Transcript of the November 20, 2000 NCVHS Joint Hearings of Workgroups on National Health Information Infrastructure and Health Statistics for the 21st Century


[This Transcript is Unedited]

National Committee on Vital and Health Statistics

Workgroup on National Health Information Infrastructure
the Workgroup on Health Statistics for the 21st Century

November 20, 2000

National Center for Health Statistics
4105 Hopson Road
Research Triangle Park, North Carolina

Proceedings By:
CASET Associates, Ltd.
10201 Lee Highway, Suite 160
Fairfax, Virginia 22030

List of Participants:

  • John R. Lumpkin, Chair
  • Daniel Friedman
  • Jeffrey Blair
  • Mary Jo Deering
  • Ed Sondik
  • Rob Weinzimier
  • John M. Booker
  • Paul Buescher
  • Janis L. Curtis
  • David Kibbe
  • Marc Kolman
  • Chris Mansfield
  • Tom Ricketts
  • Vaughn M. Upshaw
  • Christopher Mansfield
  • Kepa Zubeldia
  • Richard Harding
  • Ed Hunter


  • Call to Order and Introductions
  • Overview of Interim Reports
  • Panel 1:
    • Tom Ricketts
    • David Kibbe
    • Janis L. Curtis
  • Panel 2:
    • Christopher Mansfield
    • Paul Buescher
  • Joint Panel on NHII and Health Statistics for the 21st Century
    • Marc Kolman
    • John M. Booker
    • Vaughn M. Upshaw
  • Public Testimony

P R O C E E D I N G S (9:42 a.m.)

DR. LUMPKIN: We're going to get prepared to get started. Good morning. My name is John Lumpkin, and I am Chair of the National Committee on Vital and Health Statistics. I'd like to welcome everyone to a joint work group, the Work Group on National Health and Information Infrastructure and the Work Group on Health Statistics for the 21st Century.

I would also like to see if, Dan, you have any welcoming comments before we go into introductions.

DR. FRIEDMAN: No, thanks, John.

DR. LUMPKIN: We'll start off with introductions. I'd like to tell everyone that we are going live over the Internet, so as such, as our usual thing, we would like people to first introduce themselves, and of course when you speak, if you would tell people who you are, because obviously we know, but not everyone on the Internet would necessarily know.

My name is John Lumpkin. As I mentioned before, I chair the work group on NHII. I am also a director of the Illinois Department of Public Health, and only of pertinence to this meeting, Perviso High School, class of 1969. Dan?

DR. FRIEDMAN: I'm Dan Friedman with the Massachusetts Department of Public Health. I chair the Work Group on Health Statistics for the 21st Century, and UNC, 1972.

DR. HARDING: I'm Richard Harding. I am a member of the committee, and I am vice chair of psychiatry for the University of South Carolina School of Medicine, and a child psychiatrist.

MS. CURTIS: My name is Janis Curtis. I am the Senior Associate Chief Information Officer at Duke University Health System, and I am a 1970 graduate of Perviso High School, so that was why Dr. Lumpkin made that statement. Our paths kind of crossed a little bit earlier in our lives. I appreciate the opportunity to come and share with you some thoughts this morning about the proposed infrastructure.

DR. KIBBE: Good morning. My name is David Kibbe. I am one of the panel members. I am the CEO of a company called Canopy Systems, based here in Chapel Hill, North Carolina. I am also a family physician.

DR. LUMPKIN: Actually, I'd like to ask everyone, when you are speaking, you have to move the microphone fairly close to your mouth so it picks up, so we can get a good signal.

DR. RICKETTS: Tom Ricketts. I'm the Deputy Director of the Cecil G. Sheps Center for Health Services Research, and chair of the Committee on Health Information Policy in North Carolina.

DR. ZUBELDIA: Kepa Zubeldia, President of Clare Corporation and member of the committee.

DR. WEINZIMIER: Rob Weinzimier, Chief of the Data Dissemination Branch at the National Center for Health Statistics, and for the purposes of this meeting, though slightly outnumbered, a proud Maryland Terrapin fan.

DR. DEERING: Mary Jo Deering in the Office of Disease Prevention and Health Promotion. I am the lead HHS staff for the NHII work group.

DR. HUNTER: Another proud Maryland Terrapin fan. Ed Hunter, Associate Director for Planning and Budget Legislation at the National Center for Health Statistics.

DR. BUESCHER: I'm Paul Buescher. I work at the State Center for Health Statistics in the North Carolina Division of Public Health. I am a statistician in charge of the statistical services unit.

DR. BOOKER: I'm John Booker. I am the director of the now Center for Health Informatics and Statistics in the North Carolina Department of Health and Human Services, and I work with Paul.

DR. STEINDEL: I'm Steve Steindel from the Centers for Disease Control and Prevention, and staff to the NHII work group, and versus the Maryland people, a Georgia Tech fan.

DR. BLAIR: I'm Jeff Blair, Vice President of the Medical Records Institute and a member of the committee.

DR. SONDIK: I'm Ed Sondik, Director of the National Center for Health Statistics, and a fan.

DR. JACKSON: Debby Jackson, National Center for Health Statistics.

DR. UPCHURCH: Patrice Upchurch, National Center for Health Statistics.

DR. MACKISON: I'm Delta Mackison, National Center for Health Statistics..

DR. SURK: Chuck Surk, National Center for Health Statistics.

DR. CHRISTIE: Heather Christie with Emreddy and Harris.

DR. LUMPKIN: Great. We are here to discuss two reports. Actually, we are not here to discuss two reports, we are here to go beyond these two reports and move towards the development of the final documents.

The National Committee over the last number of years has had a mission which is to look at health-related information and health-related data and data collection. In 1996 with the passage of HIPAA, that activity took on a particularly sharp focus related to standards related to patient medical record information, standard transactions, electronic data interchange, which really brought us much deeper into the realm of health data related standards involved in health care particularly, and really became a sharp focus of the committee.

Over the last year or so, the committee has been trying to redefine its efforts. Most recently, we engaged in a strategic planning process of the executive subcommittee, which was reviewed by the full committee in its last meeting, to begin to set a new direction, which allows us to take where we have been with HIPAA and where we have been historically, and to wrap it up into a consistent whole, a vision for health information into the future.

An important document that has pulled that work together was an initiative to develop a definition of a national health information infrastructure, and we are toying with the title now, Better Information for Better Health.

This approach looks at a much broader and higher level of rendition of health information, so that we can have a national vision. It is kind of hard to get someplace if you don't know where you're going, if you don't know where the roadmap is, you don't know the direction. One could say it would be very hard for a candidate to decide where they are going to be going in their election unless they decide at some point they want to run for President. And of course, that is always a very decisive moment, and having an election is a defining moment in someone's life, even if it goes on for weeks and weeks in a recount.

But for us, there is no particular election. There is no singular event. But we do know that a significant milestone has been yet to be achieved in the area of health information, to the extent that we have established a way for us to have a system that functions at a high level, a level that allow interoperability of information, of information to pass from one system to the other. As important as it is to transport the information, the meaning must also go along with the data.

These are some of the issues that we have struggled with in regards to the most recent document and letter that we sent to the Secretary on patient medical record information. That certainly fulfills a major piece of what we are talking about in a national health information infrastructure.

But yet, that infrastructure is much, much broader than just patient medical record information. The health of a nation is best defined by many categories. In fact, you can go back as far as the Egyptians or Hippocrates to see that the ancient peoples understood that if you wanted to define the health of a community, you look not just at what was going on and what was being treated by the physicians, but in the words of Hippocrates, you look at the air that they breathe, the water that they drink, the environment in which they live and their activities, whether or not they are given to slovenliness or actively engaged in exercise.

As such, the national health information infrastructure is conceptualized as being that piece, that vision, that allows the free exchange of information that is pertinent to each individual's health and as such, the health of the nation as a whole.

This document which we have produced is really the first step. The hardest step is the one that we need to make, based on the hearings that we have been conducting here, in San Francisco and will be conducting in Washington, D.C. in January. If we have a vision and we want to go in that direction, and by and large, the feedback we have been getting is yes, we have been going in the right direction, what are the concrete steps that need to be done.

It is not clear what the future will hold. Certainly it is not clear what the political future will hold in Washington will be. That seems to be being determined as we speak in Florida. But clearly, the role of the National Committee is becoming increasingly important as we begin to identify some standards that will serve to make this vision come into being.

As the process has gone through, we have had a high degree of cooperation between us and the Department of Health and Human Services, as an advisory committee, certainly the involvement of our staff, and made sure that pretty much we are moving in the same direction as the Department of Health and Human Services.

As such, we can begin to identify some of those standards that at least will be viable candidates for adoption of the Department of Health and Human Services as they implement the HIPAA process.

The goal of this meeting is to move forward, both the NHII as well as the work group on health statistics, to begin to identify those concrete steps that we will follow. After this meeting, the two work groups will be having a joint meeting in December, so that we can better integrate the two documents and coordinate them as we begin to move toward their release sometime in the following year or thereabouts.

So that is where we are looking from the NHII perspective. I'm going to toss it over to Dan to talk about the 21st century vision for the national health statistics.

DR. FRIEDMAN: Thank you, John. The 21st century vision for health statistics is a subset of the NHII. Specifically, it is the subset of health data that can be generalized to a known population of individuals, events, organizations or institutions, and data that are specifically used for designing, evaluating the monitoring programs and policies.

The vision for health statistics as we have defined it includes three components. First of all, we are looking at a vision, looking for a vision, which includes a broad integrative definition of health, health not the absence of disease and health not disease itself. We hope to define a vision that includes the population's health, the health care system, the interactions between the population's health and the health care system, as well as the non-health care, non-medical care determinants of health.

We are also hoping to define a vision that includes not only public data systems, but also private data systems, and that builds upon and recognizes the potential strengths of each of them and the potential synergy between them.

A second component of the vision is going to be a definition of the elements, the resources, the organizations that can help to implement the vision.

Third, we are also hoping and planning to define a set of criteria for evaluating health statistics systems currently and in the future.

The process that we have engaged in is now in just about the end of its second year. We started out trying to assess where we currently are. We started out with a series of commissioned papers, authors in this country and abroad, with a series of local and national discussion groups, as well as with a National Academy workshop.

Based upon those efforts, we developed the interim report that I'm sure you have looked at, and based upon feedback from the regional hearings and other feedback that we receive, we are going to go ahead and draft a final report.

The comments that we have received so far in the interim report have been quite consistent. On the one hand, the comments have been complimentary, largely complimentary in terms of the principles, but the comments have also pointed to a need which we recognize and agree with, and the need is for much greater specificity about how we get from here to there, as John said.

One of the things that we are hoping for in today's meeting is specificity about how we get from here to there, and particularly specificity around two of the principles that were delineated in the interim report. The first principle being principle number four, information must be available at a sufficiently detailed level to be relevant to real decisions, and principle number seven, health statistics data must be provided back to communities, community groups, local governments, and so forth in ways that can maximize data access and ease of use.

So we are looking forward to the feedback that we receive today and for your suggestions about how to move forward, and then in conclusion, let me just apologize because I am going to need to leave unexpectedly early.

So thanks for coming, and I'm sorry I'm not going to be able to stay for the day.

DR. LUMPKIN: At this point, we are going to move on to our first panel. The panelists had introduced themselves when we went around the table, so we will start off with Tom Ricketts.

DR. RICKETTS: Thank you for the invitation to speak with you about the vision toward a national health statistics infrastructure.

I come to this from a background of working with a similar effort in North Carolina, as the state began to try to move toward comprehensive health reform at a time when that seemed to be possible. Early in the '90s, the question was not if, but when, we would have some form of national health insurance or national health coverage program. In the early part of the 21st century, that question seems a little naive.

But at the time, we felt that one of the keys to developing a national program of health care coverage and a state program for health care coverage required the coordination of health data at the local, the community, the practice and across the state, to understand our costs, understand our needs, and understand where to go.

The vision that you have outlined is very similar to the vision that was developed in the data and information committee of North Carolina, then called the health planning committee, which became the health reform commission. The vision produced certain vignettes such as those that are described in the NHII vision, where there is a free and easy access, appropriate access to health information for personal and clinical decision making or personal care decisions, or for the development of policy at the local and the national level.

As the committee began to work toward practical development of this program, we found that we were confronted with many technical issues, that this committee wrestles with on a day to day basis.

At that time, we had very few standards for interchange or data exchange. We didn't even have an electronic signature authority. We didn't have a lot of the necessary information infrastructure that would even make such a project feasible.

However, we did have a lot of visioning, and a lot of visioning that raised expectations on the part of the public, on the part of politicians, and on the part of entrepreneurs, who felt that they would take a chance to exploit the expectations that were being developed by these visions.

The vision is good. The vision is at some point in the future realizable. But the vision must be managed as well as the technology must be managed in its development. The expectations of the public and the expectations of the politicians must be managed to match the reality of the technology and the reality of the capacity of the system.

I give that one only because we left a lot more disappointed than satisfied people, even with the great progress that we did make in developing the underlying infrastructure.

At the time in 1993 to 1994, our biggest challenge was data merging, and data merging of existing databases. That is, the creation of standards for existing databases for which there were no standards. These were not clinical databases, but more often they were the collected information that was gathered by government, by regulatory agencies, by the myriad groups who are charged with collecting information about individual, local and government and organizational health activity.

Just those standards proved daunting, in terms of creating what was the goal in something called the community health information network, or CHIN. That was the hot name for the kind of vision that we are seeing today. Rarely do we hear the term CHIN anymore. I think it lost its value as people understood the difficulty of creating such a thing.

In this process, as people came to us to provide assistance in the development of programs, we learned that there was great value in the data, and that there were many economic opportunities for the private sector.

