do is to really start adopting or identifying what we can agree upon is a nomenclature.
So, in one dimension we have the stuff that the docs and nurses write about the patient's progress. It is observational. It is diagnostic. It is interpretive. It is the kind of thing that even if you had a camera you won't get in any sort of automated system. It requires the involvement of a human brain.
Then there is other stuff that is on there. Some of it may be drug treatment. Some of it may be laboratory results, all the sort of stuff which is already in a way that is easily automated. All of it has to be integrated, either electronically or as it is currently done it is integrated by human beings.
The other aspect to this, other dimension of this problem is that we have different sites. We have got hospitals which lend themselves to one type of solution, and we have got integrated systems, and then we have individual provider's offices who may participate in multiple type systems, and so, I guess that relates to an issue of scaleability which we haven't really discussed, to what extent can the system or does it need to be scaled and extensibility, and I am not sure that scaleability works in relationship to the system, but I think we need to discuss that and define it and see if in fact we are moving towards a standard is that one of the capacities that it needs to have, and so, when you put that in the context and then talk about the various other aspects of are we talking about a descriptive terminology or are we talking about a coding terminology or reference terminology, all these kinds of other things, you almost get a three-dimensional model that we seem to be trying to describe in one dimension.
So, what I would like to perhaps suggest to throw onto the plate is that we get better at describing what it is that we are trying to define.
MR. MAYES: John, just to jump in actually you said much better exactly what I was really trying to get at with level of abstraction, and I think we have got to nail down somehow the parameters. They can be arbitrary, either in total or individual by individual, but we have got to go forward with saying, "Here is what we mean when we say this, and here are the parameters around it within that context."
DR. FITZMAURICE: I guess some of the things that link in my mind, and I haven't gone through my notes in great detail are that first of all it is incentives. What incentives are out there operating, and I think we heard from the people who testified different kinds of incentives. Those in the private sector who have to sell systems and put them together for a living and have to work with other systems said "We need standards to work with other systems, not our own." They didn't say, "Stay out of it. We will make our own standards, and we will build our own walls." They said, "We want a level playing field. We want standards so that we can exchange data," and they didn't show us at least any fear that others will outcompete them. They think they can outcompete others, and I think that is the American way.
I heard, also, that the government should push quality measurement, maybe even regulate reporting of quality measures as a way to improve health care, maybe, also, as a way to sell some of their systems, that that is a driving force and that the more with it managed care organizations are and other large organizations are the more they don't question the cost. They question how are we going to pull the data together to manage what we are producing, better care for patients.
I heard them wanting although I don't know that they used this term a benevolent dictator, someone who knows the right thing to do and forces it on everybody, even if it is not optimal. They will take somebody's view of it in order to get something, to reduce the number of different formats they have to work with, maybe the number of different definitions that they have to work with.
I heard praise for the GCR project. I heard some condemnation of the GCR project, but all put into context which made it helpful.
MS. FYFFE: By the way, what is that project?
DR. FITZMAURICE: Why don't we wait for Richard to explain that when it gets down to him, if that is all right. He is part of it. I am not, but he is part of it.
MS. FYFFE: Oh, he is part of it?
DR. FITZMAURICE: I heard from virtually everybody, well, almost everybody increase the speed at which things happen, that it is just taking too long, and it is costing them sales. It is making the prices of their systems higher. So, increase the speed. Maybe that is part of the benevolent dictator.
I heard conflicting things in terms of the role of the Committee. I heard some say that the Committee should mandate HIPAA standards for clinical information. That is beyond the scope of the Committee, of course. I heard that the Committee should then advise the Secretary, advise Congress that this is needed. I heard them say that this should be voluntary, find a way to make the voluntary system work faster.
I heard one person say, "You know, look and see if consumers, if patients really want this, to see if doctors really want this," because there was a question in the testifier's mind of is this really needed or wanted, and yet, I reported scientific evidence that studies have shown that you give drug information to physicians and you get reduced adverse drug events. You get savings that you can count on to get a return, a benefit-to-cost ratio greater than two which is what it takes to sell systems.
So, there still seems to be confusion out there. There still seems to be not everybody knows the scientific evidence so far. Perhaps there is not enough of it. We do have it in drugs. We have some in laboratory. I would guess that we need more of it, and we need to put in here a convincing case that there is something out there. One person said to me, and this was during the break; so, this wasn't testimony, "You know, it is like seatbelts. If we know that these things save lives why don't we mandate that there ought to be an adverse drug event pop-up information system, a computerized support system for physicians. We did seatbelts. Why don't we make it mandatory?" I would like to have that on testimony, but it shows that some people feel very strongly and other people feel that it is not needed, that there needs to be some focusing of what is the evidence for this, and I will stop there.
DR. COHN: Peter, did you have a comment you wanted to make?
MR. GROEN: Two things jumped out at me. One of them was that this Committee could facilitate the process or the government could facilitate the process, but they didn't want us to set the standards, and then there was a whole series of suggestions on how we might facilitate from subsidizing who knows what.
The other thing that popped to mind was that the government could influence the process but again don't set the standards. Don't make it mandatory, but we could influence the process by identifying standards that the Federal Government agencies would adhere to as a result of let us say the GCPR project or the use of our own internal systems and that in turn might influence people to follow the path that the government takes. So, facilitate the process, influence the process but don't set standards.
DR. COHN: Jim?
MR. GARVIE: I think you have pretty much covered the list. I would just say that the GCPR project is a government computer-based patient record project that involves the VA, DOD and Indian Health Service and LSU Medical Center, and I think as Peter was saying, if people seem to be I guess benevolent dictators is as good a term as any but people seem to be wanting to take advantage of what relatively forceful direction might come from HCFA which tended to be the example in the form of tying reimbursement to compliance in terms of reporting and then the other thing as Peter was pointing was that perhaps through example the Federal Government could demonstrate where it is providing health care itself, some of the advantages, some of the benefits of some of these systems, but I think though that there were times when an issue was raised that probably needs to be resolved for the Committee and for us, so to speak, through our work and that is the extent to which there really are barriers to implementing these systems either in physician's offices or group practices or in hospitals as seemed to be implied.
In the Indian Health Service today people tend not to think of the Indian Health Service as a rich organization. That may surprise some of you in the room but we have health information systems installed in more than 250 hospitals, clinics, part-time clinics, community health center kinds of environments and so on, and they have been installed for more than a decade, in some places much longer than that, and they are clinically focused. They weren't installed to meet the business requirements of billing for services.
