This is the second day of our meeting. I would like to welcome everyone back. My name is John Lumpkin. I am the chair of the NCVHS and also director of the Illinois Department of Public Health.

We will go around for those on the Internet so we can identify ourselves, members of our committee and also the audience, and then we will proceed with the -- one of the things I am going to ask each of the subcommittees, just as a brief point that is not on the agenda, but so you can think about it now, if there are any action items that need to be taken from your reports, I just want to know about that so we can manage to quorum throughout the day.

Marjorie.

MS. GREENBERG: I am Marjorie Greenberg from the National Center for Health Statistics, Centers for Disease Control and Prevention. And I am the executive secretary to the committee.

DR. COHN: I am Simon Cohn. I am the national director for data warehousing for Kaiser Permanente and a member of the committee.

MS. FRAWLEY: Kathleen Frawley, vice president for legislative and public policy services of the American Health Information Management Association.

MS. TRUDEL: Karen Trudel, Health Care Financing Administration, liaison to the committee.

MS. COLTIN: Kathy Coltin, director of external affairs and measurement systems for Harvard Pilgrim Health Care, member of the committee.

MS. FYFFE: Kathleen Fyffe. I work at the Health Insurance Association of America and I am a member of the committee.

MR. GELLMAN: I am Bob Gellman. I am a privacy and information policy consultant in Washington and a member of the committee.

MR. O'KEEFE: I am Jerry O'Keefe. I am with the Massachusetts Division of Health Care Finance and Policy and I am speaking on behalf of NAHDO today, the National Association of Health Data Organizations.

DR. ELIXHAUSER: Hi. I am Anne Elixhauser. I am a social science analyst at the Agency for Health Care Policy and Research and I am speaking on behalf of the Public Health Data Standards Consortium.

MR. CHENG: Good morning. My name is Paul Gheng. I am the chief financial officer for the Union Health Center, the health care arm of the International Ladies Garment Workers Union and also the Textile Industry Workers Union.

DR. CARRILLO: Emilio Carrillo, medical director of the New York Hospital Community Health Plan.

MR. KNETTEL: Anthony Knettel, vice president of health affairs for the ERISA Industry Committee.

MS. WOO: Violet Woo, Office of Minority Health. I am on the Panel on Collection of Demographic Data on HIPAA Transactions.

DR. KRIEGER: I am Dr. Nancy Krieger. I am an associate professor at the Department of Health and Social Behavior at Harvard University School of Public Health.

MR. BLAIR: Jeff Blair, vice president of the Medical Records Institute and a member of the committee.

MS. WARD: Elizabeth Ward from the Washington State Department of Health, a member of the committee.

DR. FRIEDMAN: Dan Friedman, Massachusetts Department of Public Health, member of the committee.

DR. FITZMAURICE: Michael Fitzmaurice, Agency for Health Care Policy and Research, liaison to the committee.

DR. HARDING: Richard Harding. I am a child psychiatrist from South Carolina, member of the committee.

DR. STARFIELD: Barbara Starfield, Johns Hopkins University and member of the committee.

DR. FOWLES: Jinnet Fowles from the Institute for Research and Education, Health Systems Minnesota in Minneapolis.

DR. LUMPKIN: We will now have the audience introduce themselves. And, again, I apologize to those on the Internet since we don't have microphones, but we will ask themselves to introduce themselves.

[Further introductions off microphone.]

Thank you.

Oh, Clem. Welcome.

DR. MC DONALD: Clem McDonald, Regenstrief Institute at Indiana University and member of NCVHS.

DR. LUMPKIN: Now to the work groups and subcommittees, the population -- I think you will have

-- return the document on insular areas, Puerto Rico and Virgin Islands. Do you have any other action items?

MS. COLTIN: No, actually we are -- I will let Elizabeth speak to the report on the insular and territories. We are going to withdraw as an action item the Medicare report -- Medicaid --

DR. LUMPKIN: Actually, I am not asking for us to do that. I am just trying to manage the quorum. So, I just want to -- I don't think there were any action items, other than that document coming back.

MS. WARD: We have an additional recommendation that we drafted and we will read that out to you and ask you to --

DR. LUMPKIN: Standards and security, any action items?

DR. COHN: No action items.

DR. LUMPKIN: Okay.

Privacy and confidentiality?

MS. FRAWLEY: No action items.

DR. LUMPKIN: Right. And none from the NHII.

If there are people who will be leaving before the end of the day, please let me know because we may decide to -- we may need to take a vote on that document before breaking for lunch, to make sure --

Agenda Item: Panel on Collection of Demographic Data on HIPAA Transactions

Our first item an issue that has raised a fair amount of discussion. Currently, electronic mail and some discussion on the -- I can't remember what is the name of that list serve that Marjorie got me on.

MS. GREENBERG: Public Health Data Standards Consortium List Serve.

DR. LUMPKIN: It really relates to the issue of demographic data and HIPAA transactions, something that we thought was really a simple issue, but has turned in to be much more complex than we believed. We have a distinguished panel, who will give us their insight on this issue.

We will start off with Jerry.

MR. O'KEEFE: Thank you very much and good morning to the members of the committee.

My name, again, is Jerry O'Keefe and I am representing the Massachusetts Division of Health Care Finance and Policy. The Massachusetts Division of Health Care Finance and Policy is responsible for facilitating the use of health care information among policy makers, providers, researchers and others in Massachusetts, also for developing pricing methodologies for public purchasers of health care and also managing the state's Uncompensated Care Pool.

We are responsible in Massachusetts for collecting the statewide hospital discharge data set. So, today, I am representing -- speaking on behalf of NAHDO, of which our division is a member. NAHDO is the National Association of Health Data Organizations. It is a non-profit membership and educational organization. It is dedicated to improving health and health care through collection, public availability and the use of health care information.

NAHDO represents state, federal and private organizations. Our vision at NAHDO is for data sets that are comprehensive and that have indicators that add to the policy relevance and the market utility of our data and we, in general, look to the HIPAA standards as an opportunity to help facilitate that mission. However, NAHDO strongly believes that the collection of race, ethnicity information and administrative data sets provides information that providers, policy makers and researchers find very valuable and we hope that -- and recommend that race/ethnicity information be included within the X12 Core Standards.

About 40 states have legislative mandates to collect, analyze and disseminate health care data. These data systems provide useful information about the health of the state's population and about the performance of the health care system. Hospital discharge databases are a cost effective source of statewide data for policy making and they also provide a valuable source of information for private providers and health care organizations.

Some of our diverse users include providers and their consultants, researchers, payers, public health departments and others. We believe it is important that race/ethnicity be included within the Core Transaction Standards for the following reasons. At least 16 states currently have successfully argued for the inclusion of race ethnicity data through debates at the local level, convincing data suppliers, as well as policy makers in the states of the value of this data element and that the value outweighs the burden of collecting it.

These states stand to lose race and ethnicity data collection if they are lost in the standards making process. Data collection isn't free and certainly we are very sensitive to the cost of the data supplier, the burden on them to send us the data. This is especially true of today's environment.

Given that, it is very unlikely that states could continue to collect this data if it is not part of the national standard as an option for collecting race/ethnicity data. One proposed alternative is the use of enrollment files from plans. While this may be theoretically possible, proprietary and privacy issues would likely prevent this and many state data organizations wouldn't have the statutory authority to do this. Also, state data organizations not only collect data from all payers, but we also collect data about the uninsured. This route would also not give us data about the uninsured.

It is important to put the burden of duplicate collection in context. Most hospital discharge data systems are inpatient data and increasingly, hospitalizations are for higher acuity patients and thus are very significant events for which to track on a population-wide basis. We do not believe that by making this an option for collection that that would necessarily mean that, you know, every institutional or professional provider would have to collect this for every ambulatory encounter.

Rather, I think we would be sensitive about, you know, when this data is collected. Providers can actually benefit from this data as they and their consultants are key users of hospital discharge data. The ability to understand how race and ethnicity impact treatment patterns and access to care has influenced program development.

Providers have used an understanding of their community to target needed services and this also allows them to demonstrate community benefits. One concern that has been raised is whether this information is sound as this information is sometimes not self-reported and may be collected differently from hospital to hospital.

An analysis recently conducted by the Massachusetts Department of Public Health compared the hospital discharge data to mother's self reported race/ethnicity data on the birth certificate. And we view the birth certificate data as the gold standard. The results showed over a 90 percent agreement between these data for whites, blacks and Hispanics.

In addition, an academic researcher that has used our data informed us that he tracked the consistency of race code reporting across hospitals. He found that reporting was quite consistent and he was, therefore, willing to rely upon it. We believe that a uniform national standard for the collection of race and ethnicity standard will over time only improve data quality and the consistent use of the data throughout the country.

The validity of the data will increase its use and a national standard will allow for comparisons across multiple states, comparisons that can't now be made. Race/ethnicity data are used for a variety of important purposes, including health care program planning and evaluation of access to care and quality of care.

As states and communities and providers and researchers strive to learn more about variation in health cost, quality and access, there is a real need to drill down to explain these variances. Race/ethnicity data is an important element to allow one to drill down. Just as one example, the division's preventable hospitalizations project provides data about ambulatory care sensitive conditions for small area analysis and we provide it by race to community organizations. One alliance of city hospitals have used this data to understand their community and inform programs addressing the needs of populations at higher health risk, including in this instance the needs of African-American men.

The department of public health also uses hospital discharge data to provide descriptive information by race to advocacy groups. An example of that would be one community health network area, which encompasses an urban population was using asthma data by race. Another city health department has looked at race-specific hospitalization rates for falls among the elderly.

These organizations look at race to better understand the population that they are planning interventions for. In addition, race code data has been used by researchers to document racial disparities in health status, utilization and outcomes and researchers have used our data to look at disparities in incidence of disease by race and in performing discretionary surgeries. They have also looked at mortality for certain conditions by race and insurance status.

Health programs and interventions are increasingly structured to ensure that they are culturally appropriate. It will be useful to track outcomes over time to see if racial disparities have changed.

For policy making purposes, it is important to assess these differences and also to target programs effectively.

In closing, HIPAA standards promise many benefits in terms of efficiency and due to the electronic transaction of health business. But as we move in this direction, it is really important not to lose the ability to collect data elements that are important for policy, planning and the accountability of health care systems.

Therefore, NAHDO strongly urges that the race/ethnicity data be included in the national standards for claims transactions.

Thank you very much for this opportunity to speak this morning.

DR. STARFIELD: Just a clarification.

Your remarks address only race/ethnicity and not sociodemography data in general, like, for example, SES?

MR. O'KEEFE: Yes. That is what I came here to discuss this morning, yes. Yes, that is correct.

DR. LUMPKIN: I think it would be useful for us, for those of you who are advocating that it be on the transactions, if you would identify the transaction that you would like to see it on, I think that may be also helpful and if it is collected and how it is collected. For instance, is it part of the 1500 or the modified 1500 you use in Massachusetts or where do you get race and ethnicity today?

MR. O'KEEFE: Most of our data is UB-92 based information.

DR. LUMPKIN: Okay. And which transaction would you --

MR. O'KEEFE: It would be the 837.

DR. LUMPKIN: Thank you.

We will take questions from the committee as a whole on content issues to all of you once you are all done.

DR. ELIXHAUSER: Thank you very much for the opportunity to be here today.

I think all the committee members have slides that kind of follow the presentation that I am going to be doing. The document that I handed out also follows that pretty closely. If you would like to follow that along, that would be fine.

I am AHCPR's representative on the Public Health Data Standards Consortium. The consortium is comprised of representatives from private organizations and from state and federal agencies and we are working to be sure that public health and health services research interests are heard during implementation of HIPAA.

The consortium is basically an outgrowth of a meeting that was sponsored last fall by NCHS, CDC and AHCPR. Since that meeting, we have set up a list serve

-- NCHS has actually set up a list serve to discuss potential state changes around the HIPAA standard. One of the first elements that we have started discussing is race/ethnicity.

What we are currently doing on that list serve is exploring the desirability of integrating grace/ethnicity as an optional data element into the institutional claim for the HIPAA standard. So, this is the claim for the hospital, analogous to the UB-92 now. What we are doing right now is we are discussing the advantages and the disadvantages.

Today's discussion we really perceive as part of our larger discussion. Now, race/ethnicity is a particularly critical data element because administrative data, like hospital data, institutional claims data, has provided an excellent way for us to measure disparities and it can continue to provide an excellent way to measure our success in eliminating disparities over time.

Table 1 that is in the written testimony that I provided you today provides you with a sample of studies. It is like the fourth or fifth page, but Table 1 provides you with a sample of studies that uses hospital discharge abstract data to look at race/ethnicity disparities. All of these studies use data from the state discharge data systems, such as the one that Jerry O'Keefe just described.

What we see when we look at those studies is a very consistent message. We see significant racial differences in the use of procedures and other services even when controlling for comorbidities, severity of illness and insurance status and repeatedly we see that non-white populations have lower rates of invasive discretionary services.

Now, only now are we beginning to see studies that explain the reasons for those differences. Schulman's study -- it is not described there but it is in the other document that I provided you -- Schulman's study in The New England Journal of Medicine in January of this year showed that physicians recommended fewer cardiac services for blacks than for whites. This is a finding that is completely consistent with the information that we have gotten from discharge abstract data studies for the past decade.

So, without administrative data, we would have to rely primarily on primary data collection, such as surveys or prospective clinical studies or medical record reviews or we would have to rely on data that is based on a subset of patient, for which data are currently in existence, such as veterans or the elderly with Medicare data.

State data collection efforts, such as those maintained for hospital discharges are the only source of comprehensive data on all patients regardless of age, regardless of insurance status and it includes the uninsured. Now, most states compile hospital discharge abstract data based on the UB92. AHCPR maintains the data collection effort called the Health Care Cost and Utilization Project.

For the current data year that we are working on, we are obtaining data from 22 states. Of those 22 states, 16 collect information on race and ethnicity, even though race ethnicity is not part of the UB-92 standard. These states have decided that race/ethnicity is critical enough that its collection is warranted. So, they are asking the hospitals to provide that information in addition to the claims that they are submitting to the state data organizations.

But one problem that we face is non-uniform coding of race ethnicity across the states. For example, some states provide Hispanic information as a separate data element. Some combine it with the race categories. We also see problems in compliance in providing that information by the hospitals.

Now, HIPAA applies to electronic transmission of health transactions. HIPAA would not apply to state data collection efforts, such as the one described by Jerry. But based on the preliminary information that we have from the states that participate in the HCUP data project, the states plan to be HIPAA compliant. They plan to do this data conversion in order to take advantage of the tremendous opportunity that standardization provides.

They figure that standardization is going to reduce the burden on hospitals and is going to provide better data because it will be uniformly coded and it will have the same data elements. But under current HIPAA standards, as we know, there is not going to be any information on race/ethnicity. And for those states that don't want to lose this information, they would have to request race/ethnicity as an additional data element outside the standard.

Now, two factors would facilitate adding race/ethnicity to the institutional claims standard. First, even though race/ethnicity is not part of the implementation guide, it is a claims standard. What that means is that it is a recognized data element by the standard setting organizations. It just isn't part of the implementation for HIPAA.

Second, CDC staff are working with the standards setting groups right now to align the race/ethnicity code sets that are part of the claims standard with OMB Directive 15, the federal standard for collecting race/ethnicity. These two factors mean that adding race/ethnicity to the implementation guide would be considerably simpler than adding a completely new data element.

So, we have often received the question why not just collect race ethnicity as part of the enrollment file instead of collecting it at each hospitalization. We figure that there are four main factors that weigh against that. First, state data organizations do not have access to enrollment files. Many state laws require hospitals to submit hospital discharge records to the state, but enrollment files are not part of the submission.

Second, access to enrollment files would require relationships with entities other than hospitals and states don't have jurisdiction over some of those entities, such as employer-sponsored ERISA plans. We would never be able to get enrollment data from those plans.

Third, the use of enrollment files would require patient identifiers to link the data and many states don't collect identifiers on patients in order to protect patient confidentiality.

Finally, as Jerry mentioned earlier, there is no enrollment files for the uninsured or self-pay patients. So, race/ethnicity would be available for this potentially most vulnerable population.

So, we ask ourselves why do we want to add race ethnicity to the claims standard. The major requirement for adding a data element to the HIPAA standard is that it meet the business needs of the transaction. So, the question is is race/ethnicity information an important part of the transaction?

We contend that a health care transaction is more than a business arrangement between a provider and a payer. We see critical differences between health care transactions and other business transactions. The objective in health care is not just the exchange of an external commodity, but it is the improvement of a person's health and the nation as a whole has an interest in maintaining and improving the population's health.

So, we ask is race/ethnicity related to the delivery of health care? The preponderance of evidence that you see in Table 1 suggests that it is. Racial and ethnic disparities int he use of services is not just a public health issue, but it also suggests differences in quality of care. We contend that business as it relates to the HIPAA standard is not just a matter of submitting claims; rather, quality of care is a big part of the business of health care, the business of providing health services.

Plans, providers and institutions are being held accountable for the quality of care that they provide. Part of eliminating disparities in health status is reducing disparities in the use of health services. So, only by maintaining our data infrastructure will be able to track and evaluate our success in reducing or eliminating disparities.

For example, across the country, not for profit hospitals are being required by states to demonstrate that they provide significant community benefit to justify their not-for-profit status. An important part of their reporting requirements will be to demonstrate the benefits that they provide to vulnerable populations, such as racial and ethnic minorities. So, the business of running a not-for-profit hospital will hinge on showing what they do for minorities. This requires information on race and ethnicity.

So, what if race/ethnicity is not added to the standard for institutional claims. We know that the purpose of administrative simplification is to simplify the business transaction by providing a uniform standard for data collection and transmission.

Now, if states mandate additional data collection above the standard, that is, if they ask hospitals to continue to report race/ethnicity even though it is not part of the standard claim, it is very likely that we are going to see non-uniform formats in that data collection. If the collection of race/ethnicity remains outside of the normal stream of data collection, it is likely to impose a greater burden on providers.

In the future after the implementation of HIPAA, it is likely that there is going to be greater resistance providing elements outside of the standard.

So, in conclusion, what we would like to emphasize is that health care data serve broader needs, both business and otherwise than simply a business transaction by integrating the collection of critical data elements, like race/ethnicity into the standard. The burden will be minimized while simultaneously meeting the information needs of an industry that affects the well-being of us all.

Thank you very much.

DR. LUMPKIN: Thank you.

DR. FOWLES: God morning. Thank you very much for your invitation to give testimony to the Panel on Collection of Demographic Data on HIPAA Transactions.

I am a health services researcher and manage a research organization that is affiliated with a vertically integrated delivery system called Health System Minnesota. Today I am here not as a researcher, but to give you some insight into the providers' experience in collecting race and ethnicity information. To orient you to this perspective, I will first briefly describe Health System Minnesota.

Health System Minnesota is an integrated care system with approximately 250,000 active patients. These patients represent multiple payer sources, including health maintenance organizations, point of service contracts, self-insured employers and indemnity coverage. Health System Minnesota is based in the suburbs primarily of Minneapolis and employees approximately 6,000 people, including more than 450 physicians.

There are 16 clinic locations and one community teaching hospital. We have had space in our registration field -- for our patient registration field for race information for the past ten years, but until 1999, there was no attempt to collect that information systematically. Before 1999, I think it is fair to say that the space was virtually ignored.

Our research needs drove the race and ethnicity data collection agenda at Health System Minnesota. We cannot apply for federal funding, continue to accept it without these data and federal requirements, as you know, have become increasingly better defined.

There were several critical steps that we followed to collect race and ethnicity data at Health System Minnesota. First was determining what the motivation would be for the organization. For us, the dollars brought in from federal research and the subsequent contributions that we were able to make toward improving patient care through research were a compelling argument first for my own institute for research and education, but more importantly for the leadership of our parent organization.

We needed to get the buy in of senior management and by informing them of the dollars at stake and perhaps even more important, appealing to their role, of their reputation in the community, we received significant support from them. That led to the development of an explicit policy defined by senior management and then the hard work began.