In the vision for the information infrastructure, you mentioned that health data are a private resources as well as a public good. But far less attention is paid to this dichotomy between private resource and public good, and how that would be managed say for the problem of confidentiality.

The biggest problems when dealing with the balance between private resource and public responsibility is that entrepreneurs are willing to take responsibility for the management, handling and marketing of the private resource aspect of health data, where there is very unclear responsibility for the ownership, for the dissemination and for the use of the public interest aspects of health data.

At the national level, this committee, the National Center for Health Statistics, Vital and Health Statistics, have set standards and taken on a sense of responsibility for developing the things that are necessary to meet your responsibilities, which are to understand the status of health care in the United States, and progress and change that we make as we attempt to meet our needs, both as a nation and as a society.

However, as you move down the levels of government and down the levels of aggregation of population, you find less and less sense of responsibility for the coordination of standards, development of standards, and the use of data in the public interest. Many states have data coordination committees. Many states cooperate with the Centers for Disease Control and the Center for Vital and Health Statistics, but they are not the only people charged with coordinating some aspects of data.

In North Carolina, as we worked, there was an immediate conflict, interagency, over who was going to control say a new technology that was entering our vocabulary at the time, and that was geographic information systems, and how they would be handled. One state agency desired complete control over the GIS analysis capacity, while the state center for health statistics was making the greatest strides in its use for public good.

There were no rules for the distribution of responsibility and future planning for the use of those data, whereas the private sector was able to move ahead quickly in making use of the new technology, marketing it, and providing it to those people who felt the need for it.

Our problems with public coordination really become very difficult when one comes down to the personal health level. You have placed a vision for the personal data system, a virtual data entry mechanism, but there still remain some conflicts about the ownership of these data, and how they would be used again for public purpose.

For private purposes, an entrepreneurial spirit guides folks who wish to take that information that is available and use it, providing value to decision makers and to the market.

The public sector is not as aggressive in understanding the values of such data or the aggregation of such data. I believe that we need to spend some time providing the vision for the public sector as opposed to the end user vision. The public sector and those people who are responsible for the public sector and the commonwealth of the nation, the states and communities, need to be educated as to how they may use these data, rather than the gee-whiz examples that we have of personal health utilization.

It is difficult to sell the use of fairly complex data systems to politicians who don't have a particular vision for what they are seeking, and what they want to see in terms of making use of the data. I believe that we need to make more effect in educating people, that is, the politicians and policy makers at the state and community levels, as to what the uses of these data would be.

I would provide the county commissioners, I would provide state legislators, I would provide local health directors some sense of a vision for how they would use the data before I would move into promising personal health systems.

Our balance between the private resource and public interest is like the way we handle mineral rights. You all understand data mining, I'm sure you have heard the term. We license data mining in a way that we handle mineral rights. That is, you can have some sense of ownership of your property as you have some sense of ownership of your body, but you can find that that can be lost to an external organization that owns what is beneath the surface of the ground or beneath the surface of your personal lives in terms of health data.

If you understand the process by which people have exploited the mineral rights of communities in say West Virginia or Louisiana or in the Western states, you know that this is not a pretty process, when people take something that is to be shared. I think that if you recognize this problem as being real, that we should spend a good deal of time working this out and making sure that those conflicts were dealt with formally and up front.

As we moved ahead as a commission and as a committee on data information needs, we found that these conflicts and these opposing forces were never really formally addressed, or never really given a forum for developing some sense of either turf or formal division.

In your vision, you also talk about a community vision for health data. I work daily in this field, and right now am working with the division of shortage designation to try and create a new and improved measure of under service, something that is not likely to occur, but something which the demands of policy will require.

But in terms of developing a measure such as under service or a measure of community capacity, it is very unclear as to who the client is for any kind of community measure. The clients are more likely, those people who seek denominators for the use of information for investigation, or denominators for the use of information for benchmarking, for the purpose of private resource entrepreneurial dissemination in marketing.

We need to develop some better sense of the context and use for community health markers. Great emphasis is placed on this in this vision and prior visions of the utility of health data, where very little effort has been made on developing the market for any kind of public sense community indicators.

So my message to you is that it took us from the time we began in 1992 to two years ago to get the electronic signature passed, and to watch the process of the electronic data interface and interchange committees as they move forward. You recognize the time that it takes just to work out the technology.

But there is also the time necessary to work out the expectations. For that, I think we must pay as much attention to the contents and technology of our visioning as we do to the contents of the technology that will make the vision come true. That means that we must be more realistic and understand that we must develop a market, rather than just provide a vision.

Thank you.

DR. LUMPKIN: Thank you. We'll take all three members of the panel, and then we'll go into questions at that point. David?

DR. KIBBE: Good morning. My name is David Kibbe, and I am the Chief Executive Officer and founder of a company called Canopy Systems, which is based here or very close by, in Chapel Hill, North Carolina.

I am honored to have been asked to provide testimony to assist this committee to further develop the national health information infrastructure. Because Canopy Systems has several clients located in rural areas of the country, and because we have experience in the trenches, so to speak, with both the successes and barriers to success of using Internet-based software in these environments in rural parts of the country, it was suggested by the planners of today's meeting that my comments be directed towards the special requirements of rural health care and the Internet. I was also asked to suggest specific steps and actions that could be taken by the federal government to assure that the national health information infrastructure extends to rural America and Americans.

I have the sense that I am the token entrepreneur in today's meeting, so I would like to take just a couple of moments to tell you about what we are doing, and then get to some examples that I think will address quite squarely the issues of detail of data needed in these systems, and how to achieve some level of feedback to the community.

Canopy Systems is a company that designs and hosts web-based software systems for case management utilization and disease management. So in the developing parlance of the e-health industry, we are what is known as a vertical ASP, or applications service provider, combining the main expertise in health care management with software built to run on the Internet.

Our flagship product, called Canopy, is used by case managers to manage and coordinate the care of patients typically who have chronic and or destabilizing diseases, and to securely share that important clinical information with caregivers in sites throughout the community, for example, physicians' offices, nursing homes, home health agencies and not yet, but hopefully in the not too distant future, health departments.

Our products are designed in a native Internet environment and meet current and anticipated HIPAA standards for security and privacy of Internet transmission. My company, Canopy Systems, is a subsidiary of the Medical Mutual Insurance Company of North Carolina, which I mentioned only to make sure that you understand our heritage is that of the North Carolina Medical Society. We are quite literally a child of the North Carolina Medical Society as a company, and wholly owned by the physician policy holders in the state of North Carolina, which makes us I think the only company of our kind that is owned entirely by physicians.

There is little question that rural populations have very significant health and health care delivery problems that I believe can be addressed by the application of Internet technologies in these communities. As you all are well aware, the Agency for Health Care Research and Quality has documented through its work that almost one in three adults living in rural areas of America is in poor to fair health, and nearly half suffer from at least one major chronic illness. Traumatic injuries are much more common in rural areas, and residents face worse outcomes and higher risks of death than urban patients, partly because of transportation problems and lack of access to advanced life support training and emergency medical personnel.

Low population density in rural areas makes it inherently difficult to deliver services that target persons with special health needs. Groups at particular risk include the elderly, the poor, people with HIV and AIDS, the homeless mothers, children and adolescents, racial or ethnic minorities, and persons with disabilities.

Many small and rural hospitals have closed and are continuing to close, while other health care facilities are in dire financial straits. The supply of primary care providers in rural America is actually decreasing, adding to problems of health care delivery access in rural America.

The Internet as a low cost vehicle for the distribution of information and information management services can play a role in assisting health care professionals and their patients to make better decisions regarding their health, to coordinate care among diverse and geographically separated provider sites, to help avoid errors due to this fragmentation, and to help us all understand, while certain individuals and groups of individuals choose to use particular resources and choose to avoid others.

At Canopy Systems, we have seen the benefits of the Internet in rural areas in real time, so to speak. In 1998, Access Care, which is a 125-physician IPA in 25 medical practices, some in very rural parts of North Carolina, asked us to design and build a web-based information system that manages information on over 85,000 children who are Medicaid beneficiaries in the state of North Carolina.

This web based software, rather imaginatively named Patient Tracker, is used by care managers in 25 medical communities to study, record and report on the experiences of the patients in this population who visit the emergency room or are hospitalized. Case managers in each one of these communities access our software via modem connection and browser, using the system to check patient availability status, previous admissions, medications and other clinical elements.

They also track patient visits to the emergency room. This information collected in a single database is then used to generate regular, sometimes weekly, reports which are studied by access care managers, physicians and nurses, to understand their organization's trends and utilization, to compare rates of hospitalization for specific illnesses among the several practices and several doctors involved, and to test the impact of patient education and extended facility hours on lowering unnecessary ER visits.

We also designed and built a web based asthma disease management component of this system, that links reminders and evidence based guidelines with asthma management and educational tools.

We are told that our Internet based system helped Access Care and the state of North Carolina Medicaid program save several millions of dollars over a two-year period through care management. Although they did not measure patient satisfaction or service quality metrics of this program, many of the case managers involved reported numerous stories and shared anecdotes with us about how their interventions and their timely access to information to their patients' previous experiences made a large difference in the lives of patients and families served.

Another of our rural clients is using Canopy, which is our second generation web based product for a continuum based case management project, an innovative program for the frail elderly in rural Northeastern North Carolina. This program, like TLC for Transition Life Care, is supported by a grant from the Duke Endowment as a joint effort of Halifax Regional Medical Center and Rural Health Associates, a group of five rural clinics, that also owns and operates a rural nursing facility.

TLC's Canopy links nursing home personnel, hospital nurses and physicians with the web based medical records of over 100 frail elderly and nursing home patients. One of this program's components under development is a website that helps direct patients and families to community agencies and programs for the elderly in these three rural counties.

Dr. Jane McCaleb is the person most responsible for the success of this program, and someone who can rightly be called a physician champion of rural health Internet initiatives. She considers the information in the web based system to be part of a triad involving information, intervention and interaction, which she feels must be in what she calls a dynamic balance in order to optimize patient care and to anticipate future decisions among the frail patient population for whom transitions are to be expected.

Dr. McCaleb was also instrumental in the design of the TLC version of our product Canopy, including the sections maintaining up to date information on patients' advance directives and alerts for patients on multiple prescription drugs. Her system is now being integrated with the case managers who use Canopy on the in-patient services of Halifax Regional Medical Center.

Perhaps the most helpful comment I can made regarding my company's experience delivering web based information systems in rural areas like this to date is how successful they have been. That is, how few complications and barriers we and our clients have encountered. These are not unsophisticated software programs. We are not talking about web portals here. Canopy is a multi-tiered software program written in JAVA HTML and X-HTML. It is servercentric and it is built with the security and privacy guidelines of HIPAA firmly in mind.

The security subsystem alone is very complicated, involving client size, strong authentication via digital certificates and industrial strength 128-bit encryption or higher. The product provides usable audit trails. It is optimally run via high speed connections, either DSN or partial T-1 lines, and the software programs have over 70,000 lines of code.

Yet, even in rural areas, where case managers and providers have had to connect via telephone modem at relatively low speeds, they have found value in using these web based systems. For the most part, we have been able to negotiate the intricacies of third-party digital certificate authorities with minimal difficulty, and training of the providers has been much less time consuming than anticipated.

This last point I think bears some mention. Our training experience has been very positive, and I think is due primarily to the universal familiarity with the web-browser interface. Even somewhat computerphobic nurse case managers in rural areas have found our software easy to navigate and familiar because of the browser interface.

What have been some of the problems and barriers to overcome? Clearly one problem for us has been the quality of the telephone lines in some rural areas of the country. We and our clients would prefer to see the same dependability, reliability and modernity of telephone equipment that exists in urban and suburban areas extended to rural areas. It is a very practical issue. We have also had occasional problems with small local Internet service providers who may not always have the capacity needed to support rapid growth in their customer bases.

Finally, there have been occasions when just the idea of the Internet posed problems, due to lack of understanding about security or fear that the Internet is inherently less safe than the hodgepodge of paper and unsecured server systems that predominate in health care organizations, particularly in hospitals.

Every contract to put our software into operation has to go through a lawyer's office. It might surprise you, or perhaps not, to hear of the variety of opinions being expressed out there by legal firms regarding use of the Internet for health care information. This is not exactly their strongest suit. However, I want to emphasize that in large part, we have been able to overcome these problems, albeit much slower than we would like to, at times. We have found that some of the programs that we have been supporting literally get delayed months by this issue of lack of understanding about the security in the legal firms.

What can the federal government do to promote the quality use of online health information management in rural America? I would suggest that there are a range of activities that the federal government could take and that this committee could encourage to improve directly and indirectly the capture, storage, communication and processing presentation of health information for the benefit of individuals, providers and communities located in rural areas.

Here is my short list of recommendations. The federal government, perhaps in concert with the state government and other agencies, should find ways to support care coordination programs that bring together rural providers, including hospitals, physician offices, health departments and long term care facilities with the goal of improving the health status of particular groups of patients such as the frail elderly, those with specific common diagnoses such as diabetes or congestive heart failure, and those who are disadvantaged with regard to economic or social standing such as those who have no insurance.

Best practices and collaborative efforts that show promise should be showcased and their leaders made available to publish the results. I firmly believe that the value of the Internet as a means to overcome the problems imposed by a fragmented health system is becoming so obvious and well known, that coordinated care programmers and their planners will by themselves invent the rural health information infrastructure and find Internet based solutions on their own, if given the opportunity.

Unlike any other information technology to date, the Internet combines an open architecture with very low marginal costs to sustain its use, making it particularly attractive in rural areas. What rural health care providers need is seed money and models for planning to get things started.