The VA has extensive health information systems installed in all of its hospitals, 172 or whatever the count today is medical centers, and I will have to admit in my own experience in implementing systems in a relatively poor financial environment, the Indian Health Service, I think you have a very difficult time generalizing. Sometimes people say, "Well, when the new generation of physicians who have been trained in this sort of thing, when they emerge and the older folks die off or whatever, then this stuff will all come to pass."
Some of the most forceful advocates for computer- based patient records are people who look like me. They are not young kids. If we were to go to the VA Washington Medical Center and talk with Ross Fletcher who I am assuming by appearance is well over 30, twice that, you know, he is a forceful advocate for the advantages of these kinds of systems. I don't think age, and I don't think a lot of these other things that have been alluded to are really issues, and there is another more fundamental issue, and that is that you don't implement health information systems in hospitals.
You start someplace within the hospital. You don't just go in and one day implement the whole thing. You tend to start someplace. You tend to start someplace where it is doable, whether it is pharmacy, whether it is lab, you know, someone who is interested, someplace where there is a recognized need and benefit and then through example most people most people don't want to be left out. They want to have the advantages that they see their counterparts with.
So, I think this is a long way of saying that I think we need to dispel or at least be real clear that there really are major barriers out there to implementing computer-based patient records, and I, for one, really question whether many of the barriers that have been identified are true barriers.
DR. COHN: I guess I am not sure that I caught what you were saying. Are you saying that there are barriers but they aren't what you just described; they are something else, and we should investigate what they are?
MR. GARVIE: I think we need to be, my impression, my honest impression is that there are barriers, certainly return on investment and so on, some of the things that I could identify later. People have to see that this is affordable, that these things are doable and so on, but I think that a lot of what continues to be touted as barriers, that people just don't see any value to this, I don't think is real. So, I am trying to say that I think the Committee needs to be clear. If there really are major barriers, we need to be really clear on what they are, and if there aren't barriers in a lot of these areas we could just move on.
DR. COHN: So, what we need to do is to identify what the real barriers are.
MR. GARVIE: And what the real status quo is, and I think as John was suggesting going out and seeing some of these sites, if there are community hospitals where this is just impossible, we need to know about that and find out what the real barriers are, but I am just saying that from my own experience in dealing with a relatively impoverished environment there aren't barriers.
MS. FYFFE: There was quite a bit of dense information discussed over the past couple of days, very dense, indeed, and it is going to take a while for some of us including yours truly to reread these papers and get through it, but there are two sorts of visual images that have come out of these 2 days of hearings that we might think about in terms of the lay person, and the lay person including those people up on Capitol Hill who are going to have to listen to all this, and one is the communication of medical information about a patient between hospitals.
Just visualize the fact that you have got Medicare patients in this country who live up in New Hampshire and then go down to Florida during the winter, and they need medical care in both places.
Can that information be communicated? No. Okay, that is very simple, but if we don't do anything else I would like to see hospitals and physicians' offices eventually communicate with one another. The other visual thing here which caught my eye and definitely catches the eyes and ears of people is the 3- 1/2 inch diskette. Can you take your medical history, your medical information from one part of the country to another, from one doctor to another and say, "Here it is. Put it into your computer, and you don't have to test me again and poke me and prod me again because this has already been done, and you have the results."
So, those are two very very simple things and if we can think about just how to approach just those two things, and there is quite a bit behind it, then I will be somewhat satisfied.
DR. COHN: I just want to comment that describing some of the dense stuff, some of the people who just came to mind as you were talking had talked about the focus being on quote, unquote, data interchange. We may want to consider that as one of the paradigms, and I think you were just beginning to describe two sort of cases of that.
MS. FYFFE: Yes. When you peel down the layers of what it would take to finally get to the point where you would have a diskette that you could take anyplace in the country, you are talking about standards. You are talking about terminology. You are talking about -- I mean that is the end point, perhaps.
MR. BLAIR: I think that folds in a lot with John Lumpkin's comment where he was saying, "I think we need to understand what people mean when they say that they have a computer-based patient record or a clinical data repository," because some of them are talking about the information exchange.
Some of them are talking about capturing an image. Some of them are talking about decision support. Some of them are talking about just the ability to do basically what you are talking about, and we sort of have to separate that out when we go into this next iteration during this next year as we go deeper into these things to understand them as to the people that are speaking to us, what is their frame of reference that they are coming from when they say that they have a clinical data repository or CPR.
DR. COHN: Richard, would you like to make a comment?
DR. FERRANS: Several things struck me. I think the thing that struck me the most today was the figure of $250 billion, and I mean that could sum up the business case, I mean what several witnesses talked about, and I think we went back to them and even Paul Tang talked about 15 to 25 percent being the range of savings that could be achieved in long term off a trillion-dollar health care budget.
You mentioned at the very beginning of this discussion that it would be something for the Committee to study. I think it is very important that we do that because that is sort of like an elephant hiding in plain sight, and people talk about trying to control the cost of health care and trying to improve quality. The argument goes that information technology in health care is the only thing that could do both simultaneously. Here is a starting point. We can begin to investigate and maybe flesh this out.
This is the sort of thing that on one page can really impress members of Congress if it can be captured at the appropriate level, but I am not willing to stake my reputation on those numbers. I certainly don't know the sources and what all the assumptions are and I think that to make a business case we need to lay out some assumptions, but I think that even if we are off by an order of magnitude and it ends up being $25 billion certainly that can provide a lot of impetus from the return on investment standpoint for investment into helping to subsidize the standards process or investments in informatics research which is something that a number of people yesterday called for that, you know, there is this 10-year project. There is this sort of Holy Grail at some point in the future, but there are a number of basic questions that need to be answered.
So, I think that that is important. I sort of teased out of the whole discussion that I felt on the balance of the arguments in favor and against and the data that was presented against and in some of the visceral arguments, in favor in the visceral arguments against that I felt like a threshold was met for in different areas some level of government involvement.
I didn't hear a clear case of good evidence where the government should not be involved. The big question in my mind, and this is my interpretation only, the big question in my mind is what is the appropriate level and I think that someone already answered that and said that it is really going to depend on what we are talking about.