First was the identification of a process owner and a care team. In our organization, identification of a process owner is someone who can take accountability as key for making any successful systematic change. In this case, the process owner is Michael McGregor, a former registrar and ward clerk. Mr. McGregor is now a business analyst with our information management team, but he combined the front line experience and the personal attributes for a dedicated and successful process owner.

Our next step was to mount an internal marketing campaign, including internal publicity for the process. All of our in-house publications were used to announce the policy and the rationale for the policy and I will make available copies of the examples of that internal publicity.

The next step was training of the front line staff. This began in a very open conversational mode with the persons representing our receptionist and registrars. We have an organizations called our Front Line Council, which includes 50 of these representatives, representing about 250 receptionists throughout the system.

He announced the policy and collected a lot of feedback from the receptionists and registrars. The council has continued to provide a forum for airing any questions and concerns that come up with regard to collecting these data.

Interestingly, the organization had not factored in any negative response from physicians. I think a more complete training would have included communications aimed directly at not just the leadership physicians but the rank and file physicians in the organization..

The implementation required development and testing of scripts. Working with his team, Mike and his colleagues drafted a number of written materials. They included scripts for asking the questions, cards to act as prompts for patients, both in large type and small type, information sheets for staff and information sheets for patients. Again, there are examples that I will included with my testimony.

These materials were not only reviewed by members of the team, but also by patients, who attend our clinics. The categories, as you will see on these materials when you get them, did not map precisely to the current OMB definitions for race and ethnicity, much to my chagrin.

Here our good intentions ran directly into Y2K. Everybody agreed that it was the right thing to do, but allowing people to check more than one was not a programming option that would be available until we had solved our Y2K problems. So, they have a multiple category, but they can't check multiple categories.

Importantly, in our system, we could provide feedback to the receptionists about their performance. We could provide this receptionist specific information fairly quickly and, thus, identify people who were more or less on the right wagon. This feedback information allowed us to elicit further concerns that people had and work with them to identify more elaborate rationales for what was going on.

All of this began in the first quarter of 1999 and as of June 15th, 1999, race/ethnicity data are available for over 210,000 out of our 250,000 active patients. Though we have had anecdotal reports of some backsliding, which comes in the form you may be aware of, using visual inspection rather than self-identification, we have been able to pick up these issues and continue to address them through the active Front Line Council that we have.

The minimal patient questions that have arisen have been addressed by the materials and the individuals that are available to answer them. So, what have we learned in this process. I have counted eight steps. First, planning is essential. Second system changes involve a lot of different people and sometimes not always do you identify them on the first pass. That it is important to allow opportunity for feedback to all players, that the scrips and training are key. There will be always unexpected problems, such as Y2K.

Rank and file physician buy in is important. Individual measurement and feedback is helpful and like all processes, it takes time.

What is the carryover to NAHDO's proposal to collect race and ethnicity on claims and encounters? You will note that this is collected on the registration file. I asked various information management staff how much of a problem it would be to reproduce this information on claims and encounter files. The uniform answer that I received was that the transfer is not a major issue for this institution.

Because we routinely alter patient information to fit the needs of each payer that we have, this transition is equivalent to modifying a payer's request estimated to take perhaps at a maximum one to three hours of programming time per payer.

In closing, I would suggest that our experience indicates that race and ethnicity data can be successfully collected in both hospital and clinic. Our situation, of course, is not universal. We are a large multi-specialty practice with systems in place, but I do think that the lessons we have learned could be usefully carried forward to other organizations.

I think that it would be extremely helpful should you proceed in this path and really otherwise, for some organization to be identified to pull together curriculum for medical practices that want to collect this information.

I have included in my testimony examples of what such a curriculum might contain based on our individual experience. In addition to having a written curriculum, I think that seminars could be offered at any professional meetings of clinic and hospital administrators because I believe that many providers truly are interested in having these data available to assist them in the quality of care issues that Mr. O'Keefe has mentioned.

I think that if a variety of agencies and organizations are interested in having these data collected, it becomes incumbent upon them to provide some assistance in this process.

In closing, I would like to thank you for this opportunity to share Health System Minnesota's experience and be happy to answer any questions during the question period.

MR. CHENG: Thank you. My name is Paul Cheng and I am the chief financial officer for the Union Health Center in New York City. The Union Health Center is a freestanding, not-for-profit, Article 28 diagnostic and treatment facility, licensed by the state of New York.

What makes the Union Health Center so unique is that it is the health care arm of the UNITE. UNITE is the Union of Needletrades, Industrial and Textile Employees. UNITE! is the result of a national merger between the Unions of the International Ladies Garment Workers Union and the Needletrade and Textile Employees Union.

Union Health Center was established in 1914 by the ILGWU and was the first Article 28 facility in New York. It was to provide health care services to garment workers in the sweatshops, who were exposed to the rampant TB epidemic at the time. The garment workers at the time and it is still true today is composed of a very large, diverse immigrant population.

In order to properly serve and meet the health care needs of the union membership, a large database must be established and maintained. This database at the Union Health Center is composed of the standard demographic, insurance and medical data of the population.

Because of the cultural diversity of the patient database, it was decided several years ago to begin collection of race and ethnic information. The Union Health Center understood that many health care problems stem from the unique cultural backgrounds of the union members, which could be worsened by the repetitive duties of the garment workers.

The inclusion of race/ethnic data required changes in work flow, data collection, programming and sensitivity training of the staff. Our current database shows that the race/ethnic population to be composed of 10 percent black, 45 percent Hispanic, 24 percent Asian, 12 percent white, 2 percent other and 7 percent unknown.

This is further broken down by sex, age, diagnosis categories and requires constant updating and maintenance. This data is essential for the health center in the treatment and planning for the health care needs of our specific population, the union membership.

Immigrant health care problems are vastly different than those of the non-immigrant population. This information, especially the race and ethnic information, assists the Union Benefits Fund in their planning for health care expenditures for the union membership. Now, the union membership in the Tri-State area along is 150,000 members, not counting their families nationwide. It is a very, very large union and we have centers located throughout the country.

It is my understanding that the union currently does not collect race/ethnic information since the regulations do not mandate the collection of this data. That is why they rely on the health center's data.

Though the health center consciously made the decision to collect race/ethnic data, it is not without controversy. To this date, it is still very controversial, especially the reliability of the race/ethnic data. The information is gathered voluntarily from the patient or by observation of the registrars or by the medical notes of the providers or by the association of patients' names and with any combination of what I have just said.

Most difficult are the children of interracial marriages. Where do you classify the children as far as race and ethnic data if the parents don't agree, they don't know or they don't care? Not to underscore the important of this issue politically, this is a very difficult -- it is very difficult to deal with.

As to costs, the Union Health Center incurred thousands of dollars in reprogramming its computer system with its software vendors. The health center's electronic medical record system was not written to collect this type of data. Work flows had to be changed to accommodate the collection of the data, which now required retraining at an additional cost and cultural sensitivity training, as I mentioned earlier, of the staff that added even more costs.

It is interesting to note that the cultural sensitivity training took many, many months and it is still ongoing because of various cultural diversity of our population, even hand gestures, body language can be offensive to very different cultures. So, we had to make the staff very, very aware on what to do and what not to do. But it is costly.

And the Operational Policy Letter No. 93 from the Department of Health and Human Services, HCFA, dated May 1999 for Medicare Plus Choice organizations, guidance in the collection of race and ethnic data. The five categories of race, American Indian or Alaska Native, Asian, black or African American, Native Hawaiian or other Pacific Islanders and white, and the two categories of ethnicity, Hispanic or Latino and not Hispanic or Latino. I mean, that is some category, two categories. Everybody falls into one of the two.

It is helpful in one way but I don't know how helpful it would be if you are looking at the general population. It has got to be a little more specific and allow for we don't know. If it is utilized by all health care organizations, it would require a lot of reprogramming of their systems.

Besides, what if the provider is unable to determine the status of the patient as far as race and ethnic is concerned or what if the patient refuses to respond giving the proper answer. There is no category for unknown, as I said before. The additional direct and indirect costs could be enormous if providers as a whole are required to collect this data.

I am not talking just about the institutional providers. If you are going to collect data, let's collect it from all right down to the solo practitioner because not everybody becomes institutionalized.

We all know that the most expensive provider in incurring costs due to change are the institutions, hospitals, perfect example. It is my personal opinion that the business case to be made for the collection of this data is strong. I personally do not have a problem with it. What I do have a problem with is the notion of inclusion of this data in the billing/claims transmission records. A bill is a bill. It should only require what is necessary to adjudicate the claim. Nothing more, nothing less. I personally think it should belong in the enrollment transaction.

As my colleagues who preceded me indicated that they don't have access to enrollment data, well, you know, Congress is still here. They will still be around. Have them pass the regulations that require that.

Providers who currently transmit claims or who plan to do so in the near future are very leery of this idea of including it in the claims transactions and the question when I asked several of my colleagues that they raised was why -- or statement was why give a payer another excuse to reject the claim if the race and ethnic data is not included with the transmission.

I mean, once you make it a requirement, there is someone out there that is going to say, well, you didn't include it there. We are not going to pay you the claim. Well, gee, we don't know. Do you want us to make something up? Because that is what is going to happen.

If you make it a requirement and it is not known, I will tell you right now I am going to direct my patient accounts manager, you send the claim in and you don't have it there and we don't know what it is, well, put in anything so we can get paid. That is exactly what is going to happen. No ifs, ands or buts. I am going to instruct my staff whatever you do, do whatever you have to do to get the claims paid.

What is to prevent a payor from misusing the information or anybody else? What assurances do we have that this data is not going to be misused? Red-lining is the perfect example.

In addition, to a provider, race/ethnic information is not a service delivery issue. It is a public health issue and that is where it should remain. That is the proper domain. The delivery and quality of care, health care services, are the prime concerns of the providers, not public health issues as it relates to race and ethnic. Yes, we are interested in it. We are concerned, but don't put the burden on us. We have got enough headaches.

A provider may for its own purposes attempt to collect this data as in the case of the Union Health Center. And I tell you right now from my own personal experience in the industry and also doing a lot of consulting work, I have yet to come across any institutional provider that does not collect this data. They are very, very interested. They want to know. They need to know the ethnic and racial backgrounds of their patient population to provide the proper services that are necessary and required or what they feel is required.

You are not going to get that much resistance from what I can see from the provider community, but put it in its proper place.

For public health purposes the collection -- it is clearly stated. The mechanism at this point as far as collection must be uniform. It has got to be consistent. Safeguards must be in place to prevent the misuse of this very, very sensitive data. It is really a political quagmire.

My own personal involvement, for example, with trying to get health -- ethnic and racial data as far as health care is concerned, in my limited dealings with the Asian and Pacific Islanders Coalition for AIDS, they are having a tough time getting information. They are having a tough time getting information from other sources and it is very, very important information. But they face the same problems like everyone else faces and it really is the bottom line. Who is going to pay for all of this? There is a cost associated with this.

I hate to just sound like a finance guy but I am a finance guy. Who is going to pay? The feds are cutting back. The states are cutting back. Insurance companies are cutting back. Who is going to pay? Because we are not talking just a couple of dollars here. You add it up. You are talking millions and millions and millions of dollars that are not available.

So, what you decide, it is going to be very, very important. It is going to affect the bottom line. And I hate to say it, but the bottom line dictates, regardless of how we all feel and regardless of how noble the cause may be, it is still the bottom line is going to dictate.

Thank you.

DR. LUMPKIN: Emilio.

DR. CARRILLO: Good morning. Thank you for inviting me to speak to you.

I am going to speak about the collection of race and ethnic identifiers by managed care organizations. I serve as medical director of a Medicaid HMO provider-sponsored, sponsored by a large hospital health care system in New York City, the New York Presbyterian Health System.

Recently, I attended a meeting here in Washington earlier this month, sponsored by Health and Human Services, as well as the Commonwealth Fund. This was a meeting whose purpose was to discuss the issue of performance measurement in managed care and its role in eliminating disparities.

Many leaders from various managed care organizations around the country, both commercial and Medicaid, were present, commercial, provider sponsored. There were members there from the group associations in managed care. I was not chosen by the group as a spokesperson. However, I have agreed to bring to you some of my own feelings on the subject, as well as to reflect on what was said at this particular meeting, which we all felt was quite constructive and instructive in terms of what came out.

The first finding that might be surprising is that currently this data is not being collected for the most part. The Commonwealth Fund carried out its survey of 12 of the largest not-for-profit managed care organizations around the country and they found that not a single one of them collected this data.

Secondly, the group was unanimous in its desire to see this data. The group felt that there was a need for this data, echoing the concerns and the concepts that have been mentioned by the panelists that preceded me. Why do MCOs, managed care organizations, why do MCOs feel that this data is important? Well, first of all, to compare plans. MCOs find competition very important and this is another area in terms of quality improvement where competition could be advantageous to many plans.

Secondly, plans feel it is important in order to identify areas that are ripe for quality improvement and disease management. Now, there were two central issues that were tackled at this meeting is the very basic question, should MCOs collect data on race and ethnicity and, second, a less conceptual, more practical question, if so, how?

What I would like to do is just present you with the issues and concerns that were presented on each of these areas and also to present you with some of the proposed solutions and possible alternatives in terms of each of these particular items.

First of all, in terms of the question of should MCOs collect this data, there are many issues that were raised and most of the plans were quite, quite emphatic in their concerns around these issues. The first issue is that managed care organizations are concerned at being singled out and being mandated externally to gather this data. And they are concerned that if managed care begins to show disparity in its services and the fee-for-service environment is not also collecting this data and is not also processing and analyzing this data, it would seem as though the managed care plans are once again the devils.

Given the current national public sentiment around managed care plans, everyone in the group felt that it would be a real danger to take on the collection of this data ahead of the rest of the industry. Secondly, the issue that was also very concerning is an issue that already has been mentioned by Paul Cheng, which is the issue of the misuse of data. There are issues of denial of enrollment, cherry picking, red-lining. We can call it several things.

There are issues of the risk adjustment of data. Once managed care plans begin to adjust their data for risk associated with, you know, certain racial/ethnic variables, they might then feel abdicated about their responsibility for having low performance. A managed care plan may have risk adjustment of their data, showing that they have very low mammography -- accomplished mammography in their members, they might feel that, well, we have very high risk. We have a lot of minority women. So, therefore, we are not so bad.

That is a problem. Thirdly, within the area of misuse is the area of discrimination. I mean, discrimination happens in our society at all different levels and once this data is available, the society will be very, very creative in finding ways to discriminate within the MCO environment.

The third issue in terms of whether this should happen is the legal issue. Is this a violation of title stakes of some other regulatory mandate or federal or state. Now, at this meeting there were representatives

-- high level representatives of the Department of Justice and even they have disagreement among themselves in terms of to what extent there was legal issues, to what extent there were not legal issues.

So, there has to be some more clarity in that issue. And that is an issue that concerns, if not a real issue, certainly a perceived issue, just a wide perception that there is confidentiality and legal status that this data cannot be collected.

A fourth area of concern in terms of whether this should take place is the cost and I think that Paul has addressed that already. The fifth issue, which I think is a very important issue, is that in the absence of regulatory requirement from the Federal Government or the state governments and in the absence of strong purchaser requirements of this data, managed care organizations are not going to be driven to engage in this necessary endeavor.

Those are some of the more conceptual, legal issues. There are also some very important technical difficulties that were raised by the various representatives of the MCOs. The first such difficulty is the lack of accurate practical prophesies for the collection, which would be standardized nationally, in order to facilitate comparisons. Again, unless there is a standardization, sort of as you have in the HEATUS data set, unless you have that standardization, which takes a lot of work and a lot of thought and it is not something that just can be developed over night, unless you have that, you really cannot begin to compare across the country or even across the street.

Secondly, another technical difficulty is the question of validity of the data in the absence of linkage to SES and acculturation data. I mean, we have learned that -- Dr. Ramado(?) and others have learned that Latinos, acculturation is a very important indicator. Also, SES, there is just a plentitude of data on the importance of that and, yes, there are strong linkages with racial and ethnic identifiers but that is something that remains an issue.

Thirdly, there is tremendous intergroup variations. Hispanics, you know, there are Cubans, Puerto Ricans. There are graphic differences. There are differences in terms of disease prevalence and many other things. Asians, Japanese, Vietnamese and even within blacks, there is the African American. There is the African Caribbean and there is the African immigrant, of which in New York, for example, we have many many.

So, there are subtleties that are not gathered in these identifiers that are of concern to many members of the managed care organizations.

Fourth, there is another concern, which is in terms of the lack of a standardized validated method for the actual collection of the data. At what point is the data collected? How is it collected and so forth and so on?

Fifth, there is issues in terms of comparability of self-reported data. There was a feeling among many in the group that self report would most likely be the best way to gather this data. However, questions came up objective and societal considerations in self reporting. This is something that Robert Hahn(?) has written about, Gregory Robinson here in the government and Paul Cheng just mentioned some examples about that, issues of intermarriage, et cetera, et cetera.

So, the self reporting is an issue that needs to have some clarification. Now, that is all very bleak and that is only one side of the coin. There are solutions. There are alternatives. There are ways that the group felt these issues can be addressed. Again, I say the group as a whole felt that we should proceed with finding a way to collect this data, but keeping in mind all these caveats and all these considerations.

In terms of the question about the managed care plans being at a liability because they would be the only ones that would be sort of the leading horse in this equation, this data should be collected universally. I think that that is clearly something that you are all addressing in today's meeting.

Secondly, the misuse of data, this really demands that there be federal and state mandates that would be in place, not just there on the books, but also be enforced in terms of how to deal with this information. That is very, very crucial.

Thirdly, in terms of the legal issue, it is important that there is a clarification that the Justice Department take this issue, examine it and come to a conclusion on what it is, what the factor is. Once this is out from the Department of Justice, it is something that has to filter down, that has to be discussed in the state governments and has to be put out up front in a very visible fashion.

The question of additional cost comes up and, of course, the MCOs are looking for compensation. Now, where would that come from? That depends on the situation, type of plan, et cetera, et cetera.

Then in terms of the absence of regulatory requirement is important. It is imperative that the Federal Government and the state government begin to mandate, regulate these issues. HCFA, for example, is perhaps the largest purchaser in this country and HCFA has tremendous leverage. So, if HCFA takes it on and HCFA basically makes this part of their purchasing leverage, this is something that really, you know, would really happen much easier.

In terms of the technical difficulties that I mentioned, there are also considerations for solution. The first is the lack of a practical accurate process. We have the OMB Directive No. 15, which Anne has pointed out in her letter and this is something that is limited in some ways, but it is there and there is some experience; five categories of race and two for ethnicity.

There is also the CAHPS, the Medicaid CAHPS Survey, which is the Medicaid Consumer Assessment of Health Plan Survey, which is a national tool, which includes information on race and ethnicity and language, which is a very important measurement of acculturation. So, the CAHPS is one way of getting at this information within the managed care environment.

We then had the question of the validity of the data, given the fact that we have SES and acculturation as confounders and factors that can aggravate the data, the value of the data. And, basically, all you need -- we felt that all we need is basically to propose a two-step model, where we basically generate hypotheses for analysis and intervention development through racial and ethnic data, pure racial/ethnic data.

Then once you get into the level of analysis and intervention development, at that point you then gather other data, SES data, acculturation data, that will basically give you more powerful data with which to make your analyses and create your interventions. The third issue, the issue of the intergroup variations, again, this is very important but clearly there has to be amendment of such things as the OMB directive. There has to be creation of tools that are more precise.

Fourthly, in terms of the lack of how to gather this data, most of -- I think all of the group felt that self report is the way to go immediately after the enrollment. However, as was pointed out earlier there are issues with self report that have to be taken into consideration. There has to be an assessment of the literature and the validity of racial/ethnic self report. It is there but it has to be looked at, analyzed and there has to be other surveys, other studies done to get a better sense of this very important issue.

There were a few steps that were recommended by the group, which I will share with you. First is the group feels that we have to work towards a national policy on the collection of race and ethnic data that cuts across not just the managed care environment but also cuts across the fee-for-service environment. Certainly, I see that happening here in this room.

Secondly, there needs to be a small step before you take a big step and within managed care environment a model program would be interesting. For example, Latinos suffer disproportionately from diabetes. So, programs that would begin to collect ethnicity data on Latinos and then link that with disease management program to have early detection of diabetes would be a very good step forward.