I would echo Tom's remarks about the CHINs. One of the problems with CHINs of course was technological, but another problem with the CHIN was the vision. I believe it is the care coordination around definite population of patients that will provide the glue to hold these provider organizations together. Otherwise, the political and economic portions of the rivalries will win out. We have to bring these models to these communities that are successful and show them how they can create win-win-win-win situations.

Secondly, I think the federal government should encourage grants and studies in rural areas that examine best practices for Internet security and privacy, for example public key encryption infrastructure and HIPAA implementation in rural areas. These are rather obvious, but I think they need specific focus and grant monies.

There is a lot that can be done to help educate leaders in rural populations as to the value of Internet technologies and to overcome the fear or lack of knowledge I alluded to earlier. Unfortunately, I have heard on several occasions health care leaders in rural areas express the attitude that, that is fine in Chapel Hill, but the rural folks in my part of the world are never going to use the Internet. These same leaders are sometimes very surprised by how many children are online right in their own backwards, in chat rooms and so forth.

The digital divide cannot be allowed to be a self fulfilling prophecy in rural areas. We need to think about how to overcome this kind of bias with education and public relations campaigns. I would like to see the federal government support informatics educational grants for rural health care professionals of many different kinds, and I will volunteer my company's involvement in helping those be successful, if so asked.

Four, we need to develop demonstration projects using the Internet for rural provider networks that show sustainable promise of delivering value in the day to day world. I have been a rural family physician, and I can tell you that it is a very daunting task. There are policies and practices that often discourage continuity and efficiency, and without access to informational resources that could make a difference in my practice.

We cannot expect success to occur if we simply add another three hours work at a computer to each rural physician's day. Combining informational systems that work on a new level of capability which we already have with innovative forms of reimbursement and financial incentives is really essential if we are going to move beyond isolated successes.

Thanks for this opportunity to talk with you. I look forward to any questions and discussion.

MS. CURTIS: Thank you. My name is Janis Curtis again, and I am going to do a little bit of a different format than my former colleagues. I have some specific points I want to put out on the table this morning.

Let me also give you a little bit of my background, so you will have a framework for understanding the comments that I am going to show you this morning.

First of all, I had the opportunity and privilege to chair the data committee that Tom talked about a few minutes ago that the state had as part of the health planning commission a couple of years ago, so I clearly understand and support the need to have a vision. You need to know where you are going in order to get there; it just doesn't happen.

I have had the benefit of augmenting my experience of the benefit of a vision with real-world experience since I have been at Duke for the last nine years.

Currently, I work in the information systems department for Duke University Medical Center and Health System. I have responsibility for budget preparation, HR responsibilities and also strategic planning.

Several years ago, Duke University Medical Center embarked upon the integrated delivery system strategy, the way many other institutions have around the country. We purchased two community hospitals, one in Durham and one in Raleigh. We also purchased 14 primary care group practices. We also had a vision that within the health system, having ubiquitous connectivity and systems and applications that interfaced appropriately and seamlessly and would get us to where we wanted to be. Again, a vision is good, but where the rubber hits the road is when you become very knowledgeable about the challenges that you have to confront.

So what I wanted to do with the time that I have this morning was to go through a couple of bullets and give you my reaction to the document that was presented. What I would hope is that you would ask questions after I finish, because it is easier for me to respond to specific questions and give you the perspectives that come from my real experience at Duke.

I wholeheartedly support the vision. The first thing that came to my mind as I read through the report was the identification of cost categories identified in the report. There were references to network costs, systems costs, application costs. One of the things that I have become quite knowledgeable about at Duke is, once you get the capital to invest and implement a network or to have applications developed or to buy a particular application product, or even to support the development of interfaces between those applications, it is the operating costs that kill you. We generally allow about 20 percent, or take whatever the total cost is to implement something, and establish or estimate an additional 20 percent for ongoing operating costs. Operating costs really sometimes is the sleeper. Many times when we have departments from Duke Hospital come and ask for capital, they generally underestimate the operating expenses associated with supporting that particular system once implemented. So one thing to keep in mind is the ongoing operation costs, because if you get the capital to install and implement this infrastructure that we are talking about, you want to be able to have the funds to support it after the fact.

DR. LUMPKIN: Janis, I'm sorry, is that 20 percent per year.

MS. CURTIS: Yes. That is a broad -- sometimes you're going to need increased FTE, sometimes you are going to need to replace hardware, sometimes you're going to need system enhancement, or whatever.

The other area that I wanted to pick up on that Dr. Kibbe referenced a few minutes ago is training of people. I think that many times, we underestimate -- us technicians and visionaries, we all envision, once we get the technology in place, you build it and they will come sort of a deal. That is really not the case, and that is not our experience at Duke as well.

We do have a number of clinicians and providers who -- and when I say providers, I mean the actual people who provide hands-on care, who are computer savvy, but not all people are who are involved in the patient care process. I think inherent in this vision that we have is that you capture data at the point of care, and then you progressively build on it and aggregate it, to the point that we can share it back and forth and have this huge system.

The problem is, we overlook the fact that people who are actually responsible for capturing Internet data are not always trained. They are generally the low people on the totem pole, and they are a linchpin to having good quality accurate data to be able to do something with it. It doesn't make sense to have a system where you can transmit inaccurate information. You want to be sure that you have accurate information. So training of people is a very significant aspect of implementing this vision.

When I wear my public health hat, I also wanted to make a comment that when I read this report, I took away an inherent bias in the report towards people who are computer savvy. There was reference in here that we needed to make sure that we had ubiquitous access to computers. That is true, but that is only one aspect of it. People have to feel comfortable using a computer and know that they have a responsibility to share information or to allow people to have access to that information. It is not just technology and training of how to use the technology, but to educate and train people about what their responsibility is, to make sure that the vision that we put forward in this report is realized.

Another point that I wanted to make has to do with the rule of vendors. Just to give you some idea, at Duke we are truly a multi-system, multi-vendor environment. I can't tell you how many different systems that we have. We spend a lot of time developing interfaces to be able to share information back and forth between the systems.

We recently purchased Raleigh Community Hospital. Raleigh Community had been a single system vendor and they had vendored what is called Medi-Tech as their hospital information system. Through no fault of theirs, but now that they are part of the Duke health system, we are making them become a multi-system, multi-vendor environment, and quite frankly, they are overwhelmed right now and they can't handle that. So I think we need to be cognizant of, as we move forward in terms of trying to implement this vision, what burden that puts on the smaller provider.

I appreciate Dr. Kibbe's comment about rural providers, but it is really all small providers, all small hospitals, all small physicians, et cetera. They do have computers, but they don't always have the level of computer savvy necessary to support a system once a system is put in place, or they hook up with additional systems.

The reason I brought that up is because we at Duke -- and I know that this is probably true and reflects many organizations across the country -- we are trying to move away from development of applications all the time, and we are trying to buy applications, only allocate resources towards development when it is going to give us competitive advantage, or it is something unique to what we have to do.

Taking that into consideration, you have to acknowledge the significant role that vendors play in terms of having to bring about that realization. The philosophies and the standards and the discipline that they bring to developing their own applications, there has to be some uniformity to easily share information back and forth between different systems and applications. So while we can have the need to do it at a provider level, sometimes our biggest struggle is with the vendors, whose systems we are buying, and getting them to make sure they can assist us and make sure that those applications interface.

When I read the report also, I picked up on the level of detail that is needed to achieve the objectives. I think, Dr. Lumpkin, you had mentioned something earlier in your introductory comments about the level of detail needed. I think we have to be very careful to balance the data that we need in order to achieve this vision and the burden that that puts on the provider to provide that information.

Many times, you will find that the systems that exist in these provider organizations, their data structures, their data systems, et cetera, that are comprised of data elements that they themselves felt were important to meet their objectives. So as we identify the need for additional data, we need to make sure that we are not overburdening those providers to capture this level of detail and extent to which they can benefit from that, as opposed to just meeting what our larger need is.

There is this need for money also to make -- or, I would like to encourage you to acknowledge the funds that are going to be required within institutions to make the changes to the internal systems that are going to be necessary to provide data and that we will decide that is needed, regardless of what that particular data set is.

Again, this is from experience at Duke. We believe that because an organization has computers, that all the needed data is computerized. It has been my experience that that is not always the case. Some providers are going to require additional support to help take data that exists, but it exists in manual form, to convert it into an electronic format to be able to participate and exchange this data electronically.

The key role of standards. I won't dwell on that, because I am sure that you have heard that in different hearings across the country that you have had. All I can say to that is, amen, we need standards, and I think the extent to which we can identify those critical standards that are required and adhere to them and endorse them, the better.

When I read the report as well, again, I completely support the vision, but also, the vision was overwhelming. I wanted to pick up on some comments that Tom Ricketts made earlier. We had a vision that came out of the state work that we did, and fortunately, where we are now, a lot of advances in technology have occurred that put us in even a better position to help realize that vision.

But I think it is going to be important as we go forward to acknowledge that we have to break the vision down into manageable pieces. I think that Tom made an excellent point earlier about how important it is to manage expectations. This is something that we have to do at Duke all the time, manage expectations of the -- and I'm saying physicians now, because a lot of what we are doing really relates to clinical systems, to make sure that we appropriately set their expectations about what technology can do, and what the applications can do to help meet their needs, just in general, but also out of the starting block. Many times, you implement a system and it takes several years of enhancements and customization before you get to the point that a physician's needs are adequately met.

I think it is important to begin to break the vision down into manageable pieces. The other thing that is important to recognize and get across to people who are going to be significant in helping to realize this vision is that all things are not equal. Certain things have to be done first before you can do other kinds of things. The extent to which we can break it down and begin to prioritize, we can also then make sure that the dollars are appropriately channeled to make sure that those foundation pieces are addressed first before focusing on some of the high end.

I think in one of the scenarios, there was a reference to a person walking in, and all this electronic automatic sharing of information, and your Palm Pilot and your PC and your this and that are all in sync very quickly. While that is a great vision, if I had a limited pot of money right now, I would opt not to focus my money there, and focus on some of the fundamental areas, like trying to make sure everybody can convert the minimum level of information from the menu format or from a hard copy format into an electronic format.

I had a conversation briefly before this hearing started with Dr. Deering about the return on investments and infrastructure. Again, drawing on my experience at Duke, it is very, very difficult to ask people to give you money for infrastructure. It has been --

DR. LUMPKIN: Did you say money for instruction?

MS. CURTIS: Infrastructure, technical infrastructure. It has been easier to this point, because as Duke pursued the business strategy to create an integrated delivery system, the need for a network to connect all the hospitals and all of the primary care practices was a given. So it was relatively easy to say, in order to do that, we need this money to extend the network out, we need these routers, we need these X, Ys and Zs.

But as that infrastructure has been implemented, it becomes more difficult to convince people how you have to maintain that infrastructure. So I come back to my support comment at the very beginning. People believe that once you spend a half a million dollars for routers, et cetera, that once you spend it, you don't have to go back and update those routers or replace those routers.

So I would encourage you to build in the notion of a scheduled replacement of that infrastructure, so that it is not all at once and it doesn't hit you all at once.

What else did I want to say about infrastructure? Just that it is important. I go back to -- I have always had a bias in favor of networking and networks. Many times I have had conversations with people who are so interested in application development, and believe me, I believe in applications. But they are so focused on the bells and whistles, and all these features, sometimes they omit the fact that in this distributed world, if you don't have a reliable network, it doesn't matter about the bells and whistles in the applications; you won't be able to get to it, you won't be able to transmit the data that you need. So I really want to emphasize how important infrastructure is.

Just a few more comments. Picking up on Dr. Kibbe's comment about level of computer savvy, he had made reference to some of the providers taking to the use of his application because of their familiarity with browser technology. I would just underscore, that is true. We always make reference, kids can do more than what their parents can.

The bottom line in terms of our experience is that providers want to provide care. We need to make sure that when we put computer technology out there to help them, that we also recognize that the more they have to learn, or if the application itself is cumbersome, that it is taking them away from what they are supposed to be doing and what they are interested in doing, which is providing care. So we need to recognize the level of computer savvy out in the community.

I had been asked also to comment a little bit about what government can do to help make this a reality. Actually, Dr. Kibbe made a couple of comments that I would support wholeheartedly. One is to fund proof of concept projects and pilot demonstration projects. Again, that is consistent with the notion of breaking this vision down into manageable pieces, and making sure that where we invest our money long term is something that has been proven to work.

The other area is to continue to promote and in some cases to develop the policies necessary to support the technical solution. Again, sometimes the technology is much further ahead than the supporting policies that are needed to take full advantage of that technology. I'm sure you have heard that over and over again.

I think the experience that we have right now with security and privacy is an excellent example of that.

Those are my comments.

DR. LUMPKIN: Thank you. Questions? Well, I've got a bunch. Jeff, go ahead, Jeff.

DR. BLAIR: While you're asking your question, I'm going to think through so I can articulate mine a little more plainly.

DR. LUMPKIN: I guess my first question goes to David and Janis. That is, to what extent do the systems that you are implementing replacing paper, or are they in addition to paper?

DR. KIBBE: I'll take a shot at that.

DR. LUMPKIN: Please.

DR. KIBBE: Canopy is a system that helps -- it is an ASP model, a remote software application that requires nothing of the client other than the browser and access to the Internet. Utilization management, disease management and case management done in most health care organizations -- hospitals, integrated delivery systems -- is now a combination of paperwork and electronic systems. Data is taken off of machines by case managers, by utilization managers, put in paper format, sometimes scribbled down, then put back into computer systems. In some very small communities it could be 100 percent paper, although even there, there is some information coming to the case managers off of the patient accounting systems.