I, again, heard the 10-year project and a lot of that has to do with infrastructure, and I think some of that has to be funded in research. I did think that there was low-hanging fruit, also, the pharmacy and lab data but, also, I think people and perhaps it is a little bit higher hanging, but people talked about HL7 and about clinical vocabularies and what I heard people say is that is it appropriate to try to move towards reference terminologies and reference models or reference standards for messaging and with the emphasis on the word "reference," and I think that that is something that needs to be fleshed out a bit.
I heard some proposals about service bureaus or specific ways where the vocabulary issues can be addressed. I think that they are intriguing, and they merits some follow-up.
I would love to ask for in writing from some of the people who made these suggestions to get some comments back to the Committee with some more details. In particular I would also, like to try to get some more comments from Bill Stead about some of the things that he said could be done in the short term, but we really didn't have a chance to talk to him at length about it, but it was obvious that he had some specific things in mind, and I would very much like to get some more granular input from him.
I, also, heard the issues of the quality measurements in clinical data, and I think that that just is part of the business case for the vocabulary and that is really those two in my mind are somewhat inseparable.
I heard the questions about the legal issues and state issues, the state's rights issues, and I think that that needs to be addressed. I heard issues regarding as far as government involvement, I did hear things about at some level that government should somehow organize SDF. That was one sort of bring everybody together. People weren't more specific than that, but that was a theme among people who were in favor of that, and the other part that government should pay, and should pay for some of this, should pay for the maintenance of some of these efforts, and I heard from several large vendors and small vendors that it is not in their business interests.
There are financial disincentives for them to participate in the standards process and that they saw this as a potential remedy.
The one thing I didn't hear was about cost. I asked a little bit. I am still a little bit concerned about that. I get concerned when people say, "Oh, you know we have a great return on investment model." That doesn't mean that people can afford it. I would use PAX(?) as a great example. They have beautiful data on how people can afford PAX because the return on investment is after 3 years or whatever. You know the joke is the chief of radiology hears from the CFO and says, "Well, the bad news is I don't have the $2 million up front to put into that," and I think there are significant issues with the total cost of ownership, you know when you have to upgrade networks, and when we talk about doing some of these things we really have to look at the total cost of ownership, not the cost of purchasing the software.
I think that that is a significant issue. I think that there are lots of organizations including my own that question sometimes the value of these systems, that the value is oversold. I think that some of that may have to do with the lack of standards, to be honest with you.
So, those are some of the things I heard, and the last point I wanted to make was that I heard Bill Stead say very clearly that whatever needs to be done there is a big caveat about interrupting work flow, and that was echoed by the AMA when they talked about the user interface issues for physicians, and I know that in the past that information standards have placed administrative and perhaps financial burdens on organizations.
I think there is always a risk in this arena of providing real work flow burdens on the individual physicians, and that is a real danger because user interface issues are tremendous.
DR. COHN: Thank you.
Mike, do you have a comment?
DR. FITZMAURICE: Yes, what jumped my hand up was the elephant analogy at the very beginning of Richard's presentation.
MR. BLAIR: Everybody loves the elephant.
DR. FITZMAURICE: Because it got me thinking that I don't think it is like an elephant. I don't think there are many elephants out there that go under this. Maybe a medium-sized element is the use of computerized decision score system for drug system. Maybe it is being able to retrieve lab tests so that you don't have to have duplicate tests. So, there is a savings.
Maybe it is in radiology so that you keep track of the films, and you don't have to retake it because somebody misplaced the film. You have got it there, and you can bill for it, and maybe administration, admitting, discharge and transfer, get it in the system once, and if it follows the patient around, the patient doesn't have to keep remembering his name and address.
Those may be the medium-size elements, but when I look back to computerization, for me and where I work I would write out a memo to somebody important. Everybody is more important than I am so that wasn't hard. I would give it to the secretary and instead of a typewriter she has a computer on her desk. So, she types it in and prints it out and gives it back to me, and gee, that is nice. Was there an increase in productivity? Yes, when I corrected it and gave it back to her it didn't take her so long to correct. She didn't have to type the whole page. That was okay, but that wasn't the big thing.
The big thing was when I started composing in my mind and I went through my fingers. It saved the secretaries. The took one of the secretaries away, and if you look around there are many more of us who don't have secretaries or there is one secretary supporting many more of us because the productivity filtered through the system.
I think it is not elephants but ants. I think each little saving, you pull up the dirt and you scrape and there is an ant, and you have to grab hold of it. You have to do something with the savings or the ant will find something else to do, and so, I think it is the same way in hospitals and medical clinics that you take a job away from somebody, and they will find something else to do. There is always something. It will be lower priority.
Someone has to look and say, "What are the higher priorities?" Maybe saving the salary is the high priority, and then you start accumulating those salary savings, but when you go out to measure it, it is very hard to evaluate because the ants find something else to do. The boss ant has to make sure they are doing the high priority things, and then when things get below a threshold priority it shouldn't get done, and you put that in the bottom line and return it to your stockholders or return it to your boss.
DR. COHN: Mike, you are making me wonder whether we should put you in charge of the business case or not.
DR. MAYES: I just wanted to make a comment on that. Having lived through having to do an impact analysis and deal with the office of actuary on the HIPAA transaction regs, let me tell you those $250 billion are in the air, because they "ain't" anywhere else.
We had a struggle, a struggle showing that we could show a positive return in a 5-year time frame on HIPAA transactions. There is no cost model out there. There are no data, and people who have data won't give it to you. So, I think that that would be an extremely important function to do, but it is not going to be send somebody over to the library and pull up a bunch of articles, and it is going to be there. We really had a hard time even showing what would have seemed to be a very easy thing to cost justify in terms of our system, just in terms of the financial transactions in HIPAA.
We had to go back and forth several times to the office of the actuary and get them to change their model before they would agree that it actually had some positive net result and that the bottom line was a positive one.
MR. BLAIR: And especially in the clinical area where there hasn't been up until this past year very much to be able to measure and show cost improvements, but during this last 18 months, and I didn't know that Blackford was going to wind up presenting some of the business case to us, but there is a lot of other work that he has done in assembling different studies. It was published in the January issue of Advance magazine, and anyway I think that we are beginning to get early results of cost benefits for clinical improvements from electronic health records.
Now, please keep in mind that it isn't just that they stick in a system, and it has standards. They have to go in on a consulting basis and do the clinical process re- engineering as consultants to yield a lot of the major benefits.
So, it is a change in the infrastructure and the processes that yield the big results and we are just at the tip of the iceberg, but it is worth learning more. If we do this legislation and regulation it will require a formal cost impact analysis.