I mean, we know that Latinos may spend 20 years without having -- with not knowing they have diabetes. By the time it is diagnosed, it is too late. So, that is an example of a model addressing one disease specific that would make sense.

And to close off, the third point, leverage by HCFA, the large purchaser, and finally to consider other sources of data within the managed care data set, for example, getting information from the Medicaid rolls where that is possible and by carrying out small area analysis as was mentioned earlier.

I will be happy to go into more detail and answer questions regarding these things. Thank you.

DR. LUMPKIN: Thank you.

I am sorry to say this to the final speakers but you were asked to stick to a ten minute time frame. So, I am going to kind of ask you to do that or else we won't have any time for questions.

MR. KNETTEL: I am speaking this morning on behalf of the ERISA Industry Committee or ERIC for short. And for those of you who are not familiar with ERIC, we are an association of over a hundred of the nation's largest employers. Typically ERIC member companies have 50 or 60 thousand employees. So, that is the specific context from which I am speaking.

ERIC does not have an approved policy position on the specific issue of the collection of race and ethnicity data, although we do have some principles on the collection of health related information and I think that those principles are relevant to this discussion. So, I would like to summarize them just very briefly.

In a nutshell, ERIC's members believe that their employees have the right to expect that if their employer collects sensitive data, that it should be collected pursuant to a specific business need, that the disclosure of that information should be made only to those people within the company, who have a business need to use it and that the data should only be used pursuant to the specific need for which it was collected.

So, for example, if health related information were collected in order to administer a health benefit plan, ERIC members acknowledge that it would not be appropriate to disclose that information to an employee's supervisor who has an employment discretion over that individual.

One of the implications of those sets of principles, however, if ERIC members don't perceive that they have a specific business need for collecting information, they don't collect it. In fact, when I spoke with ERIC members in preparation for coming to speak to you today, few if any of them indicated that they currently collect or have an interest in collecting demographic information.

A number of the preceding speakers, I think, made a very articulate case for why, in fact, there is a business need and a public health need to collect this information. I don't disagree with those arguments, but out in the real world among the ERIC member companies, who responded to my inquiries, I think people have not focused on and not really identified those business needs.

In fact, it was much more likely when I talked to ERIC members that they expressed that, in fact, they had no business need for this data, had no desire to collect it. In short, didn't want to be anywhere near it. I think the reasons for that are fairly straightforward and reasonably obvious.

In particular the issue really isn't cost, especially for an ERIC member size company. While the cost of collection and reprogramming and so forth would not be significant -- excuse me -- while the cost would not be insignificant, that is not really the driving factor here. The driving consideration is employee relations and the driving consideration is aversion to legal liability.

Just as our members tended not to see a specific business need for collecting data on race and ethnicity, so, too, I think their perception is that their employees don't perceive a need for it either. So, a frequent concern that was raised was that if employers in enrolling their employees in health plans and in engaging in other activities requested this kind of information and employees did not see a specific need related to the health benefit plans for collecting it, that the employees would assume it was being collected for some other purpose and, in particular, be suspicious that the information was being collected for purposes of employment discrimination.

So, it is partly an employee relations issue in terms of not wanting to collect information that potentially could create employee relations problems, but also increasing concern about potential legal liability that might follow. And we are not just talking about federal or state liability. Many ERIC members are global companies, for example, and they are trying to figure out what exactly the new European Union data collection requirements are going to have in terms of impacts on them.

Those requirements are much broader than health related information. So, to the degree that employers were getting into the business of collecting and communicating race and ethnicity information that is sensitive information, that entails with it confidentiality requirements, potential liability and so forth. So, they are extremely concerned about being brought into this context especially when, as I said, many of them do not yet fully appreciate the important public health and business purposes for wanting to collect this kind of information.

In summary, I think, ERIC members are not unsympathetic to the arguments to collect and make appropriate use of this information, but they are very much nervous about the potential liability and employee relations problems it could cause. I think that that suggests that should you go forward in this direction, that there is a very significant need for public education, both of employers who would be one of the nexuses of collecting this information, as well as education for employees that people appreciate why the information is needed and why it be communicated and perhaps also find ways of indicating at the time that the information is collected that, in fact, it is being collected pursuant to a government requirement, that this is not being collected because the employer intends to make an appropriate use of it.

With that, I certainly want to -- I just want to conclude by associating myself with the comments that have been made by the prior panelists. I think they have both made very compelling arguments as to why this information is needed, as well as made very compelling arguments about many of the pitfalls of doing so. I would certainly urge caution as we move forward in this particular area.

Thank you.

DR. LUMPKIN: Thank you.

MS. WOO: Emilio gave a very good summary of the barriers that managed care organizations face in collecting racial and ethnic data from the enrollees. One of the issues that he pointed out was the legal issue for violating a state or federal legislation.

Ed and I are going to give a very brief summary on that part. We did a quick surface investigation of state laws, which may prohibit the collection of racial and ethnic data by health insurers from the enrollees. Of the 48 states that we received feedback from, we only received response from two that said there is a section in their state code, which does specifically prohibit the collection of racial and ethnic data from the enrollees.

The State of California, 1969, enacted that no application for insurance or insurance investigation report furnished by such an insurer to its agents or employees for use in determining the insurability of the applicant shall carry any identification or any requirement thereof of the applicant's race, color, religion, ancestry or national origin.

In 1973, the State of Maryland passed under prohibited inquiries, an insured agent or broker may not make an inquiry about race, creed, color or national origin in an insurance form, questionnaire or other manner requesting general information that relates to an application for insurance.

Now, it is a question neither Ed or I are state experts in the code and interpretation and application of the code. So, this is information that we required from the legal division of the State Department of Insurance.

As far as the state law in California, I was told that that applies to health indemnities and preferred provider organizations and not managed care organizations and I haven't received feedback from the Maryland Department of Insurance's general counsel for the exact interpretation and application of that law.

So, there is not evidence against the collection of racial and ethnic data from health insurers on a state basis. Ed will elaborate more on that. But one of the common responses that I received from the State Department of Insurance is that, no, they do not have a specific section of the law, which prohibits the collection of racial and ethnic data, but they do not collect the data.

If they saw that question on an application form, they would red-line that because it would probably violate a section of the unfair trade practice or discriminatory practice, which, of course, would be a violation. So, even though it is not specifically prohibited, it is kind of in that gray area of it may violate another section of the law. And Ed can elaborate more on the legal issue on that. Because this is coming from a non-attorney, non-state expert.

MR. WOO: My expertise is not in the area of data collection. I work primarily in the area of civil rights enforcement. So, I am familiar with some of the issues that the prior panelists have raised. I do want to underscore at the outset that from a civil rights enforcement perspective, the collection of race and ethnicity data is very important. I can't underscore that enough.

We are often involved in investigating complaints of discrimination where we have to go into a facility. I work in the area of Title VI enforcement, which would cover hospitals and other facilities that receive Medicaid and Medicare reimbursement, for example. So, we are often involved in investigations where we are trying to piece together what a facility's practices are and how their practices affect minorities, for example.

It is very hard to get timely data and, so, we are often at a disadvantage trying to understand the day-to-day operations and what is happening at the patient level and how the patient proceeds through the hospital system, for example.

That being said, I also want to underscore that the issues of sensitivity about the collection of race and ethnicity data have been around for as long as the civil rights laws have been around. In the health care area we are all grappling with it now, but in other areas, such as employment and education, some of these issues have already been dealt with and basically there is a recognition that this information is sensitive, that the patients or applicants on one hand regard with suspicion people who are asking for this information.

On the other hand, employers or other facilities who are trying to collect the information are rightly concerned that those patients or applicants view this with suspicion and view any action disadvantage to them as triggered by this request for information.

In the balance in those areas of employment and education, the balance is tipped in collecting information. It does take some education to assure patients and others that this information is not going to be used against them. Often that is done by making the data collection voluntary or by assurances on application forms that this is being used for civil rights enforcement purposes.

So, there are ways to provide the confidence that this is not going to be used against the patients.

I just wanted to touch briefly on this question of if we collect the data, will it be misused by people. Obviously, that is a risk. On the other hand, from a civil rights enforcement standpoint, the data that is collected helps protect against misuse.

If we can see how applicants are being treated and if we see disparities based on race that can't be explained by other reasons, that is a protection. So, I did want to respond to that because from a civil rights enforcement standpoint, that information is vitally important and it serves as a protection not as a real risk.

Now, the other thing I want to point out is, because I know there is this concern about whether people can collect the data and whether it violates laws or it doesn't. In my experience, we have never faced the issue where somebody said this is absolutely forbidden by state law and there is no way to collect it without violating the state law. Like I said, we recognize that there are concerns about confidentiality, privacy, how this information is being used.

Those in my experience have always been able to be worked out at the facility level. We have never run into a situation where a state was saying we forbid it and even for the purpose of civil rights enforcement.

I don't have any prepared presentation. I am just here more as a resource on these issues of civil rights enforcement. I would be happy to take any questions.

DR. LUMPKIN: Thank you.

Nancy.

DR. KRIEGER: Thank you. I come today as an individual. I am not representing any organization and I came specifically -- Dr. Marjorie Greenberg asked me to come -- to comment in part on the different kinds of presentations that have been given today and to raise, I think, some additional issues as well.

Just to give you a little bit of a sense, as a social epidemiologist, I am very involved in the collection and analysis of data. Having worked at the Division of Research at Kaiser in Oakland, I am very aware of issues within HMOs about collecting data and now working on projects in terms of looking at social inequalities and how using data from public administrative databases and particularly vital statistics, I am also aware of issues that come up in data collection in these matters.

I feel that a central theme that is coming up in the comments is that we do have discrimination in this society and we do have inequality and that a fundamental standpoint of public health is to want to minimize these problems in our society, to recognize them and to minimize them. Not to collect is a way of not seeing that ends up causing injury and that to not have data is, in fact, a statement saying that we will not pay attention to what we know are problems in our society. We will not have the information. We will not be able to monitor.

Monitoring is a key component and that is where having data and administrative databases is very different from having specialized studies and surveys that can go on but are not necessarily useful for ongoing monitoring to see what the problems are and whether the problems are being diminished over time in terms of inequality in access and also inequality in care itself and inequality in other components that are played out in our bodies in terms of our health.

So, I think it is very important to say that the starting point of a discussion like this is that we have concerns that there are inequalities that are preventable and that we need to monitor them to see what is happening and that we know in this country race/ethnicity is one of the important axes of inequality.

We also know that it is not unto itself the only axis of inequality. I would really urge the committee to consider -- not to repeat mistakes that exist in some of the current databases, which are solely racialized and do not collect information on socioeconomic position, that to continue to have it became a matter of only race/ethnicity, one, leads to a lack of understanding that there is economic heterogeneity among racial/ethnic groups, populations of color and also the white population. Once again, the poor whites disappear from view and that is a problem.

But also not seeing the increasing economic heterogeneity among populations of color also perpetuates stereotypes and also makes it impossible to see what the actual health care needs are of changing populations or of changing demographics. There are ways to address not only getting racial/ethnic data but also socioeconomic data. Whether or not it is through actually something that you do on the form and asking questions or using -- because you have address information, processes of geocoding and using area-based measures of socioeconomic position, that becomes another possibility and I think is a topic that is of concern and consideration that could be discussed further.

But I would urge that it be seen that the issue here is one of inequality and the components that are of concern here are race, ethnicity and socioeconomic position. Not either/or. Both are necessary to understand what is going on, what is needed and both can be addressed.

I would like to underscore also another point, again, which I think has been spoken to, but that to look at only issues of misuse misses the point about the harm caused by no use of the data. I just would like -- that point has been raised by other panelists, but I think that it has to be there. If we do not collect information in a way that is systematic, that there is public access and there can be monitoring, you don't know that there is, in fact, racial/ethnic inequalities in health. You can't understand the link to socioeconomic position as well in terms of where resources and remedies need to be directed.

So, to have silences in data can be injurious, as well as actually collecting data and having misuse of that data. I think that those both need to be kept in mind.

On the concerns I am hearing about validity of data, there is no need to reinvent the wheel in terms of all the research that was done by the OMB in terms of the new definitions and collection around data around race/ethnicity for the year 2000 census. I think it is important to keep in mind that those OMB categories are the ones that ultimately have to deal with the collapse, too. They all say that you can't collect more detailed information. They say whatever information we do collect should ultimately be collapsible to these categories.

So that you can actually ask for more detailed information and that depends on the nature of your questions and concerns that you have. You have to at least meet the standard, but you can go beyond it. So, I think it should be seen as a floor and not a ceiling.

Also, in terms of the questions about self report of race/ethnicity, that is a key standard form of collecting the data and, again, you don't need to reinvent the wheel in terms of all the OMB research on that. That said, considering one of the major databases in this country, the SEER cancer registries, they rely on an amalgam of data that is collected through medical records and charts to get the race/ethnicity data of patients.

That information is used for cancer planning even with its limitations with an understanding that ultimately end up with a conservative bias and the bias when you have problems in collection of racial/ethnic data can be assessed empirically. So, if you have a situation where you do not have -- you have concerns about misclassification, this is amenable to analysis and what the analyses show repeatedly, whether it has been with follow-up studies that have been done around the cancer registry records or others is that you tend to have an underestimate of the extent of racial/ethnic inequality.

Similarly, with regard to socioeconomic data given underreporting at the extremes, you end up with a bias towards a conservative estimate of what the problem is. From a standpoint of needing to address a problem, you know that if it is at least this bad, that you are not overstating the case.

So, I think that when you deal with concerns about questions about validity of data, you have to step back and think empirically in what direction will the bias affect the results. If it is a conservative bias, you know you have more of a problem, not less of a problem.

So, I think that these are things that can be addressed and thought about and dealt with a rational, empirical manner.

I guess the last point that I just would like to raise is that it is, I think, fairly critical at this point to have data towards what the consequences are of actions as seen in what kind of health care is provided or not.

One of the areas that I do research on is also around discrimination in health and including people's self perceptions of discrimination and also how you have to have data that is not just about what people themselves report, that is about what the actions are that are taken. If you are concerned about looking at inequalities in health in relation to racial discrimination in health care, you actually need to have that information about what have been the trends -- what have been the outcomes in terms of utilization of services.

To ask patients and to ask providers about attitudes is not sufficient. That is because there are problems when people report. People who are the ones that are on the side that would be the discriminators are more likely to underreport that they have discriminatory attitudes than that they don't.

In terms of people who are being discriminated against, it is very hard to know, in fact, you are being discriminated against unless you have a standard of comparison. You need to know information about what is happening with other people and, in fact, people's self reports of discrimination tend to underestimate the experiences that they have. In fact, there is social psychological and cognitive research on this now that show that there is also a cost to reporting that you have been subject to discrimination. It is not a fun thing to think about.

In fact, in the research that I have done, the area when asking about people's experiences of discrimination from police, on the streets, in the courts, in restaurants, in various different settings, notably, the one that had the least reporting of discrimination was in medical care settings. That is because that is a place that you are extremely vulnerable and you don't want to think that you actually are subject to racial discrimination.

So that while it is going to be important to have studies that actually look at what people report based on their experiences, that is going to be insufficient to actually capture what is happening in terms of whether procedures are being sufficiently recommended or differentially recommended for different groups, based on race/ethnicity and socioeconomic position.

So, having both the administrative data that captures what the actions are, but ends up having consequences for health.

Some of the kinds of comments that I have heard today and I very much want to urge the committee to, one, move forward in collecting data on the race/ethnicity because that is part of the fundamental public health mandate of monitoring the public health.

Secondly, to benefit from the work that was done by OMB and not recreate the wheel in terms of how these data should be collected, also to have comparability to census data in terms of denominators, et cetera. And, third, not to see race/ethnicity as something in isolation, but rather to understand what we are talking is social inequality in its manifest form and that we need to be able to link these data to socioeconomic position and that there are data that can be collected, asked of individuals, but you can also -- and their household conditions, but you can also move with possibilities of geocoding and data linkage to area-based measures and, therefore, not have to get into the question of what data you are collecting at the point of the health encounter itself.

DR. LUMPKIN: Thank you.

Questions and comments?

DR. STARFIELD: I would appreciate knowing from Dr. Krieger, you know, what your own experiences are with geocoding, what its problems might be and, you know, is it really a solution to what we are having to deal with, that is, having to collect the data in a standard way for everybody.

DR. KRIEGER: I have worked with geocoding of public health databases both collected by surveys, also in terms of records at Kaiser. For example, at Kaiser it was incredibly useful because there was no membership information on socioeconomic position, that there were assumptions being made about what the membership looked like.

To actually geocode the data at Kaiser and put -- find out what block groups people lived in and then get the sociodemographic characteristics of those block groups and then compare that to where people were in the general population showed that as you expected you had generally a good bulk range of the population across various socioeconomic strata, not many as you would expect among the very poor and not many as you would expect among the much more wealthy block groups.

So, you had a reflection of a reality that you think would be there, which we could confirm. Again to the extent that there were individual surveys that were done in Kaiser, it was a very useful technique.

I think in the United States -- so, I have been involved in work on validation. I am involved in a much larger study right now involving the databases in Massachusetts and Rhode Island to really rigorously look at what are the area-based measures of socioeconomic position that could be used in the United States.

In the United Kingdom, for example, there has been extensive research in various area-based measures of deprivation. There hasn't been any analogous effort in the United States. It is much more eclectic at this point. But there is a wealth of information from the census that can be possibly used and it is used all the time by marketing firms in one form or another to identify different market niche segments.

I think, public health, we have a different standpoint of who we are trying to identify. There are problems in terms of the timeliness of the data. Obviously, you have a decennial census. So, it is every ten years. You can look at questions about -- and it has been done that -- our concerns about gentrification, other changes in cities, but that there is a certain consistencies of bad neighborhoods staying bad neighborhoods, good neighborhoods staying good neighborhoods. You can figure out how people are moving.

There are lots of -- again, in any area there are technical questions that come up that can be addressed and it becomes a useful tool. So, I think it can be a tool for the uses of this committee and that is a whole -- you would need another whole panel to present on what the pros and cons are of geocoding.

I feel that there is much more technical capacity now to do it than there has been before with new software packages and more attention to the possibilities of using these kinds of data. I think there is an increasing interest in public health in this kind of approach. I know that cancer registries, the SEER registries in particular, are expressing interest and that there are ways that instead of actually having on the file, for example, what the person's geographic codes are in terms of confidentiality, rather needing to specify what would be the appropriate variables to append, so that you don't say this person lives in this county, the census track and this block group, but rather this person lives in this kind of block group.

There needs to be more research but it is an empirical question that can be resolved.

DR. HARDING: I was wondering, who put this panel together? Is that Marjorie?

MS. GREENBERG: Why did you ask? I have to take credit.

DR. HARDING: I was going to say that this is one of the most lucid and clear groups of individuals that we have had testify before this group and very much appreciative of the information and the clarity and keeping within time limits and all the rest that you had to do doing this. I was going to say somebody -- not only the panelists but whoever put it together, it was very interesting.

MS. GREENBERG: Thank the panelists for the --

DR. HARDING: Yes, I do thank the panelists very much.

The question that I have and it is just a brief one, you talked about that race and ethnicity is often a -- is voluntarily stated by the registrant. Have you ever seen or had indication that there is gaming of that system by the registrants for any reason, a systematic bias?

MR. CHENG: I suspect that there is some gaming. You know, the registrars -- well, sometimes when I look at the data and I compare it from the previous year, there are some disparities which are unexplainable, sudden shifts, yet the patient database really hasn't changed that much.

Yes, I suspect that it is there. I have not been able to prove it and to try to prove it would mean that I would have to take disciplinary action against somebody.

DR. FOWLES: I would be interested in what you mean by gaming.

DR. HARDING: Because it is self reporting, changing the category on the basis of some advantage to me as an individual or would there be any way of following that process or seeing a change. I am not accusing. I am just wondering if -- the temptation for me would be that if I had some advantage to be a certain minority and I had a great, great, great, great grandmother, who was, would I say it on the basis this year as opposed to next year, when there might be a different advantage or disadvantage?