In some organizations, it is much more complicated than that. They may be using three to four systems simultaneously over the course of a day, as well as interacting with information systems that are not in their organizations, such as the systems that exist in insurance companies. So it is really a very mixed bag.

We think that that is a real problem in terms of utilization in case management processes in hospitals regarding HIPAA's security and privacy components. These are activities that are by definition very insecure. Information is taken off of a computer without any session or audit trail, then it is put down on paper, put in a box. It is open to the public. Susan comes along later on, writes it up, hands it to somebody else, somebody else puts it back into a computer.

So I think that one of the things that we are helping our clients do is envision a redesign for case management and utilization management, which involves working entirely within a ASP software application.

MS. CURTIS: Where we are at Duke right now, we are in transition. Our intent is to move away from paper, but that is something that takes awhile.

One of the things that we have done is, we have developed what is called accommodator repository. It is not a data dump for all clinical information, but it is a place where we draw data from other -- key clinical information from other department processes into this accommodator repository.

In addition to pulling data from existing systems, what we are also working on and have implemented as a matter of trying to get it out to get many of the doctors using it is, we have an ability to, once they do dictation, to automatically import that electronic record into the accommodator repository. Then we build the applications so that physicians can look online and look at the transcribed dictated reports, they can look at some of the key clinical information relative to lab, radiology images, et cetera.

We are in like I said transition right now, because while we developed this to help physicians to provide care and clinicians to provide care on the spot at the point of care, there are still the JCHO requirements, et cetera, that require some stuff to be in paper. So our intent is to eventually be at a point where it is only electronic, both in terms of meeting the patient care needs, but also meeting whatever accreditation needs. We're just not at that point right now.

But that is again what I was saying, where you want to allocate your resources for development, those precious resources. So that is where we are pushing a lot of our development right now, is to make it so those documents can become electronic from the very beginning, and get in that accommodator repository.

DR. LUMPKIN: You just said something which I think is the first time we have heard this, about accreditation requirements being a barrier. Maybe not at this meeting, but I would appreciate it if you could send us by e-mail or whatever any of the specifics on those specific requirements that are barriers.

MS. CURTIS: And I don't mean to be nitpicky; rather than referring to them as barriers, I would say they are requirements that we have to take into consideration. I am really being nitpicky, I realize that.

DR. LUMPKIN: Well, but you can dance around it, but I would particularly say that to the extent that we have federal, state or accreditation organization requirements, which we certainly do have in Illinois, that mandate paper, that is a barrier. Particularly in the environment after the federal legislation on digital signatures, we should have an environment in which we can move away from that.

A recommendation from this committee may highlight those, and prompt those organizations to use an alternative way of meeting certification. I think they would be open.

MS. CURTIS: That's fine. We have JCHO coming up, so I don't want them to read that I made reference to JCHO as barriers. I'm trying to be good here.

DR. RICKETTS: This is probably unexpected by David, but I am the evaluator of some of the programs that he is involved with providing the data system support to. The Canopy system is seen as a good way to assess the clinical efficacy of some of the reorganizations and the programs that are going ahead.

However, the implementation that I have seen is on an experimental basis by one of the facilities at least that is doing this. In other words, they want to see what their paper system yields, as well as the Canopy system.

So if you were to ask does this replace -- generally, the transition I would say in medium to large institutions usually does not replace that. There is usually an experimental and high cost implementation period.

The results of these kinds of comparisons need to be distributed much more rapidly and broadly to avoid people having this duplicative cost. But I would say the standard operating procedure is to duplicate data entry at least, if not analysis for the clinical decision making.

DR. LUMPKIN: And that is standard procedure even if you have a paperless system and you're moving to another one. You've got to have the new system prove that it is actually working before you stop the double data entries.

Jeff, did you have your question formulated?

DR. BLAIR: Thank you. I guess it was this last weekend when I got my reading done, I ran across an announcement that was especially thought provoking. Like many of you, I have been immersed in all of the implications of moving towards the Internet and all the technologies that are associated with that, and trying to use application service provider technologies to be able to reach out to where Larry is and smaller group practices in cost efficient manners.

Here is this article, and I don't know more than what I am about to repeat to you here, so I'm going to ask you for your insights and perspectives, and help us interpret this.

The announcement was that the American Hospital Association in response to frustrations and complaints with health care providers moving towards Internet technologies, has -- in particular it was concerns about confidentiality over the Internet, reliability over the Internet, and limited data rates over the Internet. They have decided to provide an offering of a value added network at low cost, which would be able to provide network capabilities that would inherently overcome those limitations of Internet technologies, at least the way they exist today; we don't have Internet two yet.

I thought, my goodness, with all of the things that I have been hearing for these last several years about how revolutionizing Internet is going to be, for the AHA to make this announcement at this time, it was a surprise.

So could I please have your insights into this? Is this something we should just consider to be a minor incident on the road to the future, or is this something we have to really consider?

MS. CURTIS: Just for clarification, minor incident relative to AHA taking that course?

DR. BLAIR: I'm sorry?

MS. CURTIS: A minor incident relative to AHA taking that particular course of having a VP in, virtual private network?

DR. BLAIR: Yes. How should we understand this event?

DR. RICKETTS: I think that it is happening far more often than just AHA. This is the way in which institutions and organizations can realize the value of electronic intercommunication, and who have the capacity to set up standards and allow it to expand or grow to their needs.

The fact only is that it interfaces at some point with the Internet that makes it maybe seem like it is different because of its containment. But this is more than likely the way in which things will go as systems find that they can improve connectivity within their controlled environment, and control throughput and control the training up and the standards across that. It only makes sense for any organization which is trying to get value out of its communication system to go this way.

The Internet standard will be roughly equivalent, but I know that we are having difficulties now with Internet standard and some of the new virtual networks within our organization that has to communicate with a number of groups. LINUX is not solving all the problems.

DR. BLAIR: Are you saying that this is complementary, and not a different direction?

DR. RICKETTS: It is complementary currently, as people work from an essentially common standard. But standards are diverging, not converging.

DR. BLAIR: Really? Can you give us some examples?

DR. RICKETTS: Some of the more imaginative adaptations of LINUX operating systems or in some of the user processes that people are developing, the JAVA alternative that Microsoft is working with, have difficulty with inter-operation on an Internet standard. Anybody who does the Internet and knows when you had frames incompatible systems that this next development is creating some divergences that are going to have to be worked out.

So I think you can see it yourself. To work into a system that is say infrared based in an organization that is firewalled out, they will have built a specific translator. That is our system; we have built our own translator and our linkages to manage ourselves within our LINUX environment. That is the kind of thing people will have.

So I think that there is a potential for divergence, as people find ways to improve for the unit in which they control the information, and to make things work faster for themselves. They are not worried that maybe two steps away, there is a little bit of incompatibility that will sometimes translate into inability to use across platform.

I am not a technical person. I am the chairman of our IT system, and I am on the committee for the IT system for the University of North Carolina, or at least was until I made such a nuisance of myself. But now I know that these problems have cropped up constantly.

DR. BLAIR: Can I also get reactions from David and Janis?

DR. KIBBE: Sure. I take that action and announcement by the AHA as very positive, because I think that they are recognizing that they have to supply -- they have to get in the mix, so to speak, with their member hospitals, finding solutions for creating local Internet solutions for community based systems.

There is a hidden secret in all of this. It is sort of a dirty little secret that people don't want to acknowledge, although I think I acknowledged it earlier. People in health care really don't want to share information a lot of the time. I think that any community we have worked in, we have had to deal with the fact that people see their own information -- that is, providers see their own information in very proprietary terms. As they start to think about who might be sharing or seeing that information, they get very worried.

On the other hand, the consumer health movement is pushing very, very hard as we all know towards more sharing of information. Ultimately, I think people who have illnesses, people who are seeking information about illnesses, people who are seeking information about their own treatment, are going to push everyone in the direction of finding out ways to communicate information.

A lot of people don't want to do that still. I think that the hospital industry in general has been very tentative up until now, thinking about ways in which they would acknowledge that the Internet is going to be a set of technologies that are used in this way. I think what they are doing is acknowledging this, and trying to put their own brand on it, basically.

I do believe that there are some divergences developing within this environment that have to do with languages, like JAVA for programming, that have to do with standards for presenting data, HTML and so forth. I think those are really very minor glitches. I don't see those standing in the way of the momentum that is developing towards using the Internet in all of these environments as a mechanism for not just displaying information, but transferring the tools for managing that information into various locales.

One of the problems that the hospital industries have -- I have visited about 60 hospitals in the last year, most of them 200 to 400 bed institutions. I have not been in a single hospital that didn't acknowledge that they have a total mess with their information systems internally. This is not the place to go into the details of what that mess is, but there is a real feeling of just being overwhelmed, not having enough people to manage, too many systems. They are beginning to look for opportunities to outsource as much of that as they can.

Of course, outsourcing is a real double-edged sword. This is true particularly in health care. So I think we are going to see an ongoing set of moves and countermoves and technological changes that -- but they will inevitably move towards a more outsourced information system infrastructure for hospitals, because that is the only economic solution.

DR. BLAIR: Janis?

MS. CURTIS: Clearly, this is looking at it from a practical standpoint. I view VPNs, virtual private networks, as one option that is worth exploring and considering.

I think Tom's points were well taken. All options now always have to be considered relative to security and privacy. That is just a given. We find ourselves in a situation of looking at VPNs because it was a very cost effective alternative for us.

I had mentioned earlier that we have two other hospitals and 15 primary care practices that we owned. We were under the gun to get them connected to our common services network. For the majority of them up the starting block, we did 2-1 lines. The cost is killing us.

What happened was, at that time also, there were significant concerns associated with the Internet. Here we are, several years later, when Internet has improved to a certain extent. What the virtual private network is, is a sort of subset within that. It becomes a better cost alternative option for us.

When we were looking recently at connecting some clinic sites, we had to look at the cost of capital, how much it was going to cost, they needed work stations, they needed these applications, they needed a printer. We needed to work with the local telecom company, et cetera. The cost really started to be enormous even though there was a business need. So we started looking at VPN as an alternative, and it turned out to be much cheaper to go that route.

Again, we are still evaluating the security concerns. But the way I look at right now, it would be inappropriate and short sighted not to consider VPNs. It would be irresponsible to completely embrace it, but I do think it represents a viable option that warrants close study.

DR. LUMPKIN: We have only got time for one more question.

DR. SONDIK: Tom, you said very interesting things. All your presentations I think were very useful. In terms of having all of this process make a difference, you said some things about working with decision makers vis-a-vis their expectations, not overselling and managing expectations.

I wonder if all of you could briefly address how you think we should go about that, we as the country should go about that, beginning at the top, with making the kinds of investments that we need in this. Or perhaps we begin the other way. Perhaps we begin at the local and the state level to make these investments.

But you have raised a real concern for me about overselling, if you will.

DR. RICKETTS: Just briefly, I think that that problem does exist in communities where people are being asked to either purchase systems or to give the go-ahead or to support these systems, not really knowing what they entail.

The problem of how privacy came up and hit us squarely in the face and was very much a concern meant that there was very little basic education about privacy problems and the use of medical records, and their actual flow of how they were used. So we needed to educate people who were in some position to make either a cooperative decision to allow their system to join or to purchase or to work with even supervising a county health department to have concerns at that level.

Then there are folks who are appropriating or allowing the coordination of say vital statistics at the state level and how they can be used for the benefit -- for public policy benefit. I think that this is a situation where best cases should be presented to the folks, to see how some decision processes work through when data were available at these levels, so that they can see how it really works.

I think people are moving into an unknown land, and they would like to get a sense of what happens and where they are going, rather than what we are presenting to some people. We are presenting an ideal vision on the one hand, and then we have doomsayers on the other hand who are presenting problems of ubiquitous health information availability that is going to ruin peoples' privacy. We really do have the in-between of, here is how it worked for this situation here in this community. Those situations are the best things for folks to get a feeling of, what am I dealing with.

But then you also have to anticipate who these policy makers are and these decision makers, and that requires you to work at the practical level, of who is doing things at what level. Often an information system will start at the top and disseminate across a system that is not usual and normal for policy dissemination, and it should parallel the normal policy things.

So it is usual common sense policy systems that would be the people to target.

DR. KIBBE: My comments would be very similar to that. I think we need to have a very patientcentric focus. In other words, the real benefit here from managing information better is to manage patient care better. The tactical difficulty, the expectations I believe that have to be managed is that we have way too few examples of how managing information better helps not only the patient, but helps providers, helps the hospitals, and helps the community.

It takes a community to manage a very sick person. It takes a community to manage children who have significant chronic illnesses. It takes a community to help elderly people find the right resources that allow them to die in dignity. I believe that what we need are to take examples where information systems are used not only to improve the outcomes of patients, but are used to actually collect the data that can then go back to the physicians and back to the hospitals, and merge with claims data and cost data, and present those to community leaders in a way that gets them to buy into that vision.

I think this can be done almost as a PR campaign, provided you don't mix apples and oranges. Those things that are going to be potentially relevant and exciting to a community hospital with a nursing home, 400 beds, 58 physicians on its medical staff, et cetera, are very different perhaps than the situations that would be exciting to a major medical university.

So again, in line with perhaps what Janis was saying, we need to pick through what projects are meaningful to what kinds of constituencies, and then go about the business of convincing them that this works.

MS. CURTIS: I certainly don't disagree with anything that Tom and David have said. The one piece I would offer is, have demonstration projects. But I think it is also important to communicate when those projects are being described to people, to get them to understand that there is no one solution.

An appropriate solution for a community locale is one where you have a corps that abides by certain standards or adheres to certain principles or achieves a common set of objectives. Above and beyond that are features, functionalities, policies, procedures and whatever that are tailored to meet any particular community's problems and challenges.