DR. COHN: Kathleen?
MS. FYFFE: I am somewhat sensitive to the concerns that the vendors expressed or some of the people expressed about the business cycle, and you have got to have a relatively short, if not a very short return on investment, but that has really caused some problems. I mean somebody did a survey a few years ago of the number of different practice management systems that existed in the country. There were 2000, and I hope we can get to a point where there is a defined standard so the vendors know they have got to do their research and development toward that end and develop systems so that we can have interoperability and communication and everything else.
I mean one reason we are in the fix we are in right now is because we haven't had a standard for the private sector to really embrace and move toward over more than just a few calendar quarters of the business cycle.
DR. COHN: We have another half of the table here.
Val, I don't think you can comment too much since you just arrived, but if you would like to?
PARTICIPANT: I cannot comment on what happened yesterday, but although I just arrived, I listened most of this morning. Thank you, again. It was working brilliantly in my office. I discovered that the last people through had unplugged my speaker. We were all trying to figure out why it was that it wasn't working. I was watching you live on screen and then finally, my office is kind of a zoo. You have to crawl under the desk to get to the back of the computer where the thing is plugged in. So, it took us a while to figure that out.
So, I heard quite a bit of what was said today, and I didn't hear obviously some of the other opinions. Most of the people I was listening to this morning or the ones that I heard were primarily from the last panel, and they were saying some things which of course are said to me very frequently which is there has got to be some way for the government to have a carrot and a stick or we will never make up our minds. We will never divide up this terminology thing, and we will never go forward, and they were not saying it in those words, but that was kind of the gist of the message I was getting there from at least a couple of the people who were speaking
I must say that given my own experience over the last whatever it is, 13 or 14 years in this field, I tend to agree with them. I do think that somebody pointed out that the HIPAA model was kind of a nice model in the sense that everyone in standards organizations knew that if push finally came to shove the government could write a standard. It says that we can, I mean that the Secretary can, but it is clearly not in the best interests of the Department to do it, nor is it what the standards organizations really want.
So, I do believe that the legislation and that mandate have had a definite kick in the behind, you know, effect on some of the progress we have seen along some of the standards and the fact that some of the people have gotten down to cases, stopped disagreeing with each other and/or at least to the same extent and tried to move the thing forward, and we don't have the same stick in the area of some of these clinical standards.
On the other hand, you know, you could imagine a situation where the GCPR project would exercise a very similar thing just by saying, "Well, we want to do this by this time frame, and we would love to have you come and propose what the standard ought to be based on your coming to final agreement over there in HL7 or wherever you are, but if you don't, we will pick one or five or whatever."
DR. YASNOFF: I have two comments. One is I want to follow up on a couple of previous comments and point out that I think there was a remarkable consensus by most of the witnesses that the government should do something from people who you wouldn't a priori expect to be advocates of government action, and so, I think that is something that we should take note of, and within that consensus there appeared to be I would say, I didn't count exactly, but it seemed to me that the majority opinion was that the government should act to facilitate and speed up the existing standards processes to coordinate them.
I don't think that there is a unanimity of opinion as to how that should occur. I am a little bit concerned about the issue of evidence of cost savings from computer- based patient records.
I think that if you look back on the literature there are a number of good examples. I think Reed Gardner presented some from his institution, and they have a very long history of computerization. I think if you look at the work that has been done in the Harvard Community Health Plan, Group Health of Puget Sound they have excellent data showing that the use of automated records has been very, very helpful to them in terms of reducing their costs per patient per year, and I think part of the problem here is that without an automated record system the management of health care is like managing a factory.
If you are the factory manager, the factory has no windows, and you cannot see what is going on inside, and all you can do is count what goes in, count what goes out and count how much money you spend, and in that kind of a situation you do know whether you are losing money or whether you are making money, but you are pretty helpless in terms of improving the processes that go on inside that factory, and if your salary depends on the improving the efficiency of what goes on in that factory, then you are no longer satisfied with not being able to go inside and not knowing what all the detailed processes are, and that essentially is the electronic patient record.
However, given that you have an electronic patient record, that does not necessarily mean that you will use it wisely to improve the efficiency of what is going on in your medical factory, and so I think that may be the source of some of the problems and, also, it takes time to understand the processes that are going on and of course it takes lots of time to convince people that they need to change them.
The other observation I feel compelled to make is that none of the witnesses alluded to the potential public health benefits of electronic patient records at all, and I submit to you that major corporations such as Wal-Mart and McDonald's as of this morning, first thing this morning knew in electronic form about every transaction conducted by their businesses worldwide as of midnight last night.
They had that information in their hands this morning, and as you well know, with respect to reportable diseases as of this morning we did not have in our hands in the public health system the information about all the important reportable diseases all over the world or even in the United States.
DR. LUMPKIN: I would like to disagree. We did, but from last year.
(Laughter.)
DR. YASNOFF: Thank you. I think you are being very overly optimistic, perhaps as of midnight yesterday 3 years ago, and I think that having that information and being able to act on it sooner is clearly a public benefit that I believe the government has a responsibility to encourage.
PARTICIPANT: Even a national security benefit.
DR. YASNOFF: Yes, and actually when I am involved in the work at CDC to help local health departments be prepared for potential bioterrorism events, and one of the key issues in such an event is to be able to detect as early as possible that it is occurring, you cannot expect that a bioterrorist after doing whatever it is they decide to do will send you a telegram notifying you of exactly what they have done so that you can begin to respond.
You have to be able to detect the event, and failure to detect it early could cost hundreds or even thousands of lives.
DR. COHN: Any other comments from the floor?
Okay, obviously we have heard a lot of comments. The good question next is what do we do about all of these things. Now, in our discussions I would first of all comment we talked a lot about the business plan, and clearly if we are going to in any way approach anybody looking at the business plan issues, I think we were all reflecting that this is a complex area.
MS. FYFFE: Are you talking about the business case?
DR. COHN: The business case, I apologize. The business plan is even worse than the business case. The business case is a complex area, and I think we are, also, aware that just as we have a little bit of, I won't say confusion but lack of clarity about what it is we oftentimes talk about as the charge of this group, it is, also, reflective of a little bit of confusion and lack of clarity about what the business case really ought to be because in some ways we talk about comparable patient medical record information. We talk about EMR, well aware that there are probably 18 different definitions of EMRs out there, going all the way from clinical data repositories to clinical decision support systems, to physician documentation pieces as part of all this, and certainly anything we do in this area probably needs to reflect some of that complexity.