DR. CARRILLO: Once you already have the entitlement, you know, there really is no advantage to the gaming. I mean, at that point you are -- basically you have the insurance whether managed care or you have the Medicare card and in terms of obtaining the service, I find it hard to think how that would be a -- gaming would be an opportunity.

DR. HARDING: Thank you.

MR. CHENG: Let me clarify my thoughts as far as gaming is concerned. When anything is changed in our computer system, no matter what field on the registration screens, there is always a trail. We know exactly who or -- unless someone gave their password to someone else to use. We take pride in our database. We really do. And the accuracy is pretty high. So, if -- mainly because everything that we have in there is required. So, if you don't complete it, you are going to be spoken to.

My sense of the term "gaming" is that they are going to put something in there just to make sure that it is there so that the discipline process would not come down on it.

DR. COHN: Like Dr. Harding, I actually am very appreciative of this set of presentations, in a way sort of being reminded how important this data is. I especially want to thank Dr. Krieger for helping put it altogether at the end and I almost thought shining a little light on the subject.

Based on the fact that I think this is important data, you know, I leap to the question of like, well, how the heck do we get it. I am sort of -- by listening to the various panelists sort of convinced that I don't have the answer. I had thought, I think, coming into this panel that, well, gee, it should be part of an enrollment transaction. But after listening to Knettel, I come away convinced that it is unlikely to be accurately captured as part of an enrollment, both because the employers don't want to put it down.

People in the process of enrollment may not want to specify. So, I have to sort of toss that out. Certainly, I don't think that -- I heard people saying that the claims transaction is not the place to put it only because you wouldn't want to put it on every single claims transaction you sent around. That doesn't make any sense.

I listened to Dr. Krieger and I was sort of reminded that SES is probably as important as race and ethnicity. And I am perhaps going to ask a very naive question of Dr. Krieger. Is this something that should be in any of the health transactions at all or is this more of a census question that we should somehow be capturing through that or DPS(?) and that should be the level that we are dealing with with this one.

DR. KRIEGER: Let me just see if I understand your question correctly. Should there be socioeconomic data collected in the health transactions and the race/ethnicity data collected in the health transactions?

DR. COHN: Yes, or should it be handled in a completely separate fashion?

DR. KRIEGER: I think the point here is that people are trying to monitor certain kinds of outcomes and that you have to have the data so that you can monitor the outcomes. Therefore, the data has to be in the health records. Where they are in the health -- to have your events so that you can actually track the events. You have to have a relationship of those events to what makes your denominator and depending on if you are having a universal system whereby you are getting everyone you are comparing to census data, you need to have comparability to the categories of the census in order to look at the population.

So, you have to think about both who your numerators and where your events are coming from and who your denominators are. So, from that point -- you have to think about the census, but you cannot only rely on the census. The census has no health information whatsoever. We need -- if we are going to be monitoring the issues involved with access to and quality of care, and this is an important way to do that, then you need to have that reflected where the health events are occurring.

You can have, for example -- I mean, if you had the information in Kaiser, only in the enrollment form and there was a whole conference at Kaiser about how could Kaiser start to collect data on both race ethnicity and socioeconomic position that happened back in 1996, 1995-1996, and actually -- and Kathy Wolman(?) from OMB and others were there to try to talk about these issues, that is great but that is internal. What I am hearing also today is that not -- states can't have access to that information necessarily. The uninsured are completely missing from the picture.

So, if you are talking again about questions of equity and who you are trying to understand where the problems are, you need something that ultimately comes across the universal system. What I am hearing -- I am not an expert at all and pretend no expertise in these particular data forms that are being discussed. That is not where I have had my focus and orientation. So, I can't say specifically that it is one form or the other of those, hearing the different names of the forms go by, but I can say that if you do not have it on those forms, you will be missing important information.

So, I think that is the conceptual point and then it becomes a technical point as to how you make that issue feasible.

DR. STARFIELD: But if I interpreted what you said correctly, all you need on that claims form or whatever it is, the encounter form, is an address.

DR. KRIEGER: All you need -- for the socioeconomic data if you use area-based measures, yes. That will not get you necessarily the race/ethnicity data.

PARTICIPANT: You have got geocoding that gets you there, too, the same way.

DR. KRIEGER: Geocoding will get you -- you can get the racial/ethnic composition of neighborhoods. That is certainly true. So, you can use that information. My sense is that you will be better off -- there are different meanings that -- in here in individual and group level data on socioeconomic variables, including economic composition and racial/ethnic composition, that there certainly could be interesting use made of data in terms of residential composition in terms of racial/ethnic variables, but actually what seems to be most pertinent is understanding that in relationship to overall patterns of residential segregation, which is not the same as looking particularly at that neighborhood.

Whereas, looking at what that neighborhood is in terms of its actual economic factors, it is informative in another way. So that if you look at the health research and the sociological research in this right now, I think that there is a lot of importance to getting the race/ethnicity of people as they self report it, that it would be great to get the economic information, but I understand that can be very difficult to do in certain kinds of transactions and that going for the geocoding and the area-based measures of socioeconomic position would be useful.

The thing about the neighborhoods is -- well, on the racial/ethnic composition of the neighborhoods you end up with questions about basically you have neighborhoods in the societies that are very, very white and then you have neighborhoods are -- it depends on the level of geography that are going to have mixed populations and then you have issues of missing out on the people that are living in neighborhoods that are different than who they are in ways that play out differently.

I guess it is a whole other level of conversation. That is a panel, again, different than this one if you are going to get into that, but I would urge, looking at race/ethnicity of the individuals as they self reported and if then -- and if you could get socioeconomic information in relation to income, occupation, education, et cetera, that would be great. Absent that, to go for area-based measures, which do have a validity unto themselves in terms of neighborhood and community resources.

Because economic information is about deprivation. Race/ethnicity is not about deprivation. It is about who we are and you can or cannot live under conditions of deprivation in relation to that, but it is not the same thing and they should never be confused.

DR. MC DONALD: I know some of the worries on the other side of the argument and that it really all focuses down on the transaction you pick and it isn't just because it is done a lot. The payment transaction is the one with all the coercion on it and that is the transaction that has got everyone by the neck. So, that is a good one to get extra data on because you are likely to get it.

That is the same reason why maybe the people that have to worry about it are worried about all the other coercitive(?) things that will happen accidentally, incidentally, et cetera, like if it is not completed, the bill won't be paid.

Now, there is a third -- there is another option for accommodating extra fields and that would be an attachment and I guess the first question is whether that wasn't thought about because it is not as coercive and really what you really want is this coercive transaction or it just wasn't thought about.

DR. ELIXHAUSER: Let me take that.

That is something that we actually discussed and I think there was concern that attachments simply wouldn't be submitted for all patients. What we are really considering at this point -- and we haven't come to a final decision on this by any means -- is suggesting that race and ethnicity be included as an optional element on the claim simply so that states like Massachusetts and California and other states can ask for that information and that there is a standard format in which it can be provided by hospitals.

DR. MC DONALD: Just one more follow-up question. That is, the other, I think, concern is that if this is a crack in the door for, you know, basically a huge amount of work for the various providers. You start out with your hospitals. No one is worried about that. I am not. I am not a hospital. You know, at $10,000 a visit, you know, you can absorb some costs. Well, you used to be able to anyway. I mean, not anymore.

When it is $38 a visit, you know, it may not be easy to absorb the cost. So, I heard three things. One is in the terms of the whole conversation, it has flipped from -- the first focus is the hospital. You can stay where you are and not lose anymore. And then I heard but if it is going to be a hospital, by God, it should be every single provider of every kind. I heard that some of us. Maybe it is sort of a defense, the solo practitioner all the way up the line.

So, that is going to be a challenge. The second thing I heard is that it really isn't enough, in a couple of conversations. You really need socioeconomic status and actually -- and you really would like to know language and you really -- and it is true from a research perspective. There is no end to the good questions we can think of and that are good to be answered.

The other way that I heard it sliding is that the categories really aren't fine enough and that continued splitting as we see in ICD-9 and other things, which makes it more costly and more difficult and I pulled a paper from Science that talked about there really is no such thing as race. They have proven it with DNA. So, the split has got to go on for a long time to really get it absolutely right, almost down to individual, you know, genomic patterns. The business about if it is -- you know, the mother, the father is a quarter black, half Asian, two-thirds Indian. It can get really complicated.

It only gets tough because of the coercion part. So that if this thing was done as most things are done in America in a sort of what is good in the marketplace and it moves along and you have a place for it -- the coercion part is what makes it scary. In fact, the coercion part might make it so scary that we won't have this train to ride. That is, there will be enough resistance from other parts of the market that we won't end up with it. That is my biggest concern about trying to ride the train. We might break it.

DR. LUMPKIN: Let me say that we have got time for a few more people. We have myself, Kathy, Bob and then Kepa and then we are going to close it out.

I think I kind of would like to thank the panel because this really has been a very useful discussion and it has led me to a number of conclusions I will try to bounce off you.

First of all, a short little story. Every time someone in my father's extended family -- he had ten brothers and sisters, some of who have passed away, but others -- if any of them get sick, my aunt used to always call me since I am the only doctor in the family and have them talk to their doctor and it had nothing to do with giving that person medical device. It had to do with the fact that African Americans in the health system if somebody -- if they believe that you know somebody, you are going to get better care.

In this health system, if I am African American and all the data says I am more likely -- less likely to get coronary artery bypass and I am less likely to get reprofusion, but I am more likely to get an amputation, I think I am going to take my chances with -- you know, it is a choice between taking your chances by giving your race to the insurance company so that enforcement can occur on civil rights issues or taking your chances and not giving your race to the insurance company and maybe if they don't know, they won't discriminate.

If you are in that position, relying upon government to protect you or relying on anonymity to protect you, you are going to pick anonymity every single time. So, I get very uncomfortable and I guess as this discussion has gone on and I have been thinking more and more about it, I get very uncomfortable with putting race and ethnicity into 837. I just don't see the utility.

The public health side of me says we need to have this data and I think we need to find an alternative solution because it just doesn't make me feel comfortable saying that the only way we are going to collect this data is if we are going to put it in the billing form.

The problem is that we have a paradigm shift that we are in the process of and we haven't recognized it. That is that as we move to electronic transactions, and that is what HIPAA is going to force, the data is going to be there by the providers. I have a problem with that. What the electronic data interchange engine, as I understand it, maps these data elements in the provider's database and puts it into the 837 and sends it off to the hospital. So, there is only really a one time cost to have that electronic data interchange engine that people have to buy anyway, to map that into an HL7 message, attachment or something along those lines so that every time one is generated to create a bill, it would also generate a message that goes to the health data organization or some other entity.

I am much more comfortable with that kind of solution than I am by trying to cram everything into a bill that goes to an insurance company that I can't trust to not red-line.

Any comments? Or maybe we can just toss it out and let --

DR. ELIXHAUSER: I guess I do have one comment. Your comment about anonymity is a little bit intriguing because the relationship between a patient and a physician is not anonymous and the research that we are seeing now, the Schulman study that I referred to earlier showed that there was significant differences in how physicians treated patients based on their face with everything else held constant.

DR. LUMPKIN: I am talking about anonymity between me and my insurance company, my payer. All they know is I work for the State of Illinois and that I am an employee. If they get an 837 bill that has my race on it, now their database has my race and when the physician calls up to get prior approval for procedure and my screen pops up, how do I know that race doesn't show up and that person on that end says, well, maybe we won't approve this. Maybe they need to do other studies.

I am not saying that that necessarily happens, but you have to understand we live in the legacy of Tuskegee. African Americans in this country, and I think justifiably so, believe that discrimination exists and it exists in the health care system and that if their race is known, that there is a potential that decisions will be made by the health care provider but also by the insurance company.

You can't hide it from the health care provider, but do you have to give it to the insurance company and that is what the 837 does.

DR. AMARO: I wanted to say something about that because you and I have been in a dialogue about this. I really differ with that point of view. I think that there is lots of evidence that shows that discrimination takes place at the level of judgment the clinicians make differentially where there is -- around providing these kind of procedures or the kind of health information they provide to women at prenatal care visits and, you know, down the line.

So, the kind of concern that you have is the one exactly that the literature has demonstrated already occurs between provider and that is where it is happening. The problem is that we don't always have the data, you know, to be able to demonstrate it on broader levels.

So, it is already happening and it has been documented and I think people here have cited that literature. So, my concern is along the lines of what Dr. Krieger and others have testified, that by not having the data, we have created inability to track this issue, which is really, I think, a major concern because it is one of our responsibilities.

So, if we don't put it there, then I guess I would want to know, well then, where do we put it because we have to find a way to put it somewhere where it can really be used. So, what is your suggestion about where it should be put?

DR. LUMPKIN: My suggestion was that you create a health data organization messaging standard in HL7, just like we do with immunization records in the State of Illinois. Then when the provider fills out an immunization form, they communicate directly with us, not with the insurance company. That creates a different bill. But they only have to enter the data once because their information system will split the data into the two different locations. The record that is seen is different, depending upon the user.

So, I think our challenge is -- and I am not saying that there is a currently existing solution -- our challenge is is to use the technology that we are moving forward to resolve the problem and I think that there are solutions that ought to be available.

DR. AMARO: I guess then as a good question for the panel is what the impact of that kind of solution would be if, you know, we see that it is feasible.

DR. CARRILLO: Just to add, the CAHPS survey, which basically measures satisfaction and also some utilization issues, includes race, ethnicity and language and this is something that primarily is sort of focused on the managed care industry, but, you know, it might also be something that is feasible for a more broader factor of the health care industry because it gives you the satisfaction. It gives you other aspects of the care that are important.

MS. COLTIN: I agree with Hortensia and I am concerned that this information isn't being collected now and we are seeing discrimination. So, one of the problems that is that we haven't been able to document the level at which it is occurring and not knowing whether those decisions are being made by the provider or are being made by someone who is approving or disapproving a provider's decision. We don't know that because we don't have that kind of information right now.

So, I think that is something to be concerned about and to try to figure out a way to track.

I think in terms of the issue of cost and burden of data collection, I mean, as Dr. Fowles was describing the data collection mechanism at Health System Minnesota, it sounds like this is collected one time, inpatient registration. Then the issue becomes the one time cost to program the system to put that piece of information into an 837 each time a service is delivered and a bill is generated.

Again, we are talking a one-time cost there. We do it for date of birth. We do it for lots of other things that we collect about people one time and we put them on the bill. Why put date of birth and not put race or ethnicity if they are collected at the same way, at the same point and the mechanism for putting it on the bill is the same.

So, I think that argument is one we have to look very hard at in terms of the burden and the cost issue. It doesn't seem to me like that is really the strongest case. I think the concern I have about it being only on a bill is that we would have a denominator problem because we only get bills for people who use services. We don't know then whether, in fact, the people who are not using services, in fact, are being discriminated against and being denied access.

So, we really need to also think about collecting it on the enrollment form so that we have it for the entire covered population and can begin to identify differences or disparities in access to care. So, we need it for the denominator as well as the numerator. But I think, again, it can be collected one time.

DR. AMARO: Kathy, just to sort of pick up on your point, Schulman said he did show that it was -- you know, on the issue of whether there is sort of additional at the level of denial of certain procedures of the insurance. Maybe people can comment on that, but that study did show that it was clinicians, you know, at individual levels with similar fictitious patients presenting with the same exact clinical data and only differing on race, ethnicity and gender, that they made different decisions simply based on that.

So, the strongest evidence we have is that it is happening at that level.

MR. CHENG: If I can comment on what you just said, Kathryn, as it relates to cost. Yes, it is a one-time cost for the transaction. Now, how far do you want to go after that? We took a conscientious decision that we would go further to sensitize the staff because of the cultural differences of our patient population.

When you sat there like this as you were speaking, I am a Buddhist. To me, you were praying to Buddha, even though you are not or I don't know if you are. Okay? But if I was a devout Buddhist, I would be offended by you doing that. Okay?

We had to train our staff. When you ask the patient questions about their race, the best thing to do, hands up on the table, sit straight, look them in the race. Don't show any expression. Just ask the question. If they answer, fine. Don't repeat it. You know, are you sure?

That happens. People do that. And just continue with the next question. That was a hell of a cost. We have got 300 employees. We have got over 300 physicians and each one had to be sensitized. If it wasn't a pain at the time of registration, the next area would be when the physician examines the patient, who you know as well as I do when they chart a single, white females, married, black male, whatever, it is there. So, we can take it from there, but we had to sensitize the physicians and the nursing staff how to ask that.

It is cost. So, yes, you are right, it is collected once at the registration area. Is it accurate? What happens? You constantly update.

DR. LUMPKIN: I am going to go to Bob, who had his hand next and then to Kepa. Then we are going to have to wrap up because it is 12:09. We have another panel at 1 o'clock and we have real tight schedule.

So, Bob.

MR. GELLMAN: I don't have any views on the question that we are dealing with here yet, but I want to make a broader point and echo something Clem said. Some of what I have heard here today is making the point that more data enables more research and we can answer more questions. That is certainly true.

However, that attitude, that point of view is what produced Oasis and I don't think that people in the health data establishment have learned the lesson of Oasis, that people will say "no" at some point and the whole thing will come crashing down and people will not allow data -- and I don't know where those points are and I don't know how to draw the line, but I think you have to be very sensitive to that because more data for more purposes will not be tolerated.

MR. ZUBELDIA: I am Kepa Zubeldia. You may not see it, but I am wearing my X12 hat.

First, I would like to address the question about date of birth. Even though it is a constant, it is used by a lot of the payers for two reasons. One is as a password. If you get an ID number and date of birth doesn't match, you suspend that claim. Don't pay on it. There is a potential for fraud. Somebody made up an ID number.

In other cases it is used to discriminate among the dependents that share a common ID number. Sometimes the insured has an ID number and all the dependents have the same ID number. The date of birth is used to discriminate which one we are talking about. So, that is why it is going on every single claim even though it is a constant.

But the reason why I came up to speak earlier on was the kitchen sink syndrome. Way back in 1992, when I was writing with HCFA the first implementation guide to the 837, we saw the first effects of the kitchen sink. Everybody saw the claim as the ideal place to get information and they said, oh, I want that but I want this little additional element. Then I will take the claim with my addition. And everybody keep throwing things into the kitchen sink.

The claim grew totally out of hand. I quit that group and moved on to something else because it was totally out of hand. Now, with HIPAA we have full implementation guides with claims. We have a pharmacy claim and that is the PEP(?) and somebody could argue that you need race and ethnicity at the pharmacy level to see if they are taking beta blockers and they are black and things like that.

Then we have three claims for hospital, medical and dental, but it is all the 837 with three different implementation guides. I agree with Dr. Lumpkin that the claim in one of those four implementation guides is not the vehicle for race and ethnicity and if the health community needs this and they need socioeconomic status and language and accent, they could put in another implementation guide that is specific for that. And we could have a fourth implementation guide of the 837 or HL7 or whatever appropriate form it is as a standard, maybe even under HIPAA standards and that conveys this information to the public health community.

For instance, I don't see much use in having the cost of the office visit or the prices that the hospital is charging in such a transaction. The could isolate the two areas of payment and insurance and public health reporting. In fact, this new implementation guide of the 837 -- I am just throwing something out -- could be used as discharge summary reporting, thoroughly unrelated to the claim. It uses the same transaction for a different purpose under a different implementation guide.

It is technically feasible. The issue that Dr. McDonald brought up is how do you get it. What motivates them to send it to you? That is the attractiveness of the claim since they have to file a claim to get paid, there is no question that if you put it in the claim, you are going to get it.

Now, you may not get it right. You may just have something to fill in the block, but you will get something and if you put it as a separate implementation guide, then what incentives do you have to provide in order to get it. That may be another concept.

DR. LUMPKIN: I would like to thank the panel and I wish we could continue this discussion. I will say that there is only one thing that I started off this panel with, which I am absolutely convinced is true, having heard the panel, and that this issue is much more complicated than we thought it was when we first started.

Obviously, this is something we are going to struggle with. There are a number of alternatives, which I think we need to review. I don't think that anyone should assume anything is thrown out at this point. Obviously, we will continue to look at modifications to the 837.