I think one of the things that happens is, when Tom comes to my community and he says, we use ABC system and X, Y and Z, the first thing I'll say to him is, yes, but my community is different. We have X Y and Z, or we don't do blah, blah. What happens is, I reject it out of hand, because there is some piece of what he did in the solution that he adopted for his community and I'm saying it is not going to fit in my community. If I were to tease it back and I find out, here are certain aspects of that solution that will meet some of my needs, and then that other quarter, customized or tailored features or functionality or whatever, policies, to meet Tom's community's needs I can do the same thing to meet my community needs.

So what happens is, it becomes a solution that is towards the common good, but yet is amenable to meet my unique needs. I think what happens is, many times we fall into this all or none framework or way of looking at something and it is like, here is the solution. What happens is, people reject it out of hand because they have their own particular challenges that they are dealing with. Then they can't see the forest for the trees anymore.

So I am in support of the specific examples of how things have worked in particular communities, as long as it is explained in such a way that people understand, there is a certain part of that that was customized to meet Tom Rickett's community's needs, organizational needs or whatever, and that I have that same ability as that solution is applied to my environment.

DR. BLAIR: Should we draw the conclusion from what you are saying that it might be well if we look at a national health information infrastructure that we have to build support for this from the bottom up, instead of trying to come out with solutions and ask people to react to that? Is that what you're saying? That we have a better chance to --

DR. CUSTER: I'm a firm believer in working with systems from both ends, I really do. I think that there is a benefit and role that the federal government can play in terms of articulating that vision and why it is in the public interest and the best interest of public health to move forward in a particular direction. The reality is, health care is delivered locally. To achieve this objective the strategy involves strong support and active participation at the local level. So it is something that you can't force down, you can't mandate it, but at the sam time if you wait for all the local folks to make all the right decisions, it probably will never happen.

So I am a firm believer in the synergy between the two, and you kind of have to work both ends at the same time.

DR. RICKETTS: Could I just -- I want to get this in because it is something I have been working on. I said, to give the positive vision, but one thing we are trying to use David's work on is, you are probably familiar with the term ambulatory sensitive condition admission rates. Basically, it is when a person goes into the hospital for something they probably shouldn't.

When we look at those, and we say what are the reasons for that, we find that the most common are social services, pharmacy, internist and specialist. That is the kind of local value that people can be demonstrated. Then you can sell them what will give them that information.

I think you have to sell folks the need to communicate first, before you sell them just the information that comes out. I think we have been selling information before we sell them what they want, and they have to be shown what they want, which is how to fix this problem that I don't know how to fix.

DR. LUMPKIN: I actually feel like I might have to seek professional advice from Richard, because I am of two minds on this discussion. I think it has been a very useful discussion, and part of me wants to keep on, but the other part which is charged with keeping us to the schedule, which I have done a lousy job of, says that we have to end and move on to the next panel.

We have gone through our break and way past the break, so we're just going to push on. I'd like to thank the first panel for coming. I'd like to suggest that maybe if we do have the opportunity, if you aren't pushed to immediately leave, I certainly would like to continue some of this discussion maybe over a working lunch, since there isn't a lot in the neighborhood around here for people to go. But if we all have to leave, we appreciate it, and those of you who can't stay, we also appreciate that.

So thank you, and if we can go on to our second panel, if the panelists will introduce themselves, starting off with Chris Mansfield.

DR. MANSFIELD: First of all, I have to say I am definitely honored to be here, and a bit baffled I am here; I certainly had a difficult time finding it.

DR. LUMPKIN: Could you introduce yourself?

DR. MANSFIELD: Yes. I'm Chris Mansfield, and I am an associate professor at East Carolina University in our Department of Family Medicine and director of health services research and development.

I have been involved with data I guess ever since 1967 when I was a VISTA volunteer and had to become a health planner, to convince powers that be that we ought to have a community health center. So I learned a little bit about community epidemiology then, and then went on to direct that -- and later become a real health planner, later a professor. I teach at ECU in health policy, health administration, and direct our Center for Health Services Research, which is our population focus for ECU.

East Carolina University, a regional university serving most of Eastern North Carolina, an area that has a considerable amount of pathology that exists in North Carolina, a new medical school, but one that is charged with improving the population's health. That is part of our mission. Our university health system of Eastern North Carolina centered around Pitt County Memorial Hospital, the flagship has as its mission improving population health.

So that is essentially my background, the paradigm that I come from. I am actually trained in public administration, but all my life have been involved in public health.

I first became aware of what you all were doing when I attended the Healthy People 2010 launch conference. I have to say, it was probably one of the best conferences I have ever been to, partly because it snowed and we were stuck and we had great services. I also have to say, the best presentation at that conference was delivered by Jeff Blair. It was my introduction to what you all are doing, and it was crystal clear in Jeff's presentation. I can't hope to do as well as Jeff did.

But what I would like to do -- and I don't have anything prepared in terms of handouts, but I would like to talk to you about some of the vision issues and challenges, trends and gaps, and speak to them from a conceptual, practical and technical point of view, starting with Healthy People 2010.

I have been really amazed at how little empiricism there has been in regard to the two overarching objectives of Healthy People 2010, first being increasing life span and quality, and the second being eliminating disparities.

Now, what I have done is, I have taken some slides that I have used in a presentation just recently at APHA, and put them up here. I'm not sure what order they need to be in. I was not sure at all how you all were going to be doing this. But I did see that you had a reference in the vision document to Evans and Stoddard's model, and I would argue that this is a very good way of focusing it. So if there is nothing else up on the screen while we talk, I would like this one to be up. But I will come back to some of the issues in terms of the empiricism involved with Healthy People 2010, and helping communities pursue the Healthy People 2010 objectives for healthy people and healthy communities.

I would skip to basic challenge. This is what we have seen in the last century, incredible improvement in the survival curves. This I think is the basic challenge that we have before us.

Jeff, if you're not familiar with any of these, I can explain them later, but you can clearly see how much we have gained in the space of a century. If we look at it by race, we see a different picture. The dotted lines are whites, the solid lines are the non-whites. You can see that blue line in 1997 for blacks catches up to where whites were in 1950.

So one of the ways we need to conceptualize this in terms of disparities is, you've got a chronologic gap, perhaps because we are all getting healthier, we are all seeing longer lives. When we get to North Carolina though, this way of looking at it is a little bit more difficult to do, because we don't have historic data. So one of the things I think we need to be doing initially immediately is thinking about whether or not we have good systems at the state level that can parallel the Healthy People 2010. Some states can, some states can't.

Jeff, this is just a survival curve for whites and non-whites in North Carolina. We can't go back very far. We can take a snapshot, but we can't provide a moving picture. One of the real challenges I think is to be able to see this longitudinally, historically, and be able to project it into the future.

Now, one of the metrics that I think is lacking from this, and maybe you all can have an impact on it, is the notion of measuring life span in terms of premature mortality. The survival curves are nice and are predominantly a snapshot, but they are not particularly useful at the community level, where we work.

One of the things that we had done was to look at the years of potential life loss before age 75 and project it. I happened to have to do this for a paper I was doing for American Journal of Public Health. This is what we see for the country. Green is good, red is bad. There is clearly a geography to premature mortality. This is 3,080 counties.

One of the things that I am particularly struck by with the Healthy People 2010 is why the benchmark or target for premature mortality got dropped out. If somebody here can tell me, I would very much appreciate it. It was in the draft, and it did not occur in the final document. Since it is the first overarching objective, why? Why can't communities have a benchmark for where they are in terms of life span? I think this is probably one of the best benchmarks we have.

DR. LUMPKIN: Since Mary Jo is not here, it was her fault.

DR. MANSFIELD: Oh. Is she coming back? I want to talk to her.

For instance, if you look at North Carolina, we see North Carolina ranks 40th. GAO had decided that premature mortality was the best single measure of a community's health status or a state's health status, and it would be useful as a guide for resource allocation. Perhaps that is the reason it was dropped out; people really didn't want to quantify resource allocation, I don't know.

We thought it particularly useful. It can be done by race. We did it just simply from area resource file data. One of the things that I want to talk about is making the data available and accessible to people in easy formats. Area resource file on a CD-ROM was particularly useful.

Taking a look for instance at another objective, say reducing coronary heart disease to 166 per 100,000 population, this is how it looks in North Carolina. Is that a great challenge, given that we have half of the counties that need to reduce 20 percent or more? No, it isn't really because this is what has happened in North Carolina over the last 20 years. These are data that we have been able to get from John Booker and put into from SASS spreadsheets and manipulate, and begin to track the disparities.

One of the things that I would like you to be thinking about is whether or not you can help communities benchmark themselves on the two most important goals of Healthy People 2010 -- life span and disparity. This happens to show a regional disparity in North Carolina in the East, trailing behind the rest of the state by a factor of about 20 percent on a rate ratio. One of the things that we need to do is get out some understanding of what the right metrics are for measuring disparities, what ratios, percentages.

Rate ratios are pretty useful. Here we use them just to show for coronary heart disease how we are doing in Eastern North Carolina, the region there, plotted in red, is Eastern North Carolina against the rest of the state. So you see we have grown from a 10 percent disparity to a 20 percent disparity in terms of region, a rate ratio of 1.2 to 1 in 1998.

By gender we see some progress on coronary heart disease, but by race -- and again, empiricism virtually absent in the Healthy People 2010 document in terms of where we are with disparities, looking back for sure and even looking forward.

Here for North Carolina we see disparity, rate ratio, going from null in 1979 -- this is blacks versus whites -- to a 1.2 to 1 ratio in 1979. But that just gives a picture of where we have been. The challenge needs to tell us where the disparities are, and this is what we have been able to do. The bull's-eye there is the target for 2010, but of course that is in the wrong place. We want to move it out there.

Another thing that I thought has been lacking is any kind of projection of where we are going to be in 2010. What we see here is for coronary heart disease, all but black men are likely to make it.

Now, these are pretty simplistic measures. This is just simply a linear regression trend and an extrapolation with a two standard error bar. But I think we can safely say, except for non-white men, we are going to meet it if we don't do anything but what we have been doing. So maybe because of lack of vision, we may have set our sights too high. The goal was too high, and maybe we should be rethinking what the benchmarks ought to be.

DR. LUMPKIN: Could you go back to that slide? Are one of those lines the mortality for all races?

DR. MANSFIELD: I'm sorry, I'm going too fast. The top two lines are men. The black bar is African-Americans, blue is white men. You see there a crossover in coronary heart disease mortality, about 1985, black men versus white men. Then the lower two lines are women.

What you see is a disparity gap increasing when you look at it by gender, an increasing disparity gap. So it is not just the fact that a disparity exists, it is where that disparity begins and what the slope of the lines are relative to each other, to find whether or not the disparity is increasing or not.

Here we were able to take a look at a single county in North Carolina, and the solid line is just a trend line stuck on there to smooth out the data. The R squareds are not big, but these are the data. The blue line is Hyde County, a small county in Eastern North Carolina of about 5,000 people. The red line is Orange County, and this is the racial disparity rate ratio increasing for both, but it started off in Hyde County that whites had the greater disparity, and as a matter of fact, still do. The whites still have the greater disparity in Hyde County, African-Americans in Orange County now have a disparity rate of 1.6 to 1 on the coronary heart disease.

This just takes us back in terms of Eastern North Carolina to looking at the issue of disparity. We find that racial disparities are actually greater in the rest of the state than Eastern North Carolina, where the population is typically a third non-white. But that is only because the whites are so unhealthy also. So if the whites are the denominator, their rates are abysmal, too.

Now, we have both a challenge of burden disparity and disparity trend. What we have been able to do here is map in two dimensions, mortality and the disparity, highlighting for instance those counties with 25 percent variance from the objective for in this case stroke, shaded the white to dark green, and then superimposing upon it the counties that have a higher disparity ratio.

So these are data that are pretty easily grasped in a policy arena. The next extrapolation of this of course is taking it and showing the direction of trend as well. So which counties have the high disparity have the high burden to begin with, but which have a disparity trend going in the wrong direction, whether or not you can bend the trend.

In summary on some of these data, I think years of potential life loss, premature mortality, is a very useful metric. It is easily grasped, and certainly highlights for us in Eastern North Carolina just how poor our health is, in terms of premature mortality. If we treated Eastern North Carolina as a state, and there would be some 13 that would be larger in terms of population, or smaller in terms of population, we would end up being 51st. That is something you can show your county commissioners, your area legislators, and it is a useful way for communities to begin to see how bad they are in relation to other communities.

Another way of looking at it would just be excess mortality. If we took the western portion of Eastern North Carolina as the standard and just created a standardized mortality ratio and cranked that into excess deaths, we can express that such that in terms of all-cause mortality -- that is what this one is -- there are about 1600 excess deaths in Eastern North Carolina compared to the best part of the state.

But the real challenge is getting back to those survival curves. What we would hope is that in the future, the curves get squared and squished up against each other, whites and blacks together. That I think is the health policy goal for any population group.

Do we have the data to do that? That is the challenge that I am presenting to you all. It is not just the mortality, it is the morbidity and the disability data that we need to get at, because we are all getting to live longer, but at what quality of life. So the challenge I think is whether or not we can compress the morbidity curve up against the mortality curve and how are we going to go about measuring the disability?

The mortality is there because we get the certificates filled out by the funeral home directors, who send them to the health department, who send them to John Booker. But how are we going to gauge morbidity in a community?

Another David, not David Kibbe, but David Kindig, believes that we can purchase population health if we can first measure it. That is a real challenge. So what I am throwing out in terms of your visioning is whether or not you are helping us pursue the measures of disability, morbidity as well as just mortality.