Now, the other piece, of course, is that yesterday and today we heard from a wide variety of people when we asked about what the intent of the legislation was that we should take our charge somewhat liberally and look just not at the narrow scope of patient medical record information but that we really should be looking at these other pieces that really make this area important and enabling. So, I guess I am hearing from everybody how it is a complex issue, but there are lots of pieces to it, and it is probably something that we ought to spin off at some point and have some independent work done.
Jeff?
MR. BLAIR: Let me make a stab and get reaction from the Committee and the extended Committee here in terms of process. It may work for us maybe if we were to try to synthesize these things and figure out how do we boil them down.
First of all before I describe the process let me indicate that the work plan that we had, I don't know how many of you have seen the work plan. I know the members of the work group have. We thought we had identified the areas of focus, and Simon some months ago indicated you know, we had better test that we are on target by having these hearings this last 2 days and Simon was absolutely right because if you take a look at the things we have heard and things we started listing there, there is a lot of things that we would have missed if we didn't have these hearings.
So, I think that they have been extremely valuable. My thought in terms of the process, we have at least in my mind three major things we need to do. We need to do information gathering. We need to analyze the information and make recommendations by August 2000, probably 3 or 6 months before that to have reviews with respect to our mission, quote, study uniform data standards for patient medical record information and exchange of that information.
The feeling that I got, and see if you feel like this is on target or not is we identified about eight or nine or 10 different areas here that were new to us that we really need to learn more about.
Some of them we could learn more about with hearings. Some of them we could learn more about in other ways, but I would probably put at the top of the list and as a matter of fact it wasn't even on the list of the seven items that I listed, probably at the very top is for us to come to an understanding of the spectrum of what people mean when they say that they have a clinical data repository or a computer-based patient record or clinical information system, and the reason I would put that at the top of our work plan is that if we understand the difference then we can begin to do the business case in a relevant manner.
What I am talking about here is there are folks that are telling us that they have a computer-based patient record system in an ambulatory environment and they will even refer to the clinical data as ICD9 and CPT codes. Not everybody will agree that that is clinically specific data. Some do, and some don't. We need to understand that that is what they mean when they say that they have a clinical data repository versus the other end of the spectrum where sometimes you saw dramatic positive results in business cases because they had clinical decision support, even if they didn't have the latest technology a la University of Utah, what Reed Gardner has done and Brent James has done with some of those studies where they have shown dramatic improvement in clinical processes with clinical decision support, but they have the information infrastructure to do that.
So, I think No. 1 on our list should be not a definition but understanding the spectrum of definitions that people use when they say that they have a clinical data repository.
MR. MAYES: Could I jump in for just a minute? This is Bob. Because actually nowhere in the legislation does it talk about that. The quote from the legislation is that we are to look at uniform data standards for patient medical record information, electronic, and the electronic exchange of such information.
So, I guess I am just concerned that if we get too far off on an academic exercise of defining what people are calling electronic medical records and repositories, I mean the charge in the law from Congress to this Committee is somewhat more limited.
MS. FYFFE: Let me follow up on that, Bob. I am glad you brought us back to somewhat near center again. You said two things there. You said, "Patient medical records," and you said, "Electronic exchange of such."
You know when we make our recommendations at the very minimum we have to define what the patient medical record is, have a formal definition for that.
MR. MAYES: Right, but that is not the electronic medical record.
MS. FYFFE: Correct, and the second definition we need is the electronic exchange of that, what do you mean by electronic exchange, and so if we define both of those terms and the definitions will be lengthy, maybe that will help us in pursuing our recommendations.
MR. MAYES: I think the difficulty we have is that that charge could be met by scanning or whatever into a word processor, transmitting the record in an encrypted Word Perfect file.
MS. FYFFE: That is part of the definition.
DR. LUMPKIN: That would not yield the enhancements that we would like to see in the health care system, and that gets to another piece of it which is if you have an automated system and all it does is help you the caregiver do your job better, and it doesn't hook up to your financial system; it doesn't give better quality data to the folks that are looking over your shoulder I would call that a win, but we have different perceptions and concepts of what success would be for that sort of implementation system, and we have to understand those, too, because I think that goes into the environment.
In a hospital there is a different test for that system than there would be in a one-or-two-physician- practice office.
MR. MAYES: But the reason I bring this up is it goes back to discussions we have had internally as to whether or not CPR work group was the best title for this group, and I think actually the GCPR project is struggling with the same problem or same issue, and that is as soon as you start saying, "Computer patient record or computer-based patient record," the GCPR project for instance is not really a project about developing a desk top implementation of a medical record system in the government. Its actual focus is on the exchange of health information across enterprises in a standardized machine computerized fashion, and I guess that when I look at the charge here I don't disagree with what you are saying at all, John. I think you are absolutely right and actually since this charge is not for standards due to legislation, that is where it differs from the financial action, and this is just a recommendation, and the recommendation could be that in fact we need to explore the concept of a CPR patient record, but I guess we should be clear about whether or not the -- I think the GCPR and Jim maybe or Richard, you could speak to that, but my impression was that once you started looking at it you focused and sort of used this phrase that we have out of the law. You focused on the exchange of that information.
MR. BLAIR: I really don't think that I am disagreeing with what you are saying at all. My only thought in putting that up front, getting the definitions in place so that we could wind up understanding whatever business cases are given out there is for us to get back, focus on the standards and focus on each of those areas to understand them better, but now we could begin to wind up relating to saying that if you could not only just have interoperability but you have interoperability with the clinical specificity to support decision support, then here are these studies that show significant cost reductions in decision support. Whereas if you just hit interoperability at a certain level here you will get business results at this level.
I think the studies are out there where we have got to be able to wind up keying what are the results of standards at different levels, and I think that information is there for us. So, that was what I was trying to do.
MR. MAYES: I would argue though that exchange is not equivalent to interoperability.
MR. BLAIR: That is right. That is what I am saying, "Different levels of interoperability." You could wind up having an HL7 message. That is a certain level. If you stick an interface engine in there, that is a higher level. If you add the clinical specificity that is a higher level, and I think we need to be able to quantify the benefits of moving to the higher level.