We will look at alternative implementation guides. We will look at separate messaging, but I suspect that the one thing that I did hear consistently throughout is that this is an important issue and that it is important for us to be able to address and monitor the extent of which there are variances and access and quality to the care that are related to race, ethnicity and socioeconomic status.

So, we will recess at this time.

[Whereupon, at 12:15 p.m., the meeting was recessed, to reconvene at 1:00 p.m., the same afternoon, Thursday, June 24, 1999.]

A F T E R N O O N S E S S I O N [1:04 p.m.]

DR. LUMPKIN: It is time for us to reconvene having arrived, a quorum being present.

We have a panel, but before we get started on the panel, we do have one action item to take.

MS. WARD: To be able to ask for a vote on our insular areas recommendations, I need to know whether people were able to read that part of the report and feel comfortable in endorsing those recommendations.

DR. LUMPKIN: There is a motion to adopt the report that was submitted yesterday. And you are submitting an amendment.

MS. WARD: And an additional amendment --

DR. LUMPKIN: Is there a second?

[The motion was duly seconded.]

Now your amendment.

MS. WARD: We are adding an additional recommendation that came out of our work group yesterday responding to the questions and discussions at the full committee. Unfortunately, staff has been really busy since we finished at 5:30 last night. So, I have to read this out loud to you, but it is not that complicated.

The additional recommendation would read, "Request that the Data Council identify a single point of accountability and contact relating to health statistics, health information systems and data needs for the Pacific insular areas, Puerto Rico and the U.S. Virgin Islands." That would be Recommendation No. 15 added to on the others that were in the report.

DR. LUMPKIN: Does anyone need to hear that read again?

MS. WARD: "Request the Data Council identify a single point of accountability and contact relating to health statistics, health information systems and data needs for the Pacific insular areas, Puerto Rico and the U.S. Virgin Islands."

DR. LUMPKIN: Is there any objection to including that into the main motion?

Okay. The motion before us was the report that was presented yesterday, 15 recommendations -- 14 recommendations plus the 15th we just heard today. Is there any discussion on that?

[There was no response.]

All those in favor say "aye."

[There was a chorus of "ayes."]

Opposed, say "nay."

[There was no response.]

Abstentions?

Okay. It passes. Thank you.

MS. WARD: Oh, I am sorry. I wanted to add, which I didn't do yesterday -- fortunately, Dale walked in so he can hear this -- I want to thank him for his extensive work for this report.

[Applause.]

And the former editor, who was Joan Turlock, who then gave her stuff to Dale.

DR. LUMPKIN: Great. Thank you.

I would like to thank your committee for their hard work on this particular issue and for raising an issue, I think, which clearly needed to be identified.

Agenda Item: Smart Card Development

We are going to start off with a panel on Smart Card development. I would like to thank the panel to introduce themselves.

MR. SCHWARTZ: Ari Schwartz with the Center for Democracy and Technology, policy analyst. I work on privacy issues.

DR. KINGSLAND: Our initial speaker will be Dr. Bettina Experton. She is an M.D., specialized in public health and preventive medicine and oncology. She is on the faculty of the University of California-San Diego School of Medicine. She is a health services researcher and president of Humetrix, Incorporated, a company specializing in development of Smart Card Internet health care applications.

DR. LUMPKIN: Did you introduce yourself?

DR. KINGSLAND: I am Lawrence Kingsland. I am assistant director for applied informatics at the National Library of Medicine. I have been a researcher in medical informatics concentrating in intelligent retrieval systems and artificial intelligence applications in medicine for about 30 years and am now among other things, the National Library of Medicine's liaison for Smart Card projects.

MR. MALONEY: My name is Dan Maloney. I work for the Department of Veterans Affairs within the VHACIO field office and director of emerging technologies and I work in the area of Web technologies, security PKI, advanced technologies and also health care Smart Cards.

We have also done some other things, like originate the VA Web server back in September of 1994 and we also sponsor the Internet real audio broadcasting of this session. So, it is kind of fun to be here and actually have it run back there at the field.

I am also the U.S. representative on a G8 Health Care Data Card Word Group.

The first speaker will be Bettina Experton.

DR. EXPERTON: Thank you, Dan.

Dan, Larry and I are going to tell you about Smart Card technology and how it can meet critical data needs. I will respond in terms of the privacy, security issues. In fact, Smart Cards do bring a solution in terms of security and privacy issue of health care data access.

Going back to myself, why I am here today, first of all, I thank your committee for inviting us and bringing information about technology to you. I am a health researcher and like you very much aware of the gaps we have in collecting critical health care information. As a physician, the problem we have in accessing this essential core clinical data sets for patient care, also in collecting data for population based analysis and originally from France, as you can tell by my accent, I guess I was exposed to technology, which is widely adopted in Europe, but I felt that the technology can even better meet all needs in the U.S. in terms of risking the disparity of system, the multiplicity of the player, the disjointed characteristics of our system.

The device itself is an interactive portable tool for consumer access, provider access to -- indeed, access is critical as to data sets and give consumers, the user, control over their personal health information. Also, and very importantly, that device, which can store data and this critical data set, it will store a linkage device among those various points of care and point of service, which are not always networked and these Smart Cards is not only a data carrier, but it is also a linkage device and in a way that Dan will describe more in detail, settings of the system whether on the provider side or on the public health department side.

So, what are Smart Cards? The type of cards we are talking about for health care application are microprocessor. They are cards with a microprocessor chip in them and they are small plastic cards embedded with that chip. They are inexpensive and they provide immediate and secure access to what we should determine be the essential medical information.

It is also provides meaningful on-line access to additional health data. It also provides a means to secure transfer of those data on the network. Some of you are familiar with the technology and others may not be as familiar with the technology.

I am going to pass along a sample of the French card called Vitale, which is now an ID card, but which is going to roll out in a couple of years as a really portable CPR. So, this is a green card, the French Vitale card.

This is a card, which is distributed right now by Wrighthead(?) Pharmacies. It is only what we call an electronic purse. It contain value in the chip. It is not health care card yet, but that is one step in the right direction, indeed, building an infrastructure in the country to have this technology available in the hands of consumers.

Also Smart Cards don't work alone. They need, indeed, to be read and written on by interfitting to a P.C., computer system, with a Smart Card reader and the Smart Card readers can go into the floppy drive directly, such as this one. These Smart Card readers and the cards are currently issued on a pilot basis by Bank of America to their corporate clients. So, secure, authenticated electronic banking on the Internet.

Another example of how Smart Card can be used easily at a low cost is using laptop computers where the Smart Card readers is in the PC MCI slot, which will be right here and the reader is this device with one hand and the cards slide right above it and can be read and written using the keyboard of the computer.

How Smart Cards differ from mag stripe cards or laser cards, Smart Card can indeed protect confidentiality. They can carry encryption function through the microprocessor, still encrypted data. As we will see, typically in a health care application of the Smart Card, the data are stored in separate files. It is a file structure system and those fields are partitioned with restricted access to the various fields.

So, a clinician and provider will have ready access to write and read clinical side; whereas, clerical personnel will have no access, for instance, to the writing capability of the clinical side, but access to the patient identification and insurance information about that patient.

This capability is not found with mag stripe and it is more limited. Smart Cards last five to ten years, which is more than the mag stripe, only two years. The benefit of Smart Card is that you, indeed, can write and edit the content of the card, which is not as easy and very costly in the case of mag stripe card. This is a possibility with laser card.

Smart Cards are also really smart in terms of their ability to upload and download information to a network and, again, that should be done for health care in a secure environment. That is what the Smart Card can do as well, but Dan will cover that subject.

This is not the capability of the mag stripe card. In terms of security and fraud protection, you have data lock features with Smart Cards. If the card is tampered, the data will be erased. There is further security on the chip. You can have a pink code access function. You can add biometrics markers. You can add photo ID in the chip to restrict access to the data content of the card.

The mag stripe card is easily counterfeited and once you swipe the card, you access all the information. Laser card are not as secure as Smart Card. In fact, to make a laser card secure, you need to add a chip on the card and that has become what the industry call an ibrit(?) card.

In terms of memory capacity, the optical laser card as immense amount of memory, up to 8 megabytes. Indeed, you can store imagers, x-rays and so forth, but the Smart Card today has a memory capacity of 64 kilobytes, which translate into pages of text, depending, indeed, how the data is organized and tabulated, 30 to 60 pages of text in a Smart Card.

Mag stripe card on the other hand can only contain 226 bytes of data. That is about six lines of text. Again, a Smart Card allow you to read and write and change its functions and additional functions can be downloaded from a network into the card without having to reissue the card. The mag stripe doesn't have that feature and once the function of that card is set, cannot be changed.

The optical is a card, indeed, have that same feature as the Smart Card. In terms of infrastructure, a Smart Card today can be deployed at a very low cost. One of the readers, in fact, I did pass around, which is right in back of my laptop here, this is a reader which can be used for home use, for individual to access the card contents right from their home as the reader is connected to the serial port of the P.C.

This device is a $10 device. The floppy disk reader I passed around is a $40 and a provider -- professional reader, which will be a heavier device and this one a more sturdy device cost around 40 or 50 dollars. So, the cost of infrastructure in terms of putting readers in the hand of consumers, for them to directly access information they want to access or for providers to be equipped at every point of service to update the information contained on the card, to upload that information to a network in a centralized database, this can be done at a low cost today. There has been a change the last few years.

The mag stripe readers are deployed throughout the country, but they are still costly. They cost about $300 apiece. They are back here. That is the case as well of the optical laser card readers, which are larger device and very costly, two to three thousand dollars apiece, again, versus as low as $10 a Smart Card reader.

So, Dan, I think I am going to give you the microphone to cover the issue of security, privacy protection of Smart Card technology.

MR. MALONEY: Thank you, Bettina.

There is so much to say and there is so little time. As you know, on all of your speakers that come before you, everybody has very important things to say and there is very little time.

I am going to cover a number of different areas and my boss, Rob Kolodner, recommended that I pick out three or four things and try to convey them to you in the beginning and then in the end. So, here is what I would like to emphasize.

A Smart Card system works with the network. It can also work alone, but it works with the network. A Smart Card can be a tamper proof or is a tamper proof carrier of data for communication and also a carrier of keys for authentication.

The third thing is as the network improves and we have all watched the Web explode, as the network improves, we really need a better way to authenticate, identify individuals. That is going to be key in the whole process of enabling the user to do electronic service delivery to interact with their own personal records, et cetera.

A Smart Card can be a carrier of those keys and can enable that -- can empower the user. It is also a mechanism of protecting privacy by increasing that security and making sure that you have something fairly high level in order to gain access to that information.

Also, Smart Cards are gaining in their support, both by organizations and by vendors. As I whiz through this presentation, you will see those four points coming out periodically, and I will try to emphasize them again as I get to the end.

The other thing is in your agenda there is a paper that I published at Cardtech(?) Securitech(?), which gives you more information so you can read through it at your leisure. The slides also have a fair amount of information on them.

So, basically we have this information. We have this information system. We have the multiple players in this information system. We have the patient. We have the health care provider and then we have the medical system and then behind them we have the network. And a card can play in this environment working with the network, it can carry data. As the network improves as we have more things out on the network and we make them available, we really do need to have that better protection mechanism. We can also begin to change the location of the data.

When your network isn't so good or when you go to places like an emergency room that is not on your network, you can have a card. You can walk in with it. You can have your critical medical information and convey it there.

As we work out all of the policy issues so that you can have interconnectivity from anywhere in the world in a secure and private manner and we can change the location of that data so that it fits into this continuum of evolving and improving networks. The role of a card is it can be first of all a simple visual identifier, just as it is now, but it can also be the secure mechanism or the tamper proof mechanism of carrying keys and carrying data.

It can enable the portability of these credentials, PKI private keys, things like that. It can isolate the security related components. There are Smart Cards now which do the mathematical calculations. So, at some point in time in our future -- and we are building towards that future -- you want to be able to go to the mall. You want to be able to go to your library. You want to go to your home PC and you want to be able to communicate to important systems. You want to be able to change your addresses, et cetera. You want to access your medical record.

If you have to turn your private key over to the kiosk, that is not all that good. So, a Smart Card can do the calculations within that card so you never have to turn your credentials over. This is very important in a medical environment. It can also play a role if we wish it to -- a Smart Card can play a role in electronic payment by carrying purses, by linking with credit systems.

How many of you currently use a Smart Card in your daily life? Anybody? Does anybody use one of these, which is -- which works on our metro system. It is called a Smart Trip. What you do with this thing is you walk in and you basically just touch it near the reader and it basically carries a certain amount of money and it extracts money from the card and then you can go back and recharge it.

So, these things are coming. Some of us use them in our daily lives. Some of us don't. This happens to be a Smart Card, which is my key to access a network certification authority for the VA's PKI certification authority, which is run by a company called Verasi(?). It is out there as PKI-based. You need this key in order to access that information on a network in order for someone to authenticate somebody to authorize a transaction out there.

So, these things are coming into our lives and they are good and they are empowering. The big issue as we all see, as the network improves, we need a better way of identifying and authenticating where the transaction is coming from. We need a better way of identifying who we are talking to. If you are going to be talking with your health care provider across a network, you want a little bit more than a password in order to authenticate who they are.

This is the scenario. There are major changes. This is really the heard of the whole issue. The network is getting better. Communications are getting better. We all want to communicate. We all want to have access. We all want to be empowered. We all want to have access to our health care information, but we want it done in a way that is private. We want it done in a way that is secure.

We need to raise the bar. We need to get beyond passwords. We need to have some kind of tokens. PKI is a megatrend. Asymmetric keys is a big issues that is getting a lot of support from a lot of different areas. And a Smart Card is not a bad way of carrying around those keys. You can safely carry the keys on the Smart Card. You can also have multiple other applications on there and you can do things like carry important information like that emergency data.

PKI, you all in your deliberations have probably touched on it and are fairly knowledgeable. We wanted to mention just a couple of highlight things on this area. It is playing a big part in our future systems, secure electronic mail, a consistent standardized way of accessing information systems, is involved in encryption. It is also involved in digital signatures.

The public key is basically wrapped up in something called the certificate signed by the certification authority. The nice thing is it can be put out on a directory and it can be made available publicly. You trust that certification authority because it is an authorized and known entity that is audited, et cetera.

So, you don't have to get every other individual's private key. You can go to the directory and pull it down. That is one of the reasons why PKI systems are so scalable. You really need to have your private key and you have to protect it. The reason for that is that in the encryption process, what happens is you have medical data. The public key can encrypt that data and what you end up either communicating across the network in a message or whatever is basically unreadable.

It is unencrypted by the other key. So that if I were to be sending something to Rob, I could in an oversimplified way, but I would go get his public key. I would take what I am going to send to him and I would encrypt it with his public key that I could get off the directory. It would go across the network encrypted. When he got it, he would take his private key and he is the only guy that has got his private key. He would take his private key and could open it up, which means that he is really the only person that can read that message that I sent to him.

You can also digitally -- using PKI technology you can also digitally sign documents so that you have a high degree of assurance that that document hasn't been changed. So, it could be a summary of a procedure or something like that. You all have made recommendations related to digital signatures. So, I know that you support this general area.

You have the document. You have the hash. You use the key, the private key to sign it and you end up with a document and also with the hash. And at the other end, they take the public key. They run the same digest on the document and they feed the digest and the public key in there and they basically verify the signature, verifying that not none of the document has been changed.

These are basically key pieces in this much larger schema that we are trying to create, which allows open communications, but also private and secure communications.

So, what has happened in Smart Cards recently? These are the highlights. There has been large activity in the French area and it is interesting, a person with a French accent, who happens to be American, talks about activities in the United States and I will talk a little bit about things going on in Europe.

The French activity is within the past year has been very large. There is also some -- there are other projects, which I will talk to as we go along. The French Vitale card in the last year, between May of 1998 and May of 1999, the French Government has distributed 42 million family insurance cards. Their role is administrative and insurance so that as you go from health care provider to health care provider, you have the consistent data being given to each one of them. You save the time. Nobody has to type it, et cetera, et cetera. It is a great example of administrative simplification, consistency as you move around.

They distributed 5 million cards a month,but they did all of this in a year. Another phase of the activity that they have been -- that has been going on in France is the health care provider card. They have been issuing provider cards with PKI keys on them. This is an order to digitally sign various billing documents that are sent to the payer.

They have also instituted what they call a health care network to carry these transactions. So, at this point in time it is primarily focused around administrative simplification, signed transactions going to billing parties, et cetera, and they also want to use the same key as they expand it to access information both on a card, which would be the Vitale 2, the individual card that comes up, but also to access information that is on the network. If you are a health care provider, you are accessing this information. You have the key, which is carried on your card.

Fifty thousand of those have been distributed up through May of 1999. They are going to be moving on to -- most of those have been to physicians and they are going to be moving on to other health care providers. There is another project which has been going on for five years, which is funded by the European Union. This is an example of inter-country interoperability. There are pilot sites in nine different countries and they show both the transportability of information and also using codes that you all are working with also, using codes that shows that the contents are translated into the language of the country.

So, if you happen to be a Frenchmen and you go to Spain and you go to a health care provider, in goes your card and the majority of the coded information comes up in the local language. The application, I think, of something like that here would be if there is a Spanish or a French or whatever person, who is visiting the country, who is in the country, and you want to verify their information, they could look at it in their native language as opposed to in a language of a local country.

There are a hundred thousand cards in that demonstration. Another sort of world event in Germany, this what is called the Memory Card, which does not have a computer to protect the contents, but it does store information in a digital manner on the card. Germans in 1994 and 1995 distributed 80 million cards and the concept again was administrative simplification, consistent information as you go from place to place.

As a member of the G8 Healthcare Data Card Project, what we are trying to do is work out an international agreement on a standard emergency card and also an international professional card. There are lots of different levels of interoperability, just as you all have been dealing with. There is technical interoperability as well as different levels of standardization related to nomenclature, et cetera. And they have some -- we have a number of different recommendations and they are posted up on the Web site that is on the front of my presentation.

Also just beginning right now, this next generation, which is the NetLink project and there are four different countries involved in that, France, Germany, Italy and the province of Quebec. The concept there is to try to take idea of health care provider cards and keys identifying those health care providers and allowing them to access information on cards and also across the network. So, they are really trying to apply what it is that in theory they have been supporting.

The VA is doing a number of different things moving into this area or investigating. I am sure you have heard about the VA and its medical organization and about its hospital information system that is supported were fairly large organization. We are doing a number of things.

We upgraded from a dumb card to a slightly better card, which has everything but a Smart Card, back in 1997. It was a fairly large initiative. We weren't quite ready for Smart Cards at that point in time, but it did have -- it was done in a fairly short time and it has been accepted rather well.

There are still other things we want to do and we want to have enhancements and these are the things we are working on now. We are looking at a small scale test in home health care environment to use as a mechanism to carry important data, have it always available, have it available in emergency rooms. We are looking at new technologies from the standpoint of one of the -- Microsoft recently came out and announced support back in October of 1998, support -- they are coming out with an operating system for Smart Cards, which is a fairly large event. IBM also is selling portables and their portables have -- are Smart Card enabled. So, you can secure the contents of what is in your portable.

But these large vendors supporting the concept of Smart Cards, supporting the concepts of using these for authentication mechanisms is basically to make working with these devices less expensive and also much easier to fit into health care information systems.

We are piloting secure access from Internet. We are using one of the small -- fits into floppy disk readers that Bettina showed earlier. We are working on a PKI pilot. We have some, about 50,000 cards in an electronic purse test, which has been accepted well, both by the patients and by the staff in two areas. One is in the Bronx and one is in Tampa. The other initiative is we are trying to build a larger initiative with the Department of Defense and also working with the GSA Smart Card Initiative to try to explore wider areas.

So, did I get some of my points across, that it works with a network, can carry data and can carry keys. And as the network improves, this is really an enabling mechanism to improve privacy and security and that the cards are being supported more by some large organizations, some countries in health care environment and also some vendors.

There is also activity in the United States and Larry is going to -- Larry Kingsland is going to touch on a couple of those that are being supported by the National Library of Medicine.