These are the existing measures that we have. Can we -- and this is where statistics come in -- can we take some of these things and generalize them? Some of these don't have to be generalized. The premature mortality, one of the beauties of that is, that is all deaths. It doesn't generalize to anybody; it is just simply what is. But when we get down to measuring morbidity and disability and functional status, we can't measure that in a whole population. But we can perhaps take it from groups from which we can generalize.

So I guess that is what I have to share in terms of the graphics. Let me leave this slide up, and move on to talk about -- I wanted to make sure that somebody was addressing in the vision the collaboration with other departments of DHHS on Healthy People 2010, and providing an information infrastructure that will help not only DHHS at the national level measure what is going on, but help us at the local level.

We have for example in North Carolina Healthy Carolinians group. We have communities pursuing this. We definitely have some challenges on measuring the quality of life. I'm not sure -- Fort Collins, Colorado and Los Angeles have been able to do it. I'm not sure that anybody has really approached it here in North Carolina. Maybe we need some collaboration with John and Tom Booker and David Kibbe to do that, because it definitely goes back to the data that exist in health care systems to be able to do that.

But also, the issue of measuring health services availability in a community and their use. One of the things that we did in North Carolina not long ago, and I certainly am sorry that it happened, I didn't have anything to do with it, but that was privatizing the hospital discharge data. It has become more and more difficult to get hold of. Communities have a very difficult time getting hold of it. A hospital may be able to provide some of their data from what used to be the medical database commission, but you can't get a good handle on it system wide. So I think privatization of the hospital discharge data took us in the wrong direction.

The next thing is gauging access. We had what, 480 objectives in Healthy People 2010 and 28 focus areas, 20 leading health indicators. At the bottom of that list is access. We still have a very difficult time gauging who does not have health insurance in a community. We rely on the current population survey of the Census Bureau to extrapolate it. But I think there are ways that we can get a handle on it. Who knows what direction public policy will go at the national level, but certainly the access issue is going to come up again. I think we ought to be able to better measure it the next time we talk about health care reform and the expansion of health insurance.

Helping communities map their assets. This is something that David mentioned. It takes communities to grasp health care utilization data and manage patients through oftentimes.

When we do community health planning today, we need to do the community assets mapping, and then just not leave it. We ought to be able to use that assets mapping as a way of linking people to services.

Health literacy is going to become -- I guess it is already a buzz word. It could become more of a problem if we go more to defined contribution for folks than defined benefit. It is going to get more and more difficult just to negotiate the health insurance system. I think we can empower people with a better grasp of that, and I think we can use the Internet. Certainly in Eastern North Carolina everybody does not have access to the Internet, but I think everybody knows somebody who does.

It is not a digital divide that is the problem, because people get digital signals in the poorest, most run-down shacks that you can imagine through a satellite dish for cable TV, or digital signals to their cell phone. It is more a skills divide, and a way of plugging people in to the people who understand how to negotiate the system. I think the Internet is simply a way where we can plug care managers in and then get people to negotiate a health care system.

I am envisioning community resource directories online, available in every physician's office, available to every care manager, available to lay health advisors and faith-based groups, so that everybody can understand what options we have at end of life, or what options we have for cancer treatment in this particular community. So I don't think we have to be digitally literate to use these.

There are two other things I wanted to cover. I've got some technical issues I could go over, but definitely one that I think is lacking is, in the vision data in disaster. I go back to the flood we had last year in Eastern North Carolina as an example. Both in the immediate incident, there was a data gap. We had lots of people come in with FEMA and the American Red Cross, but they didn't know the health care system. They didn't know who to communicate with. It was lucky we had some physician directories that we could plug people into, because we were absolutely reliant on folks from Chapel Hill and all the rest of the state, but they did not know the territory.

So I think being prepared in terms of health disasters is another thing that you might consider. And following up on the health disasters, it has been very difficult to track chronic disease that may have occurred as a result of the flood in Eastern North Carolina, partly because of confidentiality issues with American Red Cross and FEMA. We could not track patients who were actually displaced, other than identify them in our own clinics. So that might be an important issue.

It gets to the final point, and that is interagency relationships. I would ask, does this vision include all branches of the federal government. We know it doesn't include all branches of the state government. We are going to have to have lots of laboratories, lots of state laboratories to work this out in. But it certainly could involve all branches of the federal government.

Specifically, I am asking whether or not you've got

DoD cranked into this, because there is a population in the military that the military wants to maintain as healthy for readiness purposes, but it is also a tremendous opportunity to track longitudinally what happens in a person's lifetime.

I don't think the military has embraced Healthy People 2010. There are three Surgeon Generals that David Satcher could be talking to, and I guess there is a Secretary also, an Assistant Secretary for Health in DoD. They are a population that we could be tracking, and it could be important to understand what goes on in the military experience, and does it carry through into civilian life. So that is another place to look at whether or not you have covered the bases.

That's all I have. Thank you.


DR. BUESCHER: I'm Paul Buescher. I'm a statistician in the newly renamed Center for Health Informatics and Statistics in North Carolina. I have worked in public health statistics for 20 years in North Carolina. My comments are coming from a state government perspective.

I have provided a copy of some brief written comments to Dr. Lumpkin, and I will summarize those now.

My comments overlap with some that other people have made already, so I won't try to eliminate the redundancy, but perhaps just reinforce some of the things that other speakers have said. Certainly it is a very good idea to have a vision, but as other people have mentioned, thinking up a vision is just the first step. The real challenge comes in terms of providing the resources and the commitment and the political will to carry that out over the next decade or more. So I'm sure you are all aware of that, just having a vision but not carrying through simply gets peoples' hopes up, but doesn't get the job done.

The report correctly states that increasing geographic and demographic detail of health data is needed, and also recognizes the expense and technical difficulty in obtaining such data.

There needs to be more discussion in the report of two issues. One is resources, and the other is statistical instability of small numbers. Who will pay for collecting data at smaller levels of geography for purposes of local needs assessment and action? If you are talking about surveys such as BRFSS, then those are very expensive, and somebody is going to have to foot the bill for collecting those. We already have vital statistics data at the local level, and there are complete counts, but if you are talking about implementing other data systems that are going to collect population data, who is going to pay for that is a very important issue.

Then once the data is collected, how useful will it be, given that small numbers of events in a small geographic or demographic category may lead to very unstable percentages and rates. Even with the complete count of data, you have random error and percentages or rates may be not very meaningful if you're talking about a very small population group or a very small geographic area. Local data users in particular need more education on statistical issues related to small numbers.

A second point is that the report indicates that data must be provided back to communities, local governments and health care providers to maximize data access and ease of use. Certainly getting data back to the people that are going to use it for local decision making is very important, but there is another point that probably should be emphasized in the report.

A key principle of data quality improvement should be emphasized. That is, giving data back to the providers of the data shows them that it is being used. A lot of times, people complain that they are submitting data to the state and it just goes into a black hole and they never see anything out of it.

If providers of data know that the data is being used for policy or other decisions, it is an impetus for the data providers themselves to improve the quality of the data, especially if the decisions affect them.

There is a feedback loop where more accurate data leads to increased use of the data, which leads to increased visibility of the data outputs, which further improves accuracy at the source. So it is important that the people where the data is coming from know that it is being used and particularly if there are decisions that will affect them. It is an impetus for them to improve the quality of the data.

Related to the issue of providing data back to local communities and other users, there will continue to be a need for hard copies of documents and reports, in my opinion. This should be emphasized more in the report. I think it is mentioned in one place, and certainly trying to increase electronic means of using data on the Internet is very important. But there is a discussion of the so-called digital divide or the gulf between the technological haves and have-nots. If information is published mainly or only via the Internet, many citizens and local users will be excluded.

As the report states, some people are able to find information and use it to make informed decisions about their health. Others lack the equipment and skills to do so. So I think that for some time to come, there will be a need for hard copies and printed publications to complement the Internet and other means of providing information, particularly when you are talking about local data users.

The vision report recognizes that data derived from systems established for non-statistical purposes such as payment systems, Medicaid payment systems, for example, can be of broader use, but these systems may not readily be converted to a useful form.

I think there needs to be more discussion in the report of making better use of existing administrative data systems such as Medicaid. With limited resources, the establishment of new large-scale data collection systems will be difficult. Improving the use of existing data systems should become a priority.

An issue in this regard is the expansion of HMOs and capitated payment systems. It is critical that these health care providers be required to submit detailed encounter data for services not paid on a fee for service basis. Otherwise, there will be large gaps in paid claims databases and important data systems such as Medicaid will become less useful over time.

There is a discussion in the vision report of data linkage, and certainly data linkage is a very important means of increasing the usefulness of health data. The report states that quote, linkages of individual record data must occur within a newly established legal framework with appropriate human subject review board approval or permission from data subjects.

This statement seems misguided to me. Implementation of such regulations regarding data linkage would result in important data linkage activities grinding to a halt in many settings. Many state health statistics staff already have access to confidential data files. Protections are in place, such as requiring staff to sign legal statements to protect individual privacy. Linking two confidential data files does not usually result in any new confidentiality problems if staff already have approval to access the individual data files to be linked. Adding legal or administrative hurdles to data linkage seems unnecessary, and would greatly impede data linkage activities in state government settings.

So those are my main comments. I would like to echo what Janis Curtis said about perhaps dividing up and prioritizing the division, so that there are certain things that are considered top priority and should be done first, and then maybe a second tier of things. As it is, it sounds great, but it is a bit overwhelming. I think to maybe come up with some priorities would be a good idea.

Those are my comments.

DR. LUMPKIN: If I could toss it back at you, and you may not necessarily want to do it right at this moment, but if one were to priortize the pieces of the division, what would be your priorities?

DR. BUESCHER: I don't know that I can respond just off the top of my head. I have to go back and review it, and maybe come up with some ideas for that.

DR. LUMPKIN: We would appreciate it if you could just e-mail it to us or send us a letter. That would be helpful.

DR. BUESCHER: To Rob, is that right? Because I have been communicating with him.

DR. LUMPKIN: Sure, and he can send it to us. Questions? Thank you both for the presentations. They have been very thought provoking.

I've got a couple of questions and comments. Chris, in your presentation, you seemed to place a lot of emphasis and analysis on county level data. Some of the challenges that I seem to be facing in Illinois is, what do you do in the inter-census years in trying to calculate rates?

Second is, when you start trying to look at rates particularly -- and I assume that you probably have the same problem here; we've got counties that go down to seven, eight, nine, ten thousand people. Looking at anything over a short period of time gets to be very hazardous in trying to make some estimations of what is going on.

DR. MANSFIELD: That is always an issue. Perhaps John Booker wants to respond to it. But I think when you have small numbers, you just simply have to go to a longer time period, or use statistical techniques such as a regression trend to look at it. Otherwise, I think you get led into areas you don't want to go into by small numbers.

DR. LUMPKIN: Let me -- unless Paul has a comment on that, or anything to add?


DR. LUMPKIN: Let me take that a little bit further. In your analysis you were using trend lines. The assumption is that the trends would continue. Now, cardiovascular disease mortality is a combination of a number of factors. One is increasing technology and preventing deaths in those who develop cardiovascular disease. Another is preventing the cardiovascular disease from occurring at all. At some point, one has to expect that there is going to be a plateau effect. To what extent do we look at in setting those goals -- can we at some point, particularly as we look at various diseases, expect that we are going to run into a plateau effect, much as we have seen in the slowing of the rate of decreasing infant mortality deaths?

DR. MANSFIELD: I'm not a cardiologist, but --

DR. LUMPKIN: That's okay, the others aren't, either.

DR. MANSFIELD: But obviously, if -- I don't know why we would want to anticipate a plateau. I think if we have to project forward, we project forward based upon the best data that we have.

Maybe there are better techniques. Those are just simply the utilities that exist in Power Point, to plot those. But the time frame is important. One of the things that of course you hope for when you start looking at this is, for instance, was there an effect of having a new medical school in Eastern North Carolina, a new academic health center there. When we plotted the coronary heart disease first on five-year averages, we saw a very remarkable downtown that began occurring in 1985. That is when we brought cardiac surgery there. But when we plotted all the data points and put the trend line in, you don't see the effect. So you have to be careful, and I think always use more data rather than less.

DR. LUMPKIN: Paul, and then I'll get to Ed. I wanted to challenge your position on paper documents. I'm not sure -- and I am kind of stuck between agreeing and disagreeing. The reason is that it seems to me those that are going to be data users at the local level, and by and large, many of them are going to be not data users, but information users, data with analysis being information.

They don't generally go to the primary source. They want to know about the data or information related to their particular community. They are willing to ask somebody for that, whether it be the state health department, or go to the library.

At that particular instance, it seems to me that the system that we conceptualize ought to not develop a Vital Records Report Illinois, which is a big huge tome that I'm not sure very many people read, and certainly the people we are talking about in the community would have very little utility. But the ability to create hard copy that is particularly customized for the requester, and that maybe ought to be the role of the holders of the data to present in a means which is useful for those data users, rather than just saying that it has to be hard copy or an electronic means.

Also, those who would be looking for that, the normal sources that would go for this hard copy data, which would be the library, and most libraries in the nation have access to the Internet, so that would be certainly a means for them to get to it.

DR. BUESCHER: Well, I'm not proposing that we continue to produce large printed volumes with tons of data. For the health professional and the savvy individual, accessing on the Internet is no problem. It depends on who your audience is. If you are talking about local politicians or community health improvements or local coalitions or this type of thing, not all of those people have access to the Internet.

Perhaps the bulk of the data could be published in electronic form, but if it is NCHS or state government, maybe there needs to be some synopses or some fact sheets or some boiled-down version of the data that can be made available in printed form, even if it is in a PDF file on the Internet. Somebody at the library could print it.