MS. HUMPHREYS: I was going to say, also, that it seems to me that if you did something along these lines, I don't say spend your life doing it, but if you do something along these lines, then, also, you may have a mechanism for making sure that when you are listening to people at subsequent hearings you know what they are talking about, that is you could provide this to them and say, "Okay, you have seen the matrix. Now, where are you on that? This thing that you are talking about that you do in your organization, where is it? Is it this one, this one, this one or some hybrid we haven't thought of that combines all three of them?"
MR. BLAIR: Exactly, it will enable us to understand the benefits. When Rick Peters tells us that he is using XML for documentation standards, that is going to yield a certain benefit.
Now, if he attaches that to codifying those data elements when they are captured in XML format and put them into a clinically specific repository and add additional applications then that is something else in addition, and we need to be able to discern the differences.
MR. MAYES: Are you talking about something in like the Zachman(?) framework?
MR. BLAIR: I really didn't want to go into that. I was thinking of something that might be a 3-or-4-month thing and then the other --
DR. YASNOFF: I don't think we want to repeat or update the IOM study on the computer based patient record which might, doing that might include getting into Zachman's stuff, but I don't think we want to get into that, but what I do think we need to do is I think that we have to understand clearly the context in which this electronic information is being exchanged or should be exchanged and I think we have quite a bit more information we need to understand in order to do that.
I will repeat a suggestion I made at lunch which is we did not hear from the consultant community except for there was one fellow from Ernst and Young and I think that a set of representatives from the consultant community, Anderson, etc., where they are dealing with many different vendors and many different organizations, I think could provide us with the insight we need in order to set up a framework in which we can understand where each of these vendors and each of these initiatives fits in, and I think we have to have that framework in order to be sure that we don't do any harm in what we do as well as assuring that we do something that can facilitate the process.
MS. FYFFE: I am going to read from the law itself because I think there are four things that it says, and I think that we might need to revisit a definition of scope or the mission of this work group.
I hate to keep coming back to that, but okay, it says here, "Shall study the issues." That is one thing, studying the issues, "related to the adoption of uniform data standards for patient medical record information and the electronic exchange of such information."
So, I see four parts to that sentence, studying the issues which means you have to identify the issues, secondly, adoption of uniform data standards. You have got to define what you mean by adoption of uniform data standards. Third, patient medical record information, you need to define that and then lastly, the electronic exchange of such information.
You know that is an incredible sentence, really. It is not even a sentence. It is sort of a phrase. It is huge, and when you peel back the onion you can get down to a whole lot of detail, but in the end in the spring of the year 2000 we had better be prepared to structure our recommendations according to those four parts.
DR. FERRANS: I wanted to make a comment on that. When we are talking about defining what CPRs meant, it is amazing that we now have sort of defined what the mission is, and it is very helpful to hear it again.
When I heard that part about relating to the adoption of uniform clinical standards --
MS. FYFFE: Uniform data standards.
DR. FERRANS: Uniform data standards, one thought that that triggered in my mind was in some ways some sort of major, if we talk about the business case but in some ways if one looks at individual business cases for some of the data standards that could be uniform, and examine them and sort of say, "What would be the benefit of adoption of this standard or these sets of standards; what would be the net effect of it, you know, what would be the cost to providers or what would be the benefit given?" and another thing is what are some of the, and since we are supposed to be studying the issues relating to this what are hypothetical migration strategies towards adoption of this --
MS. FYFFE: That is an issue.
DR. FERRANS: I mean that to me is one way that some of this could be framed. I am not suggesting that the work group ought to go with that, but one way of looking at it, saying, "Okay, if we identify certain standards," and say, if they were to become the uniform standard what would be the effect and what are three or four examples of ways in which it could be done, and I think we have heard from a variety of people suggestions as to how to get there.
DR. COHN: I do want to comment because I think we are all coming up with very good ideas that one of the things we actually sought, actually our first question at this hearing was some advice and guidance about exactly what that all meant, and we have had lots of hearings and lots of testimony over the last day and one-half, but I would imagine that an analysis of how people answered some of that first question might help us flesh out a little better all of our understandings about.
I think I started out making a comment earlier that a lot of people wanted us to take a relatively liberal interpretation of it, but the reality at the end of the day is we still need to make some interpretation of it. However we want to approach this, we had better come to a firm understanding and a way to hold it in our hands what we think of all this, and then I think we begin to talk about the various steps, and I think pieces that Jeff had commented on earlier are all really parts of all that. I mean we recognize that data quality unfortunately is an issue that runs through all of this. You cannot have uniform patient medical record information if you have poor data quality, no matter what standards you have out there, and it is an unfortunate part of it. I wish it weren't so. It makes more complex the whole issue, but it is another piece of this whole thing.
Kathleen, did you have another comment? We are beginning to move forward.
Oh, Jeff, I am sorry?
MS. FYFFE: I am sorry.
DR. FITZMAURICE: No, that is all right.
MS. FYFFE: You were talking about CPR clinical data repository, clinical decision support and then we sort of --
DR. FITZMAURICE: The essence was how do we proceed from a work plan standpoint during this next year, and I think I had indicated the definitions, and I think you broadened them a little bit because you went right back to the law and I think that is even more specific than just saying what the CPR is, but I do think we do need to have some map at least in our own minds of how the different standards that were discussed in the last day and one-half relate to application functions that yield rates of return because I think that is what Congress, the DHS and the rest of the community is interested in in terms of prioritizing and in terms of whether or not they support helping the private sector in those areas.
Now, the other parts of my proposal after we wind up being able to quantify this slightly, even if it is at a crude level in terms of showing the business case relative to the different standards, that the whole idea was to relate it to the standards, different ones is that when I listed those seven or eight items those were all areas where I felt we needed to understand those areas better, the documentation standards, the coordination of terminology, the quality of data, data capture area. All of those were things where points were raised where I think we have to have a better understanding before we could determine whether or not we need to be involved and if we do need to be involved how we could be involved in a constructive way as many people have said without doing damage, and so, I think that many of them will turn out to be hearings that we will schedule, but I think some number of those might be other paths. So, that was what I was going to suggest.
DR. COHN: Comments?
Mike?
DR. FITZMAURICE: I would guess one of the charges before us, like Kathleen I pulled out the law and I took a look at it. The purpose of the HIPAA section is to improve the efficiency and effectiveness of the health care system by encouraging the development of a health information system to establish standards and requirements for the electronic transmission of certain health information. Now, there they were probably talking about the HIPAA transactions. When they talk about studying the issues, when Congress charged the Committee with doing that they may not know what all the issues are. They may know what some of them are, and part of our charge is to lay out what are the issues we want to look at on our way to writing a report to the Secretary that says, "Here are our recommendations. Here are some legislative proposals, and here is how we got there. We started with these issues and worked our way up to what ought to be done."