DR. KINGSLAND: I am going to speak very briefly about a pair of health card project examples in the U.S., two pilot projects in particular. We have chosen them because they were designated examples of the G8 pilot projects in this country and they are sponsored or co-sponsored by the National Library of Medicine.

Our director, Dr. Donald Lenburg, was appointed the health care representative to the G8 international group by Dr. Shalala.

I am going to speak briefly then about the Health Passport Project by the Western Governors Association and a project titled Secure Collaboration Technology for Rural Clinical Telemedicine by West Virginia University.

For those who might not know, the Western Governors Association is an independent non-partisan organization of governors from 18 western states, two Pacific flag territories and one commonwealth and the Health Passport Project is one of a number of the initiatives of that group.

It is the largest health care demonstration in the U.S. for Smart Cards. There will be 25,000 cards released in three communities in the west, Bismarck , North Dakota, Cheyenne Wyoming and in the fall, Reno, Nevada. The idea is that the Health Passport is a citizen-based portable medical record, that people will use the cards to give up-to-date information to their health care providers, including physicians, nurses, nutritionists and early childhood educators.

The demonstration project is designed to simplify and improve access to health services for multiple groups, resulting in healthier moms and kids. There are a number of state and federal programs involved in this project, including Medicaid, immunization through the CDC, the Women, Infants and Children Program, Head Start and Maternal and Child Health Services.

There are other partners also closely involved, private partners, including primary health care providers, specific groups of practicing clinicians and pediatricians, the grocery stores to handle the nutrition benefits, the food stamps and note that in this context, the infrastructure as ever becomes really important. We need to have readers in the right places and also interfaces of the card applications to Legacy Systems in the childhood immunization, Women, Infants, Children Program, Head Start and Blue Cross/Blue Shield. This group is unusual in having created public kiosks to put in places like the Wal-Mart Store, for instance, where some of their clients might be going so that the persons holding the cards can check and print out the information on the card, take it to a physician if something needs changing or has changed. Interesting approach.

There are four goals to this project, four primary goals. The first is to reduce health care costs, both in terms of time and of money for patients and for the providers by having accurate information where it is needed, when it is needed. The second is to improve the quality of care by giving patients better access to the care for which they are eligible and reducing gaps and duplication in their records so they get the right care.

The third goal is to empower the client, give individuals more control over their information so they can take more responsibility for their health and the health of their families.

The fourth is to improve what might be called customer satisfaction with public health services. The rollout of the project is occurring as we speak, in June 1999, after several years of preparation and development. It is really quite a challenging project with many players. There are 17 public and private public health services involved in this project. There are 502 data elements on the card. Every element is there because it was used by one or more of the agencies that was going to work with the clients in these cards. Nothing is there because some one group wanted it.

So, there really is quite a lot of information there and with luck it should make a difference in improving the efficiency of the transactions these folks are involved with.

There has been a rigorous independent evaluation planned from the outset of the project. This is underway. The outcome, obviously, is not yet clear at this point, but it looks promising.

The second project I am going to tell you again very briefly about is done by a group at West Virginia university. The goal in this project is to produce a secure collaboration technology for rural clinical telemedicine using Web-based access to patient information with encrypted transmission over the Internet and the Smart Card as a key element of the project.

As Dan mentioned and Bettina mentioned, the card is used for identification, authentication and data storage. There are a number of health care delivery partners involved in this project. The largest perhaps is called Valley Health Systems, involving 15 community health centers and public health programs in southern West Virginia.

Another is the Cabell Huntington Hospital, a 300 bed tertiary care center that is affiliated with the Marshall University School of Medicine, St. Mary's Hospital in Huntington, West Virginia and several other cooperating groups.

The operating hypotheses for this project are that in rural areas where specialized medical care may be unavailable, telemedicine can have an especially large impact on the quality and speed of patient care and the current techniques and data security can make secure telemedicine viable over public networks.

Note, as Dan mentioned, that the French declared that the Internet was not sufficiently private and not sufficiently secure. So, they built their own national health network. This group is trying to use current software and security technologies to see whether we can do this reliably over the Internet.

Major elements of the project include providing authenticated health care providers with secure access to health care information, facilitating the authorized flow of health care information between providers and systems and supporting secure teleconsultations by health care providers for patient treatment. If nobody knows you are a dog, then probably nobody knows you are the intensivist at the other end, wanting to see your ICU patient information from somewhere. So, this really is an important key.

As in Europe, two types of cards are used in this project. There are PIN-protected patient cards that include among other things data based on the European Union G7 health card interoperability spec and there are provider cards for the exchange and validation of digital certificate during telemedicine consults.

With the patient data card, role-based access restricts the information that can be viewed or modified by the patient himself or herself and by differing categories of health professionals. So, for instance, the health care business staff may be able to get only to your demographic information and perhaps your insurance ID numbers.

The medical information on the card would be available again differentially to physicians, nurses or other practitioners as appropriate.

The group in West Virginia is testing three application scenarios. One is to provide the intensivist with secure, remote access to ICU electronic patient data with the physicians authenticating each other using their provider cards.

The second is to improve delivery of health care through mid-level practitioners, such as physician assistants and nurse practitioners. These folks perform their treatments, make their reports add to the record and then physician using his card for authentication can sign off on what the mid-level practitioners have done.

The third scenario and the most recent just getting underway is to facilitate secure telemedicine access by the nurses of the Mon County Health Department for home care of a small group of diabetes patients, who are receiving frequent care or are in some cases making frequent visits to emergency rooms. The group has further card-related plans involving the addition of childhood immunization tracking components to the patient health card for use with their pediatric patients and adding an HL7 capability to the applications around the patient Smart Card to facilitate the card integration and deployment at health care sites having Legacy applications.

So, the current status then is that the end to end architecture is fully realized. There are at this point several hundred cards deployed now. It is just getting underway in terms of deployment. There will be a thousand cards in this pilot by the end of the summer. And, again, it is looking promising thus far.

I would leave you with one conclusion. Though slow to take hold in this country, I believe the smart cards are clearly worth exploring as a means of helping both to secure patient information and to facilitate its accessibility when and where needed.

That concludes the presentation.

DR. LUMPKIN: Thank you.

Ari.

MR. SCHWARTZ: I actually have a slight technology problem here, which I should have -- I just realized how I could have done this differently, but I didn't do it that way. So, I will have to wait while this loads up.

Let me just start up. My presentation is titled "Smart Cards at the Crossroads," because I take a slightly different view than Dan and Larry and actually Bettina mentioned it as well, which is that Smart Cards can enable users to have more control over their data, but that is not necessarily the case. They could also be designed to give users less control over their data, things about them, and also on the other end as well.

What I am going to be talking about is both the pitfalls and the advantages, the potential Smart Card policy and Smart Card implementations. To start with, I am going to go through different types of authentication because we have been hearing about the different uses, things that have been used -- that Smart Cards are being used for with the potential out there for Smart Card use.

I am just going to give three examples. The first one is identity, which is the one that Dan into the most detail and there is good reason for it. It is because, as Dan said, it is the one that is most needed in the network world today.

When you think about identity, you can think about, you know, a birth certificate, could be a name, but it could go further. Could be a biometric. Could be a PIN number, some other -- something that someone actually has to validate, something that is stored on the card. That is what is important.

These can be separated out on a Smart Card as we heard. You can think of them as different applications within the Smart Card. So, there are subsets as well as just identity. The next one that you can think of is just eligibility.

The eligibility, the various keys that we have that get us in and out of places, numbers that aren't directly connected to more information about us. Value, the metro card example, something that is just cash. And all of these are going to be tied together in the Smart Card, but they can be separated out in different applications. That is really what we are discussing.

So, now I want to talk about the side that we didn't see over here, which is the worst -- what are the pitfalls. What are we concerned about? The first one is the greater collection of information, the idea that all of this stuff will actually be stored on a single card and not only will this information be stored on the card, but all the transactions that we use with the card will be stored on the card as well. So, all this information gets placed into the card. Yes, it can be fireballed out and be separated out by different applications, but it is still all within one card now. It is still all within one piece of plastic basically, which is something that we haven't had in the past. And that is the centralization aspect.

Now, along with those two comes the idea that if there is a biometric on the card, if there is a stronger form of identification, why don't we just use that stronger form of identification. It is already on the card. Why doesn't everything -- so, you see this kind of devolving down to the greatest kind of identification.

But that is not what we have in our world today. We have a wallet full of different kinds of identification and that is the kind of policy decisions that are going to have to be made. While the Smart Card can protect from something like that happening, that won't be the case if we decide that that is the best route to take.

Lastly -- these are in the order of paranoia, by the way -- lastly is the means for new social controls. You can think of it as you -- that if you haven't paid your parking ticket, which you are paying the cash out of your card and you don't have enough money for it on the cash on the card, now, can you not use the card to go get health care? Can you not use the card to do other essential things? Maybe someone wants to use the example of parking ticket and opening up your car door.

They were built for one purpose, used for other purposes down the road and that is really what we are talking about with all three of these things, but, you know, the last paranoid example over there is what people are going to be thinking about and I will get back to that more at the end.

I need to think of this as two separate models is the analogy that we have been using; a single key, which has the convenience of opening up every door in our lives, the convenience of being able to use a biometric or being able to use a single form of identity for everywhere we go. But the disadvantage is when we lose that, when it gets corrupted, we have lost the -- we have to change all the locks in our lives or you can think of it as keys on a key ring. This is the way that most of the Smart Card implementers, people that are very -- that are bullish on Smart Cards and for good reason in a lot of cases is that you can think of it more as keys on a key ring.

We have different keys we use for different purposes. They allow for different kind of authentication. They do give the user more control and it reduces the ability of identity theft down the line, but we have to make sure that that is put in on a policy level, as well as at an implementation level.

Here I have a list of questions for implementers. I don't want to take too much time because I know we are running -- we are already over. But the question is, you know, what do we really need to do each application? How can we limit the information? What has changed over time to make these applications different? Is there a way to segment these out far enough so that they won't be used for disparate purposes further down the line, technologically speaking.

Then, of course, we have those same kinds of protections -- and I have these slides if anyone wants these slides afterwards, you can ask me for them or maybe I will put them on a Web site

The questions for policy makers, which is, of course, the most important for you folks is what are the different information handling practices of the card itself, what are the policies around that? What is the ability of the Smart Card company or vendor to warehouse for a third party, a major concern up there on Capitol Hill right now. What are people doing with third party information?

Even internally, what happens if a company that vends this card also has separate practice on the side, has merged with another company? What happens to the data in a merger, et cetera? Can the card be used in the way that I was talking about earlier, unintended ways?

Lastly, and most importantly, what are the internal practices for the usage of the card? Yes, we can block out that person in billing can only read the billing records. But what if other people have free access and walk right into it where billing is? What are the kind of protections that go on in that level?

Those are issues that I know you have dealt with in the past on privacy protections within health care facilities, et cetera. Those are issues that you will have to deal with all over again in a Smart Card arena.

That is the short version.

DR. LUMPKIN: Thank you very much. I think we have seen an awful lot about Smart Cards.

We unfortunately don't have much time. I will take two questions. I have one, which was for the first presenter because it was raised by Mr. Schwartz and that is what do they do in France, in Germany, if people lose their cards?

DR. EXPERTON: Indeed, the card can be replaced. Does your question imply if the card is lost, data are lost?

DR. LUMPKIN: So, in other words, is there a database somewhere that has all the data that is on the card?

DR. EXPERTON: Yes. Also, the French are not using again the Internet but they are using a dedicated network, but I think as Dan exposed a Smart Card has to work with coordination with the network and a combination of network use is a private or public network, such as Internet with the security functions Smart Card provide and the use of Smart Card as a portable device for immediate access off line, which also permits -- making it possible to solve some of the problems we have today.

DR. LUMPKIN: Any other questions from the committee?

MS. FRAWLEY: In the case of the VA, what is going to be the cost of a card for veteran -- for dependents?

MR. MALONEY: That is always a good question and it is one of the reasons why we want to watch what is going on, as well as run some tests and pilots. Right now, the costs run around four or five dollars for a card. If you read some of the literature that some of the vendors are putting out and talking about projected increases in volume and decreasing costs, then that would be more attractive. A part of this, I think, also, at least in the earlier years, will be -- in addition to user empowerment, there is also going to be user opt in and it is similar to what the banks are doing now. If you want to do electronic banking, then you can do it. Okay. But it is up to you.

If you want to have faster entrance and exit from the metro, you can do it. The individual has to pay the five bucks for that. The user pays the five bucks. There have been surveys that have shown -- and Bettina was going to talk about some of them -- myself, as an example, I am happy to pay five bucks for that convenience. I would be happy to pay the amount of money to be able to walk into the different health care providers that I was going to for the first time and just not fill out those silly forms again and make sure that the transaction that goes back to the insurance company has the right stuff on it the first time.

So, I would be willing to pay $5 for that.

DR. EXPERTON: Did we mention the survey of the American public and their perception of Smart Card, not only about 75 percent welcome Smart Cards. They are ready to spend even $50 a year to have access to multiple application in that Smart Card in terms of carrying multiple cards and more important for our purposes, Americans rank first in terms of needed application to serve their need health care access, health care information data access.

A point I wanted to make with regard to the cost, I think you have to look at the cost also in terms of the cost return on your turn on investment. If you were to look just at one application and I didn't have time to cover those in a practical world of what can those do for ourselves, just storing medication, prescribed medication on that portable device, up-to-date medication, which is brought by the patient at any point of care, whether it is the emergency room or one of those multiple providers of chronic care at those patients' visits. If you just can use that application, we did a cost return analysis.

Given the fact that, for instance, for the elderly one-third of hospital readmission are caused by adverse drug reaction, polypharmacy issues. If we were to simply reduce by 10 percent that risk by providing immediate access to this current medication listing, we will by far recover the cost of investment in providing in the hands of patients that card and the hands of providers the tool to update it, read it and update it.

DR. LUMPKIN: Thank you. I am sorry that we have time, but our next presenter is here. Thank you very much for coming. It has been a very interesting presentation.

MR. MALONEY: Thank you for the opportunity and if additional information is needed, we are available, as well as other associations and vendors. Thanks.

DR. LUMPKIN: Thank you.

Agenda Item: Discussion with CDC Deputy Director for Science and Public Health

Our next presenter gives us an opportunity to

-- I don't think this is your first time before this committee, is it, Claire?

DR. BROOME: It actually is.

DR. LUMPKIN: Oh. Well, for the first time, we get the opportunity to bring Claire Broome, who I have worked with a number of times at the Centers for Disease Control. She is the deputy director for science and public health and has a new position soon that maybe she can also tell us about.

DR. BROOME: Thank you, John.

Even though this is my first official time here, I have certainly followed with a great deal of interest the actions of the committee over the years and appreciated what you all do to try to move us in the area of health and vital statistics.

I am actually here partly representing myself, but also partly representing Jeff Copland, our new CDC director, who sends his greetings and also his appreciation for what the committee does.

I will just say a few words about my new position because that is sort of my basis for saying that I am here representing myself. On July 1st, I will be taking up a new position as senior adviser to the director of CDC for health information systems and this is a management commitment to accelerate moving forward with integrated health information systems at CDC.

It is an activity that has been encouraged, discussed for at least ten years, but Jeff has made a major management personnel and budget commitment to actually making it happen. So, I will be working with many folks throughout CDC and outside of CDC to try to facilitate this process.

I thought it would be useful today to maybe make a few opening comments, but try to use most of the time for discussion. I believe at your last meeting you had a presentation from Denise Ku(?), who went into quite a bit of detail about some of the nuts and bolts of at least our surveillance functions. So, I thought I would do a sort of very big picture overview of statistics and health information as it relates to CDC's mission and then a few words about our new initiative and then, hopefully, leave enough time for some discussion with the committee.

I know many of you know CDC well. I think some of you are not as familiar with our agency. So, just let me say a very few words about our basic mission.

The Centers for Disease Control and Prevention, which was added to our name, to indicate our responsibility as the nation's prevention agency, and people tend to think of us as infectious disease focused, sort of disease detective stereotype, but our responsibilities actually extend very broadly. We certainly still have a lot of activities in the area of protecting the public from infectious diseases. We also have very large activities in preventing chronic diseases, promoting healthy behaviors, preventing environmental exposures, working with our state and local partners to tackle toxic exposures, working on injury prevention. That is about a fairly broad range of responsibilities.

We address those responsibilities through activities that span the spectrum from pretty basic research, applied research in risk factors, etiologies, determining new infectious agents, all the way to working with our state and local partners to actually implement public health programs.

So, we also have very substantial programmatic responsibilities we fund directly our state partners to deliver everything from breast and cervical cancer screening for underserved populations, the national immunization program, HIV/AIDS prevention. So, we have got both research responsibilities and programmatic responsibilities.

So, as you can imagine from this kind of broad overview, data is central to essentially everything we do. We need effective systems for detecting problems, for giving us early warning signals that there is an outbreak or that there may be a new infectious disease. We also need data systems, which can determine whether our programs are working. Are we reaching the populations that we need to reach?

Are we delivering services effectively? Are we having the impact that we need to have to improve the health of the American population?

In order to do that, we use a huge range of different data systems, sometimes daunting to sort of look at the number of different systems, some of which we create and some of which we use secondary data sources as available and created by other entities.

Certainly one of the centerpieces for the data that we need and use is NCHS, which is one of CDC's centers. Obviously, you are very familiar with what NCHS does and its responsibilities go well beyond CDC, but we have very much appreciated the opportunity to have NCHS affiliated with CDC, both as a source of data and also as a way of meshing NCHS surveys and statistics, with the programmatic needs that we see surfacing from our work in the public health area. I think that has led to some very productive collaborations.

The second major area of data systems are the traditional surveillance systems, which I think is what Denise focused on last time. And these traditionally have been a very imperfect tool, shall we say. John knows well what I am talking about from the state perspective. They frequently are incomplete. They are not timely. They are labor intensive.

On the other hand, they can serve incredibly important functions in alerting us to problem areas. The integration project that I will describe particularly focuses on enhancements in the surveillance systems, not just to make them do better what they currently do, but to really move them to the next step of being able to interface electronically with primary databases in the health care sector so that we will have more complete, more timely, more accurate reporting on issues of public health interest.

Then we have program management databases and I have already alluded to some of the use we make of those for assessing our programs and then a whole raft of traditional and novel secondary sources. We are certainly big users of hospital discharge data bases. We also work a lot with the criminal justice side in our injury prevention activities. We need to know something about homicides, suicides, domestic violence in the area of injury prevention. We work with the Department of Transportation and some of their databases.

So, there is a huge range of data areas. Now, it has been extremely important to us, the efforts that you all have been making and we have been making with the Department in implementing HIPAA. We see this as a major step forward in our ability to work with a number of these data sources. Clearly, we have actually committed, as Marjorie well knows, huge numbers of people to work with HCFA on the implementation team for developing the HIPAA EDI standards, both for the primary core elements and also for the claims attachments.

We are also very engaged in the privacy and confidentiality issues. We see this, as I think you do, as essential to being able to progress in what we do. We have been particularly concerned that people remember that public health uses need to be considered in, for example, the privacy legislation that has been proposed and I think we have had a fairly successful dialogue with the Department and with Congress to consider the needs, public health needs for access with appropriate confidentiality protection.

We also very much look forward to the committee's efforts on the computerized patient records. Obviously, to have health databases be truly useful, we need to move from the cottage industry era to something that more resembles a database that can be used.

So, we are very supportive of what you are doing and we are also working with the Data Council and HCFA to try to move forward in those areas.

In terms of our initiative for the future, I think we have got a very promising confluence of technology development in terms of Internet and opportunities with, hopefully, coming with distributive databases, other tools, put together with the national standards that should be growing through the HIPAA process.

And then maybe the biggest component is we have actually received some substantial resources to try to motivate change in the system. In 1999, those have come primarily to assist in preparedness against the bioterrorist episodes and we have felt that key to public health system response to a bioterrorist episode is the ability to have two-way communication with our local and state partners.

We need to both be able to get information out and we need to get about rolls and public information and information that a local health department would need. We also need to be able to get information back about any suspected incidence that might, in fact, represent bioterrorist activity. So, we are working with a number of partners outside of CDC and inside CDC to design a next generation to our surveillance systems. And the core functionality would be designed around this preparedness for bioterrorist event.