I'm just saying that not everybody is going to always be able to go to the Internet, and there needs to be some consideration of those folks as far as data dissemination goes. Not that all the detailed tables need to be continued to be printed, but some synopsis.

For the people that are working in public health professionally, it may not be that necessary. I'm thinking more in terms of citizens and advocacy groups at the local level.

DR. SONDIK: Chris, I have several questions -- through the Agency for Health Care Research and Quality, through their HCUP program?

DR. MANSFIELD: No. I think Tom Ricketts can probably speak better to this. He has followed that and he had tried to catch up and plug the gap. But there was an interesting privatizing -- one thing that they could find to privatize, and I would have to say, if health care is not going to be seen as a public good, damn it, health data ought to be a public good.

DR. SONDIK: I'm surprised.

DR. RICKETTS: The state requires the reporting, and the reporting goes to a qualified data vendor. The data vendor then makes the data available to the state center for health statistics or health informatics, and the Shep Center does the quality control for the Division of Facilities Services, which actually regulates this.

We just started an agreement with the HCUP people to flow the data to that system now. It took a while for that agency, as I call them to day, would you like our state data, we have received approval from the private vendors and the hospital association, which politically controls this, they said, no, you don't have a state database system.

Well, we have a continuous hospital discharge database system since 1989. It is just that it is hard sometimes to get people to believe we do. But the data are available if you purchase them from HCIC SACS for a private user, for a nice little chunk of change. But they will flow into the HCUP on a restricted basis next year.

DR. SONDIK: They will?

DR. RICKETTS: Yes, they will.

DR. SONDIK: In which case, they can be retrieved by the state.

DR. RICKETTS: Well, the state gets them. We do analyses. They can be retrieved by a community. We respond to communities who request it with permission of the Division of Facilities Services, that provides it for a research purpose. So there is some availability, but we are creeping along as the private owners of this don't want to see their market abused or trampled on. If we tend to be a loss leader for them, they are happy.

DR. BUESCHER: We may be in a little bit of a special position, in that the center gets the data through a legislative provision, but it is probably not as available to others as it used to be. But I will have to say that under the new system, the data that we have gotten from HCIA has been of higher quality, there has been more variables available to us, and the data has been much more timely than it was under the old system.

So as far as public health goes, I think we are better off now. That leaves maybe some other people out, but we have been fairly pleased with the change from the standpoint of the data we have been getting.

DR. SONDIK: So you say it is value added?

DR. BUESCHER: I think so.

DR. SONDIK: Mary Jo decided to leave again, so she can't answer the question about Healthy People. But I would think that if you calculated instead of years of life lost, you calculated longevity, you would see different numbers, but you would see the different kinds of relationships. That is a question.

In other words, basically the same numbers are going into calculating years of potential life lost, which you are cutting off at 75, but if you calculated longevity --

DR. MANSFIELD: Oh, you mean the survival curve.


DR. MANSFIELD: Those survival curves -- that one that you saw for North Carolina was done with the '89-91 data that CDC had. I don't think states routinely do those. And of course, it is difficult to do at the local level.

DR. SONDIK: Because of the denominators?


DR. SONDIK: The demographics.

DR. MANSFIELD: Right. But premature mortality is a very simply graphed concept, and it is all the deaths.

DR. SONDIK: Well, it still requires the demographics to do it.

DR. MANSFIELD: But to do survival curves, don't you need census data? And you can calculate premature mortality just off of vital records data, and you can do that on an annual basis. Once you get into the intercensal years, the population estimates particularly in the latter parts of the sevens, eights and nines, tend to be less and less reliable, particularly if you are trying to look at disparity, because the ratio makeups at least in our experience, our estimates have not been reliable in the latter years of the decade.

DR. BUESCHER: Can I comment on that? We have one of the census user's offshore sites here. Nationally, we are finding some -- the 2000 census as a census, the intercensal estimates may be more accurate in and around certain -- as we go forward in the 2000 census, because of lack of response and problems with applications of the census.

So don't sell your intercensal estimates short, and don't compare them directly to the enumeration, because the enumeration has a margin of error probably greater than the estimate error. Especially for small area statistics, the imputations that are done are almost always statistical, anyway.

DR. MANSFIELD: And I don't think you have to do it every year. To do it every year is to be tyrannized by the data or the technology. It is a gross population descriptor.

DR. LUMPKIN: But I think the point being on the survival versus the years of productive life lost is that you can make those calculations. You don't require censal data to do that. As a measurement, we ought to take a look at that as perhaps being more useful, and take that suggestion and look at it a little bit more.

DR. MANSFIELD: Either one are easy graphics.

DR. SONDIK: I do think to make sense of a time stream of YPLL measurements, then you need the demographics. But one point in time, you could certainly do it. But then when you start comparing populations for which the demographics are changing, you need some way of being able to compare the population at one point versus the population at another point.

I think that the reason that longevity was used rather than YPLL had to do with picking an age at which to calculate YPLL. In fact, John just said years of productive life lost. Some would say years of potential life lost. But it really is a value judgment as to whether or not you want to choose 75 as the right age, because clearly people can be productive between 75 and 80, 75 and 85 and so forth and so on.

I think there is a kind of a value judgment that comes in on that that it makes me nervous to make. Whereas, longevity is not something in which you are making a value judgment. I think it is important to use those kinds of measures in looking at the country as a whole or in looking at counties or whatever. But I think that was an issue.

DR. BUESCHER: Well, the quantification was just simply greater longevity and no disparity. There are no numbers applied to either.

DR. SONDIK: Well, we probably could talk about that offline as to how quantitative one wants to get, including setting goals and that sort of thing, quantitative goals overall, versus setting a more general goal of eliminating disparities.

Paul, I wanted to ask you if you could say a little bit more about your next to the last point, which had to do with the linkages of databases and confidentiality. What about in those situations or those states, for example, in which this is a difficulty at this point? Do you think that these guidelines or -- I'm not sure what the right word would be, other than guidelines, to allow one to in fact link databases but still preserve confidentiality would be an improvement in other states?

DR. BUESCHER: I guess maybe I am biased in North Carolina, that we have been linking Medicaid and WIC and other data to vital statistics since the early 1980s. Compared to what I have heard about other states, we have had a relatively easy process. I'm not trying to minimize the confidentiality issues. Certainly we have had written approval and agreements and MOUs with other agencies that give us permission to access their data files with names for purposes of linking. It is just that as a health professional, if you have the appropriate clearance to access two different data files, then to me, just linking them together doesn't create any new confidentiality issues.

I think that to talk about having institutional review board approval or seeking some kind of written consent from patients every time you link databases, I just really don't see the need for that. There needs to be some administrative approvals, but to talk about IRB or informed consent from the individuals, that would just practically halt the activity, in my opinion.

DR. LUMPKIN: If I can maybe pose a counterpoint, and let you respond to this, because this is something I have been struggling with. Part of me agrees with you and part of me disagrees. The part of me that disagrees is, those of us in state government have no checks and balances on what we ask for and why we ask for it. I'm not saying that it is done maliciously, but there needs to be in any sort of evaluation of information policy, the need to know should be weighed against the potential risk.

In Illinois, we are now in the midst of a lawsuit with one of the newspapers who wants to get access to our cancer registry data. They want to get access to three variables, which we claim would jeopardize the confidentiality of the people as part of the cancer registry. In fact, we hired an expert who we gave these same three data elements in 50 cases, and they were able to identify by name half of them.

So the ability of people to do record linkage and then subsequently from those to link those to the data that may be derived from that to publicly available data sets, increasingly jeopardizes the privacy of the people in the states.

So at some point, it seems that we -- not necessarily an IRB, but there should be some entity within state government that reviews these kind of practices and says, yes, the benefit is worth the potential risk, rather than just having this be an internal decision to each operating unit.

To that end, we are doing some things. We created a chief privacy officer. We are creating a privacy committee in our agency to do that sort of review that is external to the program.

DR. BUESCHER: To some extent, I think we may be talking about different things. I was thinking more of taking databases that exist in state government, existing administrative databases, and linking those together within the state agency.

I think what you were referring to is providing a data file or some data to someone outside of the state government, who perhaps then could take it and link it to someone else and identifying it. That is more a data release issue. I am talking about taking data internal to state government and doing data linkages, not releasing it to anybody outside, and not in any way contacting the individual patients. It is all strictly an administrative database linkage internal to state government.

I think any time you release a data file to someone outside, then there need to be very careful provisions about the potential for identifying people.

DR. HUNTER: Can I add a comment to that, John? I think one of the issues is how apparent is it, or a term that is increasingly used and not understood, how transparent is it, and what are the ground rules.

I think to a large extent, there are linking activities going on on a routine basis within agencies of government that are not well known to the public. Because there is no set of uniform ground rules and because there is no agreed upon review process for those, when people find out about them, they get upset because they feel as though they have provided information to one side of government, on the assumption that the other side wasn't going to see it. There is something about a combined database that puts an individual at greater exposure in a privacy setting as opposed to confidentiality, using those terminologies not interchangeably, but differently.

I think to the extent that we can use this process to help sort through what are the ground rules, it may not be that there is a specific IRB as the solution, but some greater clarity as to what the conditions are that a public agency would engage in this kind of linking, whether that is different depending on the types of protections you have against disclosure. I think there is a whole host of issues that currently are muddied and discussed as if they were all the same type of issue, when they need to be parsed and sorted through.

The part of me, the patient side of me, wants there to be greater clarity. The part of me that is the public health agency side wants greater clarity also. Maybe I am working toward some synthesis in my own mind, but I think people have different objectives.

I'm not sure that what you are saying is inconsistent with the desire for ground rules review. It may be at such an aggregate level that the ground rules are clear and you don't have to go to an IRB for each individual case.

But I think it is the intersection of all kinds of different issues that have clouded this issue and made it very difficult for people to deal with.

DR. BUESCHER: What you say makes sense. I think that maybe the document or subsequent regulations need to be specific about what types of data linkage would require IRB approval and what kinds would not. But just to say that any data linkage that occurs need to have IRB approval seems to me to be going way overboard. There needs to be maybe some guidelines about what types do and what types don't.

DR. SONDIK: You certainly made me want to go back and look at it again, because that strikes me -- the way you are recounting it makes it sound as if there is a decision made every other second, if you will. I think the importance here are the safeguards that need to be there to assure that the patient's privacy is safeguarded in whatever happens to the data. The more it is shared, the more probability there is that it could become a matter of public record in some fashion.

DR. BUESCHER: We are very conscientious about making sure that any databases we have within the division of public health that have patient identifiers are closely guarded. Staff sign confidentiality agreements. To my knowledge, we have never had a case of inadvertent disclosure of patient data.

But I am talking about using the data within the division of public health, not releasing data files to people outside. There are provisions to do that. There are certain circumstances under which we can release identifying data to people outside of the division, if we go by certain guidelines.

DR. WEINZIMIER: Chris, I am curious about how you react to this discussion about sharing information among agencies. You were talking about FEMA not knowing from the health agencies what they have got, and you were talking about federal agencies not talking to DoD, the health side not talking to DoD.

What do you think when you listen to this discussion about the rules we have for sharing information among ourselves? Are we crazy?

DR. MANSFIELD: Well, it is partly who is the architect of the silo that you live in. For instance, I doubt that James Lee Witt has ever really had to think about going through an IRB. FEMA just is not thinking about people taking data about a disaster and doing research. They do have somebody in FEMA who does research, but she has never relayed it to an IRB, I'll bet you.

DR. BOOKER: I'd like to comment. I've been to a number of these discussions about the confidentiality of data and linking data, and the pressure is always on the public entities which are doing this, and protecting the public's health for quite some time, with a high degree of trust.

I am a member of a public-private consortium in North Carolina, the North Carolina Health Care Communications Alliance. Rarely is such pressure or focus provided on people who are working in the private sector, providing them some ethical guidelines.

There are situations that require that sometimes government is the wrong place for data, which also has to be kept private, and there are compromises that need to be made.

The thing that Paul is concerned about are blanket considerations that seem somehow to include those public trust organizations which have earned the trust and earned the ability to do this kind of thing. I think that it amazes me constantly that the degree of self reflection and concern of the jeopardy that the government agencies are placing -- I know that we have a great deal of public trust, but the organizations such as this committee must be balanced in their concern with the use of these data, because rarely do they come from inside, they often come from outside of the government agencies or public trust entities that have some specific mission and recognition of their responsibilities.

DR. LUMPKIN: I think I'm going to have to disagree a little bit. The concept that we have as a nation that we are working towards in access to health information is based upon the concept of consent. Public agencies are the only ones who have access to health information without consent. We do that on cancer registries without consent. Certainly our communicable disease surveillance are without consent, and have to be.

So in that regard, we have a very long tradition of maintaining public information, private health information in a confidential way and protecting it. But we also have the obligation to make sure that as public expectations are changing and technology is changing to make sure that we are being consistent with those changes.

We have to be careful that we don't just look at our past successes and say that that is going to be evidence of success in the future. Rather, we need to constantly re-evaluate our systems to make sure that we are taking all the steps that should be taken.

DR. BUESCHER: One point I was trying to emphasize is that data linkage does not in itself by definition create new confidentiality problems. Certainly, a lot of the data linkage that we have done over the years in North Carolina have contributed to the public good, because it has allowed us to evaluate various public health programs, we have provided lots of data back to Medicaid on outcomes such as infant mortality that they would not have otherwise had without linking Medicaid data to the birth certificates.

So I think there has been quite a bit of public good created through the data linkage. To just squelch that by unnecessary regulations would have some disadvantages.

DR. LUMPKIN: We're going to take Kepa and then Ed, and then we're going to break for lunch.