So, I think one of the first things in our work plan needs to be identifying the issues that we are going to focus on, identifying some issues that we won't have the time or the resources to focus on.
DR. COHN: Okay. I agree. Other comments? I have a proposal to make for our next hearing, but I am not sure if it is really quite, I mean I am sure there is more that needs to go on in this particular discussion. I feel some of this needs to be synthesized and put together with a lot of the hearings.
I guess what I was going to comment as I am listening to all of this stuff I am still feeling that there is some need to be asking the country basically more questions about how we can be of help because I don't think we have the answers. I would comment that I think we have heard some low-hanging fruit, and I would put that out as an item that I can sort of grab and hold onto, but we have begun to hear from the vendor community and the business community and from the users at this point. I think that we need to hear in some way from the standards development organizations at our next hearing, from the government, VA, DOD and others, from the standards development organizations both what their needs are in this area and what they see as the opportunities in this area, recognizing that they deal a lot with this issue of uniform data standards for patient medical record information and especially the electronic exchange of this stuff. So I think that they can begin to tell us what they have, what they see as the issues, where potentially they need help, and it may be an appropriate time to begin to get some of their thoughts.
MS. HUMPHREYS: I agree. I think the combination of that really or just using that approach to doing at least some of the issues that were laid out by Jeff and others to follow up on his suggestion that we need more information about those issues, and it seems to me that you had a certain group of people here, and they said, "Look, you need to do this and this and this," and so, a logical question is to ask the next cast of characters do you agree with that, and what do you think that means. To you what would that mean, documentation of standards or quality measures or whatever, because it does seem to me that it seems highly probable that you will learn more by asking them potentially to address some questions that relate to what you heard from all the rest of this crowd just to find out the degree of agreement or disagreement with some of the views that we have heard already.
DR. COHN: Jeff, would you like to comment?
MR. BLAIR: Yes. Let me piggyback on Simon's thoughts and Betsy's thoughts because yes, I very much agree that we didn't have as many SDOs represented directly at least as we would have liked. We went for diversity which I think really was very valuable. If as you were to suggest that we include more SDOs I would only suggest that we do make it more issue focused and if we were to pick out two or three out of I don't know, 10 or 11 or however many areas we want to focus in on, we not only invite a panel or more of the SDOs that are concerned but the users, the users from the standpoint of those folks who use the data after it is in and, also the folks that have to put the information in. Sometimes that is the same people, but sometimes it is different, and there might be on a particular issue it may vary, but there may be secondary users of that information that, also, have a stake in that game, and I think that we then begin to get kind of a full spectrum, a 360-degree focus on that particular issue and maybe it is more than a two-dimensional thing.
MR. MAYES: We heard over and over again that the government should be involved either by example in incentive, and I can tell you the government is already doing a lot of things in these areas, maybe not in a coordinated fashion, but I think it would be useful for you to hear what is, you know, we have heard the GCPR; we have heard HCFA is an incentivizer all day long, but there are things going on in CDC and FDA and NLM and such.
So, I think that they are users in a way, but they have got a little bit more power than your normal users because they can sometimes mandate as part of their business process not just some general government mandate but as the business that we actually do, whether it is an insurance company, public health agency, a bibliographic agency, whatever. I was just extending what you were saying.
MR. BLAIR: As well as they may understand some of the issues a little better.
MS. FYFFE: Do we need to hear from any consumer groups on this? I mean everybody keeps talking about the diabetics of the world in the United States and we are always hearing about the doctor saying, "You need the eye test. You need this or you need that." Do we need to hear from any chronic disease groups?
MR. MAYES: I would suggest and this just happens because we are doing a lot of the work, end-stage renal disease is a group that of course we have a lot to do with in HCFA because we pay for much of it, and they are actively involved as an entire community in development of an information system that we are hoping to put in place nationally in 18 months, an actual information system linking all those groups together, and they are participating fully in this. It is a really interesting group because they are actually trying to do this from their community perspective.
PARTICIPANT: Are they linking the consumers together?
MR. MAYES: The providers, the payors, the regulators and consumers.
MS. HUMPHREYS: We know there are at least some people who are involved in the notion, I mean consumer groups who are involved in the notion that they should carry around the data with them and they should only allow it to even be displayed in health care institutions for the briefest possible moment. They should carry around their own erasing software so none is retained, and they carry it with them, and this may be an extreme group, but the fact is that in addition to the people who are very concerned about their data being in electronic form because their privacy might be violated there is the other side of the coin where people are talking about exactly what you are talking about, Kathleen which is wait a minute I have had all these horrible tests; here are the results, take them, use; don't do this to me again. Let me carry this around with me.
MS. FYFFE: Actually you all might recall an article that was either in the New York Times or the Wall Street Journal about a physician in West Virginia who put the medical records of his chronic disease patients on the Internet, got permission from all of them. These patients said, "Do it, because next time, Doctor, you go out of town, and we have to go to the ER, we want to be certain that ER doctor knows what our medical history is," and these patients gave permission for the records to go on the Internet with password access by the local emergency rooms.
In fact, I think it was over in West Virginia or Kentucky or someplace.
MS. HUMPHREYS: I am glad he asked permission because we funded some of that work.
MS. FYFFE: Yes. You know, I hear so much about the confidentiality of the medical information but by golly we have got some very sick people in this country who want the medical community to have the information to take care of them.
DR. COHN: I think we need to begin to wrap up. I know we said, "Four o'clock." So, we are getting close to the edge. I think we talked about a number of issues for the next hearing. I see Richard has a comment.
DR. FERRANS: I just wanted to agree with your idea of bringing in the SDOs and I think that we should group them sort of by subject matter. For example, vocabulary, and of course I think it is also useful to have them all together and I heard a theme of we cannot do it on our own; we cannot seem to come to an agreement. There is real value to have them all sitting right next to each other and to ask them questions about where the problems are and see if they can all agree on an appropriate role for government or not and I think that will be very valuable.
DR. CROHN: Okay, the other piece, I mean this needs some more design and I think this will be done off line as we try to design the issues.