However, as you can understand, the kinds of information you would want to know about a case of anthrax or a case of brucellosis(?) that are quite similar to what you might want to know about emerging infectious disease and there are many components of the core system, which will have a multiple functionality. So, we are looking at this as an opportunity to design a system architecture that would have multiple functionalities for other surveillance systems that we use.

Jeff Copland has made a commitment that we will be migrating our Legacy Systems to use this system architecture over time. We have also got an effort underway to develop the data definitions and data standards that would be used for this system. This will, of course, be consistent with the definitions and with SDOs that then will be taking on variables that are of particular relevance to public health. So, there is a data standardization effort as part of this.

We understand that the Secretary is quite committed to supporting enhancements in the public health data infrastructure. So, the President's budget for 2000 and, hopefully, for 2001 will continue a level of support for these efforts that we think will help in both the development of the system and also providing incentives for state and local departments to participate enthusiastically in the process. The intent certainly is for a lot of those resources to go out to the state and local areas.

I think I will stop there and see if people have questions or comments.

DR. LUMPKIN: All right.

DR. MC DONALD: I would just like to comment, as I have before, about how sort of forward thinking and powerfully effective CDC has been in pushing forward with kind of standardizing and gathering data. So, it fits perfectly as their mode. They have really been doing a lot of good work all over the place. I would like to wave the flag a little bit for you.

DR. BROOME: Thanks, Clem. I didn't plant that comment.

DR. MC DONALD: I said it once before. Spontaneous. I will say it again if it comes to pass, if I have another opportunity.

DR. BROOME: One of the projects that Clem has been helping us with is electronic laboratory reporting and one of the objectives, as I have said, is to be able to have electronic data interchange with primary data sources, such as large commercial laboratories. So, we have been working on pilot projects, using HL7 messaging to get notifiable diseases reported electronically. We are very optimistic that this will be a very useful pilot, not just for lab data but potentially resolve some policy issues related to other primary data sources.

DR. MC DONALD: The tumor registry is cranking up and, of course, immunization has been -- it is coherent and consistent and it is not like a lot of government or some government organizations.

DR. STARFIELD: Could you comment on the extent to which your whole variety of data sources contain information on race, ethnicity and -- I am particularly thinking of, you know, especially with your long history on epidemic control, the potential for geocoding.

DR. BROOME: The race/ethnicity issue, obviously, we have been very involved in the reactions to the changes in OMB Directive 15 and we also -- clearly, part of the core variables, which are part of our public health data sets are the race/ethnicity. We would -- you know, we are not 100 percent collecting that in Legacy Systems, but the intent would be to have full collection of race/ethnicity data, hopefully, down to the more precisely specified subgroups with the new systems.

We have been following with a great deal of interest the complexities of what will be available on the enrollment and encounter forms because we see this as very key to our ability to have that data. Many of our systems were not directly interacting with the patient.

We are dependent on what the health care system recorded as the race/ethnicity and we are dependent on access to that. So, the issues of what is on enrollment and encounter data and the availability of that will be very critical for our ability to track race/ethnicity.

You had a big discussion on that yesterday, right? Did you solve it?

DR. LUMPKIN: We identified it.

DR. BROOME: I would very much encourage -- and actually I have spent a lot of time talking to a lot of people about the importance of thinking effectively about that.

And the geocoding, we have actually had a -- Chuck Kroner(?) at NCHS keeps all of us up to speed on developments. There is a very enthusiastic GIS user's group and there is actually an Atlanta branch that is just starting. We certainly could use more, but I think we are recognizing a lot of creative applications and we do need to work on having data available.

We have no problem using it when it is available, but I think there are some issues in, again, having the capacity to --

DR. STARFIELD: Well, for example, the risk behavior survey gets information on at least race, I think, but not any attempt at geocoding. Now, maybe the survey includes that, but in general there are some data sets that could do it and I think aren't doing it.

DR. LUMPKIN: Other questions or comments?

I would like to -- Kepa.

MR. ZUBELDIA: Kepa Zubeldia.

In your estimate, what percentage of the research you do could be done with deidentified information as opposed to fully identifiable information?

DR. BROOME: I think it would be rash of me to make a blanket statement. Currently, a lot of what the states send to CDC do not have personal identifiers. Those are taken off at the state level before records are sent to CDC. So, for a variety of reasons we have certainly -- our databases do not have names. We usually have some sort of an identification number, which permits us to link back to states and to work with them on eliminating duplicates, et cetera, et cetera.

I understand that an identification number is many settings does not count as deidentified. So, the -- I think we basically need to look in individual settings as to what is the purpose of the research and what level of detail is needed. I do think we are very sensitive about the issues around confidentiality. I think we also have public health responsibilities that have to be met so that I am -- and I think we also have to put equal effort into using the most sophisticated technologic protection for confidentiality and data security as another approach in settings where we need to have identifiable information.

DR. LUMPKIN: I think in a way, though, that may have been a trick question. My guess would be that the answer to that is most.

DR. BROOME: Most are deidentified or most --

DR. LUMPKIN: No. Most of the research that you do can use non-identifiable information but CDC does more than research and that is probably where the discussion occurs. You know, when you are doing an outbreak investigation, you, obviously, can't use the identifiable information. You know, someone has got to be on there on site, has got to be talking to people who are coming down with this new and strange disease and doing the case investigation.

Part of it may be a definitional issue in describing exactly what CDC does and some of the efforts which are more directly related to working with individuals and identifying what is going on.

I would like to thank you for coming. It is certainly a thrill to hear all the neat things happening. I know it has been probably almost nine years that I have been working with CDC on issues related to integration of information systems and the progress that has been made is really thrilling.

We just want to thank you for the long partnership that we have had and are looking forward to working with you in your new position for a much longer partnership.

DR. BROOME: Well, I think I will be back.

DR. LUMPKIN: And you are welcome anytime.

DR. BROOME: Thank you.

DR. LUMPKIN: Thank you.

Agenda Item: Reports from Subcommittees and Work Groups

The next item on the agenda are the reports from the subcommittees and work groups.

Let's go with the Subcommittee on Population. Are there any additional things to report?

[There was no response.]

Any reports from the Work Group on Quality?

MS. COLTIN: No. Nothing to report. Yesterday, we didn't have a breakout of the Work Group on Quality. So, really, nothing has occurred beyond the panel that we had yesterday and we will be trying to plan for future panels electronically. Get suggestions from members about where they would like to focus attention with our work plan.

MS. GREENBERG: I should mention that the executive subcommittee is having a conference call -- I believe it is July 26 -- but in any event we are going to be really planning the September meeting, the late September full committee meeting. So, the subcommittees and others should be thinking both about what you want to be doing at your breakout sessions, but also anything that you want to bring to the full committee if we are going to have any -- you know, panels on particular issues. The sooner we can start planning those, the better because August is an awful time to try to reach people also.

So, we really are going to start before that July 26th conference call. So, please get your information and ideas to me and I will share it with everyone.

DR. LUMPKIN: Thank you.

The Privacy and Confidentiality.

MS. FRAWLEY: We met this morning and one of the things that our group worked on, if you remember the Medicaid Managed Care Data Collection and Reporting Report that we saw yesterday afternoon, our subcommittee was asked by the Subcommittee on Population to actually look at contract specifications, which was an appendix, which was not part of the document that you reviewed yesterday afternoon.

We spent some time this morning discussing the appendix and had some major concerns with the actual specifications for data that were in the document. Part of the concern was the fact that it was calling for collection of data and it was not clear who would be collecting the data, for what purpose, what types of data and how it would be used.

We felt that the section of the report we were looking at did not adequately address privacy and confidentiality concerns. So, we are going to be sending a letter back to Lisa Iezzoni on behalf of the subcommittee identifying the fact that we feel that the section of the report has to be expanded to discuss privacy and confidentiality and that before we start discussing any particular data collection, we really need to focus on the importance of privacy and the key principles that we think should be a foundation and that information should be aggregate or disidentified and that we should not assume that information should be identifiable and collected without really understanding why that information needs to be in identifiable form.

We did have some discussions of some key points that Bob Gellman brought forward to us and we also are going to recommend that this section have some specific discussion of anonymity of privacy and confidentiality enhancing technologies, that for each data collection activity and data sharing activity we need to really look at the risks and the benefits and really understand that if we are asking for an element to be collected, what are the benefits and what are the risks to privacy and really understand what the impact is in terms of current state and federal legislation.

We all had some concerns about the quality of the contractual language that we will share with Lisa and we just felt that overall that it wasn't really adequately positioning this. When you read it, it sounded you could collect anything and everything for any purpose without really putting it in the proper framework.

So, we decided that we would just send those recommendations back to Lisa and I guess to be shared with the contractor to see if the report can be adjusted.

The other thing that we felt was important is that there wasn't an adequate discussion of administration simplification and HIPAA, which also is another thing that needs to be identified.

The second part of our meeting we spent talking about our work plan. We have completed all of the items on our current work plan. So, what we started discussing is what we would like to do in 1999-2000 in terms of activities. We came up with a number of activities and one of them was on Smart Cards, the discussion we had this afternoon. We may take another look at that, looking at activities ongoing at the state level, linking health information in terms of what is going on in terms of health identifiers at the state level.

Probably interacting with staff here at HHS on the privacy recommendations and regulations that will be forthcoming. So, what we were thinking of doing is having a briefing on that and seeing if there were areas that the subcommittee could be helpful to the Department.

Also, looking at the use of the Internet in terms of health data and focusing on whether there is adequate privacy policies in that area. Also, looking at what is going on in the various agencies in terms of privacy issues. Nancy Ann Min DeParle mentioned that HCFA gave greater attention to privacy concerns. We know that the VA, Social Security Administration, here within the Department, there is heightened awareness. So, we may very well at some point down the road see if we have some of the representatives talking about what is going on in various agencies.

So, that right now, and the issue of identifiability of health information, which is something that we have flirted with before in the past and had some attention. So, we have about six or seven items on our work plan. So, we are planning that over the summer and then approving it at our September meeting.

DR. LUMPKIN: You used the term "privacy-enhancing technology." I think the meaning of the term is immediately obvious, but I am not sure that we have ever had any discussion on the committee that really looked at what is the state of the art of privacy-enhancing technology.

So, if you do have time in that work plan to perhaps -- it might be worthy to sort of do a milestone or at least if there is some literature out there to share that with the rest of the committee.

MS. FRAWLEY: I will just defer to any of the other subcommittee members if there is anything else.

DR. MC DONALD: Just to comment on that. There is sort of a ying and a yang to that. Some of this deidentified data doesn't really help that much for any serious attack. You have got things like dates and times and some of -- what it really amounts to, if someone knows something else about you, they can match up against any of the variables or you have a rare disease and they know what city you are from and that sort of thing.

To really do it right is fairly hard and, so, I think we ought to have a balance in what to do, how much we have to have regulatory and how much we have is technical because some of the technical work, it doesn't give you as much as you think in terms of absolute security.

DR. LUMPKIN: I think that is probably the issue that we need to have addressed because I thought it was very important in our discussion on Smart Cards that we had a balanced approach because any new technology has a lot of promise, but there is also some downsides. I think we need to as a committee make sure that we understand the ramifications.

Anything else on privacy and confidentiality?

MS. FRAWLEY: No, that is all of our report.

DR. LUMPKIN: Simon.

DR. COHN: I will present on both the Subcommittee on Standards and Security, as well as our Work Group on Computer-Based Patient Records. I will ask Mike to probably join in on the piece. Actually, we probably should start off with the Computer-Based Patient Records Work Group since that is the majority of the activity over the next nine months.

The work group, I think, reviewed both its work plan and looked at the time line related to what will make a successful and the committee itself successful in submitting recommendations to the Secretary on recommendations for patient medical record information and the electronic exchange of such by August of the year 2000.

If we move back from that, we recognize that we have to approve at our June meeting the recommendations that will be forwarded on to the Secretary. Based on our experience at this meeting with a draft -- some draft recommendations on Medicaid. One of the things that the work group and subcommittee would recognize is that we need to have at our March meeting a set of draft recommendations for the whole committee to vet and begin to talk about, which begins to mean that there isn't very much time to have all of this together.

So, one of the things that we are going to be doing in terms of activities is in September -- well, first of all, we are going to be having hearings this fall, I think, as you all know, to future identify the issues and begin to come to conclusion about what the recommendations should be, but we are anticipating in September to be bringing forward to the committee a letter to the Secretary, not quite an interim report, but a brief letter identifying the focus areas and the issues, which will come to a vote at that point.

We also expect at the November meeting that we will be presenting a briefing to the full committee where we will identify the issues that we have identified and just beginning to, I think, generate some discussion by the full committee on some of those issues without solutions, but just sort of the issues we are conceptualizing them and getting everyone's input.

I think that should allow us by March to come forward with a draft. So, obviously, the first item there was assisting the committee with the development of that report and I think that is moving forward in a very expeditious manner. I just want to thank everybody because I know the amount of the work that the subcommittee and work group are putting into, as well as Mike Fitzmaurice has been extraordinary. And I think will continue on at least until March.

Mike, before I move on to the other agenda items we are doing, do you have any other comments about that?

DR. FITZMAURICE: That pretty well covers it. The work between now and probably the end of July, because as John mentioned earlier, we have to get the people before August because people leave town in August. If you have any questions that you want asked of users of patient medical records standards, please feel free to contact me, Jeff Blair or Simon Cohn to make suggestions on areas or issues that come up.

As we develop the issues, it is going to be important to get a good list of them in September so that we can report to you in November, but also to start fleshing it out in preparation for writing the report. So, issues dealing with patient medical record information and standardization, please get to us on that.

DR. COHN: There will be a set of hearings on September 16th and 17th by the work group.

DR. STARFIELD: I suppose one of the possibilities for a meeting is February instead of March. Will you still be ready?

DR. COHN: I don't believe so. That is one of my concerns. Anything is possible though.

MS. GREENBERG: We will see what we get back from people as to their dates. It is sounding to me that March would make more sense for this report. There may be some other dates that we didn't explore and if we really can't seem to -- you know, we will get the best dates possible and try and take into account also what the timings are.

DR. COHN: Sure.

I think the only other issue I would put on the table is this one and I would actually direct that question to the population subcommittee in particular, only because I know you are doing some work on functional status. I don't know the timing. One would think that there might be some synergy potentially there.

So, I would just want to throw that out as just a thought ten months before the report is due.

Now, the overall committee review -- excuse me -- the subcommittee review its work plan, the overall Subcommittee on Standards and Security, with the exception of that first item, it actually also completed its work plan and so reviewed its charge and came up with thoughts about what its work plan will be over the next 14 to 16 months.

Certainly, one of the charges that is ongoing is to produce recommendations to the full committee, the full committee's annual report to Congress, particularly the progress of implementation of administration simplification and the subcommittee recognized that there was going to be a need for some intensive work to occur to identify what the preimplementation status is.

We initially were thinking about in terms of the administrative transactions alone, which would take a fair amount of work to identify that, stage zero, the preimplementation phase. When one begins to think about recognizing the security standards, the security rules and the implications, it begins to take on a much broader context and will, I think, turn out to be a fairly substantial activity.

This will be something that the subcommittee will talk about more in September, but I think there is a realization of that being a relatively involved activity over the next while.

In addition, the subcommittee has a responsibility to help advise the full committee on both how the standards are being implemented on an ongoing basis, the -- taking to the full committee our recommendations for new standards and changes to currently existing standards. So, this will be an activity that will occur during -- through hearings into early and mid year 2000, hopefully, occurring after we have at least our draft copy of the CPR work group report.

DR. STARFIELD: Is it reasonable to ask the subcommittee to deal with the issue that we surfaced this morning about the transmission of race/ethnicity data, I mean, what we might recommend in terms of that, the implementation standard?

DR. COHN: That is a good question. I guess I personally believe that before that is handed to the subcommittee for further evaluation, probably the full committee needs to vet the issue of GIS and better understand where that fits into the whole piece. But I guess the --

DR. LUMPKIN: If I can interject. One of the things about GIS and this is -- you know, we didn't have a lot of time for discussion, but our demographers, who are the people in our state who do the population estimates in the intercental(?) years, basically don't like to do estimates of race/ethnicity too far beyond -- one or two years beyond the censal(?) years because what we have found from experience is that our population projects or estimates are not accurate because there is such a dynamic change. So, I am just a little bit concerned that depending on the study, some GIS studies may be useful but I think we as a committee need to look at to what extent can we use the current EDI technology to perhaps evaluate the feasibility of alternative methods of reporting, other than the 837.

If there are some viable alternatives, then that certainly may be one of the ways to do that.

DR. COHN: Certainly it is within the scope of the subcommittee to evaluate and make recommendations regarding proposals for new HIPAA administrative simplification standards and that if this is within that scope, then that his appropriate. But I think this is something that probably needs to be talked about at the executive committee level to reach a firm decision about how we should proceed on this one.

DR. LUMPKIN: Sounds reasonable.

DR. COHN: That is the report. Are there any other comments from the members?

I really want to thank all the members of the subcommittee and the work group. We are all one and the same in the sense that it is -- this is a tremendous amount of work and I want you to know I appreciate it and we are going to be doing a lot more work over the next six to eight months. So, thank you for your help.

DR. LUMPKIN: Thank you.

Dan, the Work Group on Health Statistics.

DR. FRIEDMAN: I am beginning to feel that some of us are sort of seekers wending our way towards a somewhat uncertain goal. I don't mean that to sound discouraging. We have got sort of milestones along the way and we are hitting those milestones and at the same time it is not as if we have --

DR. LUMPKIN: You are not supposed to hit them. You are supposed to pass by them.

DR. FRIEDMAN: Rolling over them. But it is not as if we have a final date certain for product. So, we are going to roll into the next millennium still visioning. Specifically there are five papers that have been commissioned and we should have first drafts of those by early September. We have been having regular conference calls with the authors of the papers. We have had three discussion groups here in D.C. over the last several months that several -- that Marjorie has participated in everyone. Barbara has. Kathy came down for one. Ed Sondek has been there. Very, very good representation, really excellent representation from a whole variety of people, who at the risk of forgetting most of their names, I won't list them.

It ranged from -- I won't name them because I will forget too many of them. But excellent representation. We should have -- we have drafts of reports of those sessions that have been prepared and we are working through the drafts and those should be available within a couple of months.

Since the Committee on National Statistics Workshop has been scheduled for early November and we received a second draft of that agenda. So, we will have sort of an interesting process working through the rest of the agenda with CINSTAT(?).

We are in the process of developing a -- something between a framework and a plan for more formally incorporating professional and more public input into the process. I think this is really absolutely essential. I think it is essential at the beginning stages. That is -- and particularly Rob Wienzimmer(?) and Ed Hunter and Marjorie Greenberg from NCHS have been really working on trying to develop that outline and that draft.

We are not exactly sure what that input is going to consist of -- it will probably consist of a combination of sessions at -- fairly open sessions of professional meetings. We had one of those a couple of weeks ago at the National Association of Public Health Statistics and Information Systems meeting. It was very successful, much more than I frankly had thought.

Ed Hunter is going to be out at AHSR or maybe there this week, soliciting input as well, basically laying out the process and getting input on it.

We are hoping to have -- I am hoping to have several sessions in real places. We keep going back to Fargo, but not necessarily Fargo but hopefully not Washington and hopefully someplace where we can actually elicit input from local government officials, consumers, health care providers who are not steeped in Washington and our state capital processes.

We have identified in a preliminary way three products that we are going to be aiming for and this is sort of the wending our way towards a somewhat uncertain future because we have not put a time frame on these. One is a vision document. The second is a, quote, road map and a third, hopefully, will be a collection of -- an edited collection of papers, which would include the commissioned papers because there really has not been an adequate collection surround surveillance. There hasn't been anything around health statistics.

Having said this, I think one of the things that we really need to work out over time is how -- and we have started with these discussions, too -- is how this is going to fit in with the national health information infrastructure discussions. I think sooner or later we are going to have to deal with some of the -- continue to deal with some of the definitional issues, which I have sort of tried to put aside because we have health statistics as a nice easy label. At the same time, increasingly I think health statistics -- I am not going to give my -- Gail has already heard my speech on this. I am not going to repeat it -- Gail Janes has.