DR. ZUBELDIA: I just want to make a quick comment for the record. I think that in order for this dialogue not to be misinterpreted, you stated that you don't see any problem with linking data that the state has, at any level that the state has. I want to make sure that there is some clarification here. We are not talking about financial data, for instance, from tax returns and so on.

DR. BUESCHER: Our experience has been mainly just with health information databases. I couldn't comment on tax data. We thought it would be nice to be able to link the income tax database to add some socioeconomic data to our health measures, but I was really focusing mostly on health data.

DR. SONDIK: When we are formulating public policy about this, one way to think about any of these things is in terms of the risk, risk of disclosure. Do you have any thoughts, not necessarily now, but thoughts that you would like to give us about what the appropriate criteria is for that risk? That I think would be very useful.

Quite often, these discussions are discussions in absolute terms. There will be no disclosure. We have put in safeguards so that there will be one. But in point of fact, with probability there will be. What sorts of measures should we use for assessing the policy, a priori? How much risk is one way to view it.

We in NCHS for example have a certain rule in terms of the number of individuals in a cell. There must be a certain number of individuals in a cell before we will release the data. That is one way to look at it, but given the wide variety of databases that exist and the number of different ways to ask for data from a public agency, it clearly is possible that sooner or later someone will be able to triangulate and identify individuals.

Now, why they would want to is another story. But if you have any thoughts about appropriate criteria for evaluating public policies, it would be very useful.

DR. BUESCHER: I'll give that some more thought, and if I come up with any more ideas, I'll let you know.

I do think, as has been mentioned before, there needs to be some balance between the right to use the data and protection of individual privacy. You could clamp down so much on use of the data that there would be virtually no chance of there being any violation of confidentiality, but then you would be missing all the advantages of using the data through linkage.

So I don't think you can ever guarantee that there will never be an inappropriate release of the data. As you say, you've got to balance the risk.

DR. LUMPKIN: Chris, did you have a comment?

DR. MANSFIELD: Well, universities deal with it all the time. For instance, on some of the other agencies that have data, maybe just giving an IRB human subjects model to FEMA for instance -- this is the way we do it, and this is when we get concerned about an individual being identified because of a small cell.

But then let's look at the possible benefit of having this knowledge available, and making sure you are sharing it with somebody who is a health professional or somebody who is a professional and is not just going to turn it over to the guy who comes around and says he will get the model out of your ventilating system.

DR. LUMPKIN: Well, if they do that under contract to the local Medicaid agency, then they would be a provider and they would have a provider number.

I think I would like again to thank both panelists. At this point we're going to break. I would like to suggest that we use this as a working lunch. Dave has stayed on over lunch. I have a couple of other questions I would like to ask him, and I think others may have, and Chris and Paul here.

Lunch is available. First of all, I have to say this is a government meeting, which means of course that lunch is not provided, sort of. But there is lunch available for nine dollars.

(The meeting recessed for lunch at 12:40 p.m., to reconvene at 1:50 p.m.)

















































A F T E R N O O N S E S S I O N (1:50 p.m.)

















































DR. LUMPKIN: This is a joint panel. We are going to start off with Marc Kolman, who is -- are you the Director of Person County Health Department?

MR. KOLMAN: Yes, I am.

DR. LUMPKIN: Administrator, or whatever the term is here.

MR. KOLMAN: My name is Marc Kolman. I am the health director in Person County, the Director and Administrator of the Person County Health Department.

I'll just give a little background on who we are, where we are situated and all that. About an hour north of here, 40 to 45 miles, so on the fringe of the Research Triangle area, but we are still considered a rural community.

DR. LUMPKIN: Are you in Eastern North Carolina?

MR. KOLMAN: I am in Piedmont, I would say.

DR. LUMPKIN: Piedmont, I would say.

MR. KOLMAN: John is shaking his head yes, because he is also local like I am. I consider myself local.

My department has got a staff of about 55 to 60, about three something million dollar operating budget per year. I'm trying to think about what is going to be most useful for you all to hear, so I'll shoot from the cuff, and please ask me any questions.

I have read both of your draft documents and tried to understand most of what is in there. I think largely I am very impressed. I think something that I think might be useful and will try to impress on you is where local public health systems are nation wise, and how some of our staff might plug in, where some of our current automation systems might plug in.

I have also been chairing a committee trying to develop a statewide automation system for the state. I think that might give a little bit of perspective on some of our dealings with private vendors versus what we can do with state government, and some of what we are faced with on the local level. This is of course only in the North Carolina experience. I can't really speak about any other states. Though sometimes I get the sense that North Carolina is impressed as being on the forefront in a lot of the areas that we are talking about.

I don't know if this is going to be much of a surprise, but automation wise, we have got staff who are still very scared of touching a mouse, using a computer at all. When I read the NHII document, I was like, wow, this is cool, this is an incredible vision. I was scared to show that to some of my staff, for fear of what they might think they have to get involved with.

We don't have a whole lot of automation. We have a number of PCs and we have a network, and we are connected to a state mainframe via a T-1 line, but we don't have a lot of variety of automation going on in local public health. I can train people, but until they have a whole lot of application, as David had said, what is the value of information sharing. We are still a fairly small department and people talk to who they want to talk to, and they don't necessarily need to use e-mail, though I am trying to encourage that more and more. Until they have some real application, that makes sense. I don't think that that training is going to be of much value.

My thinking as I was reading through these things, training seems to come up a lot. It is also one of the essential public health services functions. It needs to be addressed I think very much on a local level. In North Carolina we also have a PHTIN, which is a CDC funded deal. For my staff to travel an hour to find parking in Chapel Hill to attend a PHTIN presentation on something they are marginally interested in is going to be a real challenge. So we need to get things localized and applicable for people to start understanding how to use technology and also what the implications might become when you are talking about getting a long term vision.

We don't have -- again, this is my experience in Person County, not necessarily statewide, but we don't have technical interface with other pieces of the health care system. In Person County this is true. I know North Carolina has some communities where there is a lot of integration going on between the hospital and the health care and local providers and that kind of thing. But for the most part, I think that is really not very true.

DR. LUMPKIN: How big is Person County?

MR. KOLMAN: A population of about 35,000. So we are a fairly small county.

DR. LUMPKIN: Medium sized small county.

MR. KOLMAN: Medium-sized small county. I think the 21st century health statistics report had a lot of ideas on disaggregation of data, so taking it down to a very small level. We actually do a two-year community health assessment in North Carolina, so we are looking at comparing county data to state data. I'm sure this is not unique to anywhere else. We are looking at data that is several years old. There are a lot of questions; as soon as that comes out, how come it has to be four years old, five years old. So the issue of real time data access is I think very useful.

Also, a lot of questions come up about smaller municipalities, or what about the county seat of Roxborough, in my county as opposed to the county as a whole. Looking at a lot of environmental determinants is an issue that has come up a lot. We started a report about a month ago on these questions that are coming up.

Having real-time access to that data and having the skills to interpret it is another large issue. In my county I am essentially the person with a master's in public health that is able to analyze where the system is from state folks, where possible, but it is a challenge to figure out how to collate all that data, present it in a usable form, present that as a document in a way that the community can understand. So again, it is a whole set of training skills in getting some resources available to the local levels.

I'll talk a little bit about our attempt to get a statewide public health automation system. North Carolina public health, we have 86 departments that cover 100 counties, so there is a handful of district and multi-county departments. All these departments have to report data to a statewide data system.

DR. BLAIR: Are you saying public health departments?

MR. KOLMAN: Yes, 86 departments covering 100 counties. So population size is anywhere from about 9,000, 10,000 as the smallest up to whatever Raleigh, Gilford County and the larger communities.

They all have to dump data to -- they use what we call online, so meaning real time access where they dump data on a periodic basis into a statewide data system.

I first got interested in automation solutions when I got to director level in North Carolina about four years ago. I walked into a department a little bit more raw even than what I am now and asked, how many people do we serve, how many people come into our family clinic and things like that. It was really difficult to get an answer out of that.

We were using an old system that was designed in the late '70s, and is really pretty antiquated. Since Delton is here, I'll refer to him. Delton had a vision of revamping that system a number of years ago. There was a proposal made in order to do that from the state level, to revamp the system. That proposal was reviewed, versus consideration of looking at recruiting a private vendor to provide a solution, a statewide solution. So I am talking about getting it across the state of North Carolina, let alone integrating with any national types of data systems.

The decision was made to, rather than have a state developed system, to try to recruit a private vendor. A lot of effort and time was spent in developing a request for proposals that we feel covered a broad gamut of public health services at least in local public health in North Carolina, and provided a system mechanism for the private vendors to propose whatever type of solution they might think would best fit our entire system.

We received only 10 proposals out of what we were hoping would be the best of the best for the solutions that would cover public health. After a pretty extensive period of review, I decided to eventually terminate the RFP. We didn't think any of the private systems, either that we could afford -- there was some systems that wowed us, in the sense that we thought public health would get catapulted by the amount of information that we would have available and the types of things that we could use it for that we don't have at this point.

We didn't think we would ever be able to afford that, and we had questions about it fitting statewide architectural requirements that are dictated through the state. Some of the other products that we thought might well fit our -- provide a solution for us; we had questions about whether they would be able to manage implementation across the variety of departments that we've got on different platforms and different kinds of programs that are provided.

We do have in North Carolina -- all counties have to use the statewide data system. There is also -- most of the larger counties if not all of the larger counties do operate proprietary software, so we've got three different packages to operate in the state.

That has left us know -- we just terminated the RFP a few months ago. We are now at a state -- we have joined forces between local public health and state public health, Department of Health and Human Services, to look at doing a joint development effort, at trying to develop a comprehensive automation system for us.

DR. LUMPKIN: Thank you. We really do apologize about the Person joke.

MR. KOLMAN: They call them Personians. I'm not native, so I don't take it personally.

DR. LUMPKIN: Oh, okay. John?

DR. BOOKER: Let me give you a little bit more information about me and my background, that I sit and listen to all this and keep reminding myself that I have a little bit of idiosyncratic view of things.

I have been in North Carolina as the director of the Center for Health Informatics and Statistics for almost three years. But before that, I spent 15 years in Alaska at the University of Alaska, working to some considerable extent with state government and state public health in Alaska, as well as with native health corporations.

Before that, another 10 years or so in the Southwest part of the United States, working from a university environment, but working with tribal entities and at one point stopping out to be director of health statistics for the Navaho tribe.

What brought me to North Carolina was the promise of the quality and capacity of the health statistics agency in the state. That all proved to be true. In comparison to Alaska, it was a lot like seeing both extremes of a bipolar disorder. We in Alaska made do with relatively small resources in state government, and tried to pick up some of the slack at the university. But even so, we were a long way from what would appear to be relative riches in North Carolina.

At our agency now, we have about 125 people. That includes the vital records program as well as vital statistics registries for cancer and birth defects. We have a health services analysis unit that is the TA arm for the state Medicaid agency as well as, we do the assessment analysis of Medicaid managed care throughout the state, and several other programs. As was mentioned earlier, we work with hospital discharge data, things of that sort.

By comparison, Alaska has no hospital discharge data set. The cancer registry is in its infancy. There is no birth defects registry. But on the other hand, one of the things that Alaska has that North Carolina would like to imagine having is a trauma register that tracks the events from the first call all the way through the hospitalization and ultimate outcome.

One of the reasons for mentioning that is that my sense is that many of the things that we develop are very closely tied to their purpose. But the purpose drives the opportunities to produce connections and systems which capture things that we are interested in. So that I think tends to be the place where I see opportunities. When there is interest, we figure out a way.

Shifting from that, let me just offer some comments, both about the documents and about some of the things that other people have said. I'll try not to repeat factual or informational things, but just some commentary.

I guess I am catching on to the national agenda in wanting to lessen expectations, that maybe we would be a lot happier with some of the things that we have if we didn't dream so high in terms of the things that look possible.

One of the disconnects is that what looks strong from the information technology side perhaps has not really infected the user side of the equation in health statistics, and I am talking here about the statistics document. Some of the things that may be of immediate experiment in health care and in the provision of services may not have as much immediate attraction to the people who work with and use health statistics -- really, a reflection of the fragmentation of our health sector in the United States.

We have many political, economic and other barriers that are there because of the way health care is organized or not organized in the United States. That really came to mind in reading the documents with the references to Canada and to Europe and other places, which had very different systems than the one in which we participate.

The second observation that I would like to make following from that is that technology is not really what separates us. There has been allusions to the proprietary nature of data. The purpose of those data is not necessarily public health or the public's interest at all, but to run systems.

We talked about privacy and confidentiality. Yet, when I visit my managed care organization and sign on the bottom line, I am also giving them authority to use the information about me for marketing and planning and many other things which don't have a direct connection to my own personal care.

I do think that the documents and the work that is being done offers a lot of promise, because given the fragmented nature of things, there is obviously a need to use standards and to use recommendations to herd things in the right direction. People have spoken along those lines. Without that, there is not much within the current system that is likely to bring it together. The incentives are not within; they are public incentives without.

So in terms of things that I think we can do, I would argue that an important consideration that has been part of the mission and purpose of the National Center for Health Statistics for more than 20 years is to try to support the development of local -- and in local, I am primarily talking about state level capacity in health data and health statistics.

There is still a lot to be done. There are still the Alaskas and there are also the North Carolinas. There is a lot in between, where the development of information infrastructure may not have an immediate payoff, and as a result may not necessarily whet the appetite locally to support the development necessary on the other end.

Another point is that in one place, there is some reference to essentially a unique patient identifier. I realize that that is an awfully touchy subject in all sorts of circles. North Carolina is full of that controversy as well.

In another sense, perhaps, we need to see the development of -- and I think it is referred to as a research identifier in the document, but we need to see the development of something as a proxy to the patient's name and other personal data that will allow us to put some distance between