Now, the other piece which I think is actually, I am going back to Jeff's earlier concept that we need as quickly as possible to come to some conceptual frame around this issue of what do we mean by the four things that Kathleen had commented on, and I don't know if we have enough time at the February meeting or how we need to approach it.
As I sit here thinking about it, some of the work has to do with getting and putting together all the answers to Question 1. I will be asking others that same Question 1 about what exactly is their idea of all of this, but we, also, I think need to somehow look at this issue of what is meant by patient medical record information and maybe how that differs from the electronic medical record or is it the same thing or what and come to some sort of conceptual frame on this and it may be that we may have some thoughts of either a speaker or a small panel put together to explore some of those issues or get it in terms of some sort of a written set.
MS. FYFFE: I think one thing that needs to happen is there needs to be some off-line discussion of defining our scope, and actually defining our scope might take a few paragraphs but we need to do it soon so that we can not have scope creep.
DR. COHN: Okay, and certainly I think overall some of the focus areas that we have are actually not bad even despite all the hearings of the last couple of days.
Mike?
DR. FITZMAURICE: I guess I would not want a long discussion of the computer-based patient record because as I look at the charge and the electronic exchange of such information to me a computer-based patient record is just the electronic repository. What you do with it is the system, and so you have got to have a computer-based patient record to send, and you have to have one to receive it. It may not be the same thing as the patient's record, but it is a storage of this patient medical record information.
So, I would rather focus on the issues and not focus on what is a computer-based patient record.
DR. COHN: I think the better question actually is rather than what is a computer-based patient record, it is what is patient medical record information and what parts of it are important, and I actually have some things that I think even I can bring to the table on that one because I mean once again what is important which is related to what is around the business case probably is something that we could even just exchange because there is information within any record system that always remains within a record.
MR. MAYES: But I would suggest once again that in all of this I think we needed to be guided by the sort of view of what is important which is I think that whole business case issue. What is important in all of this.
DR. LUMPKIN: I think we are grappling with a lot of issues that we are going to need to grapple with over the next period of time until we write this report, but the language there as I have been trying to deal with it is so broad as to allow us to look at a number of things because it requires us to report to the secretary about issues related to that and the key issue that we have been discussing in trying to deal with the scope of it is part of it is we have to define what the medical record is, but you know that somewhere in that medical record is that piece that I mentioned earlier which is the cognitive output of the human caregivers that are found in the medical notes, nursing notes, the assessments, and one of the issues that has to be addressed is that while these things may not be mandated by that language of the medical record because you can take them from one place and move them to another, that the viability and the functionality and the importance of doing this is related to how that enables some value to be added to the process, and that is where you move from just having a medical record to an automated record, a computerized patient record that gives value added to the process, and whether or not we decide we can do that and make a recommendation how it should be done or recommend that that is an issue that needs to be added to the place gives us a lot of leeway in making our final recommendations depending upon how successful we think we can be in clarifying the issues between now and when we have to make the report.
DR. COHN: We probably ought to reference that if we are indeed going to prepare a letter to the Secretary. We ought to in some way reference that piece potentially.
DR. LUMPKIN: Depending on where we are in making progress because you see that is where the whole issue of vocabulary gets in. If you put everything in an ASKI file and you transmit it from one clinician to another there is some value to that, being able to get all the medical records. It doesn't help me prescribe medications better when I am seeing that patient. It doesn't help me not forget to do things that I indicated I wanted to do. That is how my shop operates and how my information system in my shop helps me whether it be a hospital or an individual provider, and those are kind of different issues.
Our problem has been that we haven't been able to separate those two because doing one doesn't really get us as far as we want to go, and I don't think it is inappropriate that we want to go far, but we can accomplish our task and depending upon how far we get into the issue of vocabulary and making any decisions about whether there are things that can be done, we can still meet the task set to us by Congress in reporting on our progress on these other issues of just electronicizing or whatever the word is "ain't" enough.
DR. COHN: I think this is an area where probably some of the SDOs may be able to provide further illumination because if one thinks about it that is sort of how they have had to go through and do some prioritization based on the potential value to the industry which I think is what you need to reflect on.
Now, I would actually like to wrap up at this point. Are there any final comments? We actually have 2 hours, I guess at the February meeting as I understand for our work group.
MS. FYFFE: The February meeting is February 2, 3 and 4.
DR. COHN: Right.
MS. FYFFE: Which day do we have the 2 hours?
DR. FITZMAURICE: The third and fourth. The second is kind of open for hearings. Is there any time in that agenda?
MS. ARAKI: Actually we were planning to have subcommittee breakouts on the third and the Subcommittee on Privacy and Confidentiality is having a meeting and, also, the Subcommittee on Populations now is having a meeting. So, there are two subcommittees meeting the day before.
DR. LUMPKIN: And when is privacy and confidentiality meeting?
MS. ARAKI: Tuesday and --
DR. LUMPKIN: I am sorry, what time?
MS. ARAKI: I haven't seen an agenda yet, but I figure it is an all-day meeting, so, probably from ten to three or ten to four. We blocked out at least that much time.
You are having two round table discussions. So, it depends on how full that agenda is.
DR. COHN: John, what is your thought on that?
DR. LUMPKIN: I was just seeing if there was time to squeeze in a few more hours somewhere.
MS. FYFFE: We have the afternoon of Wednesday, February 3.
DR. LUMPKIN: I was thinking if there was time to squeeze in after privacy and confidentiality because there is a lot of overlap between that Committee and this Committee.
MS. ARAKI: You might want to work your agenda. That is a good suggestion because what you might want to do since that agenda is not yet gelled to consider having a meeting following right after that or even preceding it depending on how early privacy meetings.
DR. COHN: That is a good idea to further discuss, and hopefully at that point we will have a draft of the letter to the Secretary regarding progress.
MS. ARAKI: I think they are scheduling a conference call amongst the members of the Privacy and Confidentiality Subcommittee. So that might be something you might want to discuss.
MS. FYFFE: Would you mind sending an e-mail to Kathleen Frawley?
MS. ARAKI: In fact, I will send an e-mail to all the members of that subcommittee.
MS. FYFFE; So that we are certain to include that in the conference call. Thank you very much.
DR. COHN: Then our next full set of hearings is scheduled for March 29 and 30. Those are the right dates, and then we have one after that scheduled for May 17 and 18.
With that, thank you to everyone here on the staff for helping to bring together an outstanding group of witnesses and for the wonderful support for the last 2 days.
(Thereupon, at 4:08 p.m., the meeting was adjourned.)