But I think increasingly that, you know, surveillance and health statistics are really just essentially obsolescent terms that are confusing more than they are illuminating and sooner or later, maybe around the NHII or maybe around the end of this process we really need to, I think, finally join that issue and really try to explore it.

So, at some point -- I can't give a date for a final -- for the endpoint of the process.

MS. GREENBERG: I was just going to add that part of this public process will be -- we are talking about would be hearings that some of which would be done under auspices of the National Committee.

DR. FRIEDMAN: Yes, this is a completely collaborative process, particularly with NCHS. We have monthly conference calls with Marjorie, with Gerry Hendershot, with -- Barbara has been on most of them. And it has been a completely collaborative process and really continues to be.

MS. GREENBERG: I am going to have to make this clear in an e-mail to everybody because, obviously, isn't here now, but the entire committee is invited to the CINSTAD(?) workshop, which will be the 4th and 5th of November, following a one-day meeting of the committee on the 3rd of November.

DR. LUMPKIN: Any questions?

[There was no response.]

The last working group is the National Health Information Infrastructure Working Group. Dan, you want to report -- oh, wait. No, I am chair.

I think we have had some very important discussions as we are trying to put a framework on this and conceptually within the committee it is -- I mentioned yesterday about how a lot of our standard work is left brain stuff and a lot of what we are doing in populations is right brain and the privacy and confidentiality is like the cerebellum because it is balanced.

But the NHII activities really are what brings that altogether, kind of like the mid brain does and I think that is really -- the goal is not to reinvent the wheel where it is not necessary but to put a framework around all of our activities that we can work towards filling in.

We spent a good chunk of the morning trying to identify what it is we were exactly talking about and looking at the dimensions of health information. We are beginning to develop a multidimensional model, which will help us define some of the groups that we will need to involve.

The intention is is to look at it as a two-stage process. The first thing is that we have sort of a conceptual model that we are working towards. We have seen what has been done in other countries, but the goal isn't just to have a national data dictionary, but it is broader than that. So, we want to try to define those people who we can bring in in the first phase to help us put the flesh on this skeleton.

Then at the -- essentially the goal is to have by the June meeting a document that will describe what this process will be and what the ultimate goal is and then after that to begin to bring in others to help us try to achieve that goal. So, that is just kind of a rough description.

Is there anybody else who was there want to add more?

Marjorie.

MS. GREENBERG: Just one thing to keep in mind. June 2000 is seeming very soon. I guess it is less than -- it is a year away but I know -- you know, we are talking about a symposium and reception probably on June 20th or 21st around the 50th anniversary. We saw that as a milestone, not an endpoint, but as a milestone to maybe roll out some of the -- on the NHII and the 21st Century vision. Neither of them will be completed then, but just thinking of having something at that point that we can share with others.

DR. LUMPKIN: And that is the goal.

Did I skip over any of the committees and work groups?

PARTICIPANT: Subcommittee on Population.

DR. LUMPKIN: No, I started with that.

MS. GREENBERG: The committee approved the one report.

DR. LUMPKIN: I asked Kathryn if she had any --

MS. GREENBERG: The other report was with --

MS. COLTIN: The work we have done to date was really discussed yesterday in the committee and you have already approved a work plan, so you know what is coming up. We do have a conference call that we have scheduled to plan the work around functional status. But that is really -- there is nothing else.

MS. GREENBERG: The only meetings we have scheduled are the CPR Work Group meeting in September, the full committee meeting in September. We have a conference call for the executive subcommittee and some of the other subcommittees might have conference calls, but if there are going to be any meetings other than those --

DR. COHN: In September or to the rest of --

MS. GREENBERG: To the end of the fiscal year, which is October 1st. We need to know because we really have to plan, knowing if there are going to be any other meetings and then, of course, whatever we can know through the end of the calendar year also. I know you have yours scheduled, but I think other than the CPR, we really don't have anything than the two full committee meetings -- the Subcommittee on Populations was tentatively talking about meeting the 29th and 30th or one or both days after the 27th and 28th September meeting. If that is going to be the case, we really have to block that in. I know we are polling for it. But I am not quite sure where it is.

DR. LUMPKIN: One final point, that the NHII work group will be holding conference calls between now and the next meeting because there are things that we need to do. So, those of you who are -- and other work groups or subcommittees, who couldn't make it to the meetings, but are interested, please let Marjorie know and then we will make sure that you are notified of the conference call so you can participate.

DR. STARFIELD: There are a lot of conference calls and it is getting hard to keep them all straight so it would be nice to have a list of them.

DR. LUMPKIN: Put them on the Web page.

DR. STARFIELD: There are those of us who don't consult the Web page.

DR. LUMPKIN: Well, that is my bible.

Agenda Item: Healthy People 2010 -- Leading Health Indicators and Evolution of Healthy People and its Data Challenge

Okay. Next item on our agenda is a presentation from ODPHP.

MS. MAIESE: Thank you, Dr. Lumpkin. It is nice to be here today. I am joined by my colleague, Linda Bailey, who has been working on leading health indicators. And I can't believe it has been a whole year since I sat before this full committee. Last June I was here with Oliver Carter-Pokras from the Office of Minority Health and Richard Kline(?) from the National Center for Health Statistics and, you know, promising you that we were going to get draft objectives out for public comment. Certainly we were successful at publishing sort of what we began to call the Yellow Pages of Health last fall for public comment.

And comments we got. In fact, I think there is some material in your briefing book that shows that we heard from people in every state in the United States and Puerto Rico. In fact, one of the things that I did just before coming here is I decided to go into the comment Web site and it is completely searchable. So, I decided to put the chairman's home state in the field of sort by Illinois and found out that there were 644 comments that either come from people in Illinois or identified something about Illinois.

So, I sort of put this out to you to say that this comment Web site is there for the using and for the asking. It is completely word searchable. So, in the event, I plugged in the word "data" and believe it or not there were 5,844 hits on the word "data." Data is very much on the radar screen of those people who were participating in the public comment period.

We got more than a thousand comments, for example, on the chapter on access. We heard not only the concerns about health insurance and having a primary medical care home, but we also heard from EMTs around the country about response side in emergency encounters. We got more than 500 comments on the educational and community-based programs chapter, again, because school health nurses were really concerned about the pupil ratio with school nurses.

It is a robust comment set. We even got comments on things that weren't in the book. In particular, we heard from virtually every nephrologist in the country and many kidney dialysis patients and family members of kidney dialysis patients, who said you really have not addressed chronic kidney disease appropriately in this initiative. As a result of those public comments and that outpouring of support, the Healthy People Steering Committee recommended and Dr. Satcher has added a chronic kidney disease chapter to the 26 that we have already published.

The other subject that was suggested during public comments was that we had not appropriately addressed vision and hearing impairment, something that really hadn't been covered very well in Healthy People 2000. Now I really understand why. It is because we have so little data. Almost all of the vision and hearing objectives that have come before the Secretary's Council this spring are developmental objectives. In other words, those objectives that we don't have national baseline data, but you know -- some of you know us very well in ODPHP and we are sort of risk takers and brave folks and so we are putting forward a chapter on vision and hearing impairment that is a significant data development agenda.

Which probably brings me to the issue that really is of most concern to this committee and that is where we stand with the data for this particular next set of objectives. This yellow book contained 531 objectives of which 43 percent were developmental and roughly that same ratio exists in the set of objectives that was brought forward to the Secretary's Council this April 23rd.

In that meeting -- and that transcript will soon go on the Web for you to see the proceedings and read verbatim what was said, but there was a lot of soul searching about the extent to which we can really go forward to the American people really saying that we have the resources to make these commitments to get all of these objectives measured. So, we continue to struggle.

The lead agencies for each of these chapters are really working through, you know, where are these in their data development stages. We all know there is a lot of work that goes on trying to get these through the cognitive labs, really make sure that what we are going to get measured is really going to be reliable. Then also there is a great concern about sort of the competition, if you will, for finding places for all these various subject matters to be covered in the National Health Interview Survey. Discussion is underway about putting some of this in NHANES and certainly then the development with the behavioral risk factor surveillance and the YRBS, the National Household Survey on Drug Abuse and the list goes on.

So, there really is a significant and tall data development agenda still being worked on in the Department as we attempt to finalize these objectives for clearance.

One of the things that we heard during the public comment period is that people said data not available was insufficient and actually one of the things we have begun to do in the template that exists for every racial and ethnic group -- and by the way, we are using the new 2003 standard of multi-race and breaking out racial and ethnic groups even to distinguish Asian Pacific Islanders from Native Hawaiians. So, the template is really designed for the 21st Century.

We are going to disclose whether or not the data have been collected, perhaps not yet analyzed or are too small to report reliably. So, this is going to be a full data disclosure in each of these templates. The standard template includes not just race, ethnicity and gender, but some measure of socioeconomic status, be it income or education in a gradient.

So, we are really trying to put out there under each of these population-based objectives what we know and what we don't know and thereby enable people to really see themselves and then target the disparities.

We have remained committed to the one target, better than the best, for all population groups in risks, behaviors, in services. So, that eliminate health disparities goal in Healthy People 2010 really remain very much central to our initiative and we continue to discuss some of the outcome measures and as we go through clearance, I presume that those better than the best targets will again be reviewed by all of the various constituencies to this initiative.

We anticipate or we actually know that we are planning a launch on January 25th of the year 2000. In fact, one of the things we are about to do -- and I will pass around this request for call for abstracts. This will go up on our Web site within the next two weeks. We really are encouraging people to send forward proposals for breakout sessions, for caucuses, for poster sessions that really showcase what has been done with Healthy People 2000, those lessons learned, as well as what is being planned for Healthy People 2010.

We are focusing not only on the goals of Healthy People increasing years and quality of healthy life and eliminating health disparities, but we are also focusing on partnerships and particularly on using technology for communicating health improvement.

The other thing that is news and next steps is the Public Health Foundation under contract with us is releasing a Healthy People 2010 tool kit this summer. This is really going to be a tool kit designed for state translation, state adaptation and adoption of these objectives and certainly a number of states have their activities underway on Healthy People 2010 and we certainly are trying to do everything we can from the Department to support that local adaptation.

Agenda Item: Leading Health Indicators

So, maybe I will stop here and turn to Linda Bailey who has been working in many domains but specifically is here today to talk about our leading health indicators.

MS. BAILEY: Thank you for the opportunity to come and talk to you today. I reported, I think, on Halloween to one of the subcommittees about our work in leading health indicators and we have made some progress since then.

What I wanted to review just briefly was, first of all, why we are developing a set of leading health indicators for Healthy People 2010. Secondly, how we approach the development of that and then to report to you on our progress there.

I have a one-page summary that I just wanted to hand around. The idea of doing leading health indicators for Healthy People 2010 really came out of some focus groups that Debbie and ODPHP did about four years ago. The purpose of the focus groups was to find out why some people used Healthy People and why other organizations, who we know are important for health improvement were not using Healthy People.

What we heard was really very conflicting. From the people who were heavy users of Healthy People, state and local health agencies, a number of medical groups, public health groups, what we heard was they liked it the way it was. They wanted it to be bigger, if anything, and they used it as a menu. Really, they wanted to make sure their state plans were reflected and what we had is a national document. So, bigger was better to them.

What we heard from a number of the people who weren't using Healthy People -- and this includes business, managed care organizations, a lot of the partner groups that we think are very important for health improvement over the next decade, was we don't use it because maybe we don't do public health everyday and 300 objectives is just so big we can't get our hands around it. We don't really understand where to start with Healthy People.

So, what the deputy assistant secretary for health conceptualized was having a small set of health indicators that would reflect the breadth and scope of Healthy People, but be small enough that even non-health professionals could understand it and use it day to day and find it easily accessible.

So, we set about developing a set of 20 or less health indicators knowing that we needed to select most of them from the Healthy People document and maybe had a little opportunity for innovation, to get a couple more things into Healthy People that in some way would reflect the concerns of these non-health professionals.

We started this about 20 months ago with a work group internal to the Department of Health and Human Services. We did a report that I think I gave out to the subcommittee. It is this report here. It is available on the Web. Just identifying what we wanted out of leading health indicators.

We then contracted with the Institute of Medicine and they put together a committee and delivered three reports to us in less than a year. So, it was a very fast paced Institute of Medicine committee. The final report is up on the Web. It is a very nice report, very small, where they identified three separate models for us of what we could do with leading health indicators.

They had a health determinance model. They had a prevention model and they had a life stages approach. We had designed a new model for Healthy People that was based on evidence in Stoddard, as well as -- a determinance kind of model. So, that concept from the Institute of Medicine really resonated best with the Secretary's Council and here within the Department, as well as with our partners. So, we have moved forward with a determinance kind of approach and what we are seeing is that we have two goals for Healthy People.

We have probably ten to twelve leading health indicators and then we have the 28 chapters of Healthy People. One of the things that was very important to us was to make sure that each of the 28 chapters was reflected in at least one of the leading health indicators. We have that kind of coverage.

If you look at the handout that I gave you, the 12 indicator topics that went to the Secretary's Council in April are bulleted down there and you can see that they really do represent the scope of health issues and concerns and the scope of Healthy People. One is on environmental health. One is on teen tobacco use. One is on healthy weight. And with that one we are really struggling with are we looking at weight or should we be looking at nutrition.

The next one is physical activity. The next one is safer community and there we are looking at preventable injuries versus violence; you know, which of the measures is most compelling to people and we will get them acting. The next one is immunizations for children, as well as adults. Access to health care and we are probably going to use insurance, present to people with insurance because it did so well in focus group testing and understanding that we are probably going to track regular source of care as a complementary measure.

Levels of poverty and there we are struggling with is it children's poverty, children living in poverty or all people living in poverty? I think we are getting pressure to go with all people and it would be interesting to have your inputs there.

Levels of education, and that we are using high school graduation. Substance use, and there we are using teens who use either alcohol or illicit drugs and probably a measure of adult heavy drinking.

We have a measure on disability that we are still working on. It probably is going to be physical and mental well-being. We are thinking about using the SF12 there and having the two summary measures and then sexual -- responsible sexual behavior. And, again, there, that measure especially changes with life stages. So, it is difficult to find the right measure there, whether we want to focus on teens or adults.

But those are the types of measures that we are considering. It may be that one or two drop out because we are trying to get it down to a list of ten.

To finish up the work, since the April Secretary's Council meeting, we have done two things. First, we did focus group testing of these 12 measures and the focus group testing was some of the most interesting conversations I have heard since our regional meetings.

We did a total of eight focus groups, two in Bethesda, Maryland, primarily white. It was a mixed group but it was primarily white, upper SES, middle and upper SES. We did a Native American group in Tucson. We did a Hispanic male and a Hispanic female group in California. We also did a mixed group of Asians in California. We did a mixed African American group in Chicago and a group of Chicagoans over age 55, just to get a sense of how they think about health, where they think they can take action and can control their health and where they don't.

It really was very interesting and will feed into how we phrase and frame and communicate these leading health indicators.

The second piece of follow-up work we have been doing has been data consultations, really looking at this set and trying to get a sense of what proportion of premature mortality and of disability can we address through this set of measures. That is a very difficult thing to calculate.

But we have been trying to put together some data so that we better understand if we achieve these, if we make progress here, you know, what are we going to see is the effect of that progress.

So, that is where we are. We are finishing this up and this will clear the Department, together with Healthy People 2010 and be published in January with 2010.

DR. LUMPKIN: Thank you.

Any questions? Dan.

DR. FRIEDMAN: You know, one of the nice things about the consensus health indicators for the year 2000 with which for people who don't speak CDC, speak more of a set of, I don't know, 17, 18, 19, 20, indicators that essentially were scalable -- almost all with one exception, they were all scalable down to the local level. So, essentially, they -- I think they are real useful to a lot of folks and these clearly are not. Some of these are scalable down from the state up. Some of them aren't even at the state level or will you be using -- is there going to be a replacement for the consensus health indicators? Are those going to be ported over to 2010? What is doing with those?

MS. BAILEY: We have been working with NCHS on that. Most of these actually are either in the consensus set or the priority data needs set. One of the things we really struggled with was trying to find indicators that were measurable at least at the state level and to the degree possible at the county level as well.

But there is actually quite a bit of overlap there with the environment, tobacco, physical activity, the poverty measures. I haven't gone through and counted them, but we did consider all of those there. I am not sure how NCHS really is planning to use the consensus set, but what will probably work with them on is to the degree that something was in the consensus set and has changed slightly, can we coordinate that so that we are looking at the some of the same measures here.

MS. MAIESE: There are a couple of answers. I don't know if the committee has received their Healthy People Review yet, but one of the things that I found really interesting is like with Health U.S., this time and for the first time, the Healthy People Review has the 18 health status indicators listed and you can carry it around with you. It is sort of a variety.

My understanding of NCHS's interest in maintaining the health status indicators and they are imbedded in one of the public health infrastructure data and surveillance system objectives in 2010 -- one of the concerns that I have had and it is interesting to hear you express such interest in their use is the fact that 9 of the 18 of those health status indicators focused on mortality. It has always been something that has been of concern to me from the standpoint that, you know, does my mother or my mother-in-law really truly understand what is actionable on those mortality measures.

I think that is one of the things that on this list that has been emerging on the leading health indicators is very few death measures there at all. But certainly I think your point is well-taken and that we certainly need to do some cross walk.

The other thing that we have been trying to do, though, is look at this list of leading health indicators to be measured at minimum at state levels and preferably at local levels as well. Certainly, that is one of the criteria that is being examined as we try to finalize these leading set of health indicators.

DR. FRIEDMAN: I am not tied to the consensus health indicators in and of themselves, but I do think having measures that are scalable -- in the state level for states is not very useful. As a matter of fact it is barely useful at all because most states don't particularly care what is going on in the neighboring state, as I am sure you know.

What is really useful for us are measures that are scalable up from whether it is the county or the city or the town, whatever, that we can then compare to the state, et cetera, et cetera, to other benchmarks.

MS. MAIESE: I think that would be an interesting thing for everyone to be involved back at home in their state 2010 action and see the extent to which states can really do -- advance our thinking and our tracking on a set of objectives that really can come from the grass roots and local level. So, certainly, I think we can learn from states as they do this process, particularly looking at those measures that really work for localities.

DR. LUMPKIN: I have a question on developmental objectives. How many of those from the year 2000 objectives got developed?

MS. MAIESE: Ah, good question. In this book all but eight of the objectives that had no baseline have been measured. So, we started with 91 of the 300 had no national baseline and we are down to eight. So, we really have succeeded in part because of the CDC School Health Programs and Practices Survey, in part because of a work site health promotion survey and in part because of a provider survey, we really established baselines in sort of three categories of objectives that we had nothing to start.

So, I think we really have used the objectives as the driver for data development over this decade and that we have been successful enough that certainly Ed Sondek was unequivocal at the Secretary's Council meeting and actually at a recent Data Council meeting saying it is really important to not look just under that lamp post, but really use this.

And I, once again, would encourage people and state and local and development of objectives to do the same thing.

DR. LUMPKIN: But, of course, that is where the light is.

National Committee on Vital and Health Statistics

June 24, 1999

Proceedings By: CASET Associates, Ltd. 10201 Lee Highway, Suite 160 Fairfax, Virginia 22030 (703) 352-0091

PARTICIPANTS:

Committee Members:

Liaison Representative:

TABLE OF CONTENTS

P RO C E E D I N G S [10:00 a.m.]

Agenda Item: Call to Order

Agenda Item: Panel on Collection of Demographic Data on HIPAA Transactions

A F T E R N O O N S E S S I O N [1:04 p.m.]

Agenda Item: Smart Card Development

Agenda Item: Discussion with CDC Deputy Director for Science and Public Health

Agenda Item: Reports from Subcommittees and Work Groups

Agenda Item: Healthy People 2010 -- Leading Health Indicators and Evolution of Healthy People and its Data Challenge

Agenda Item: Leading Health Indicators

Proceedings By:
CASET Associates, Ltd.
10201 Lee Highway, Suite 160
Fairfax, Virginia 22030
(703) 352-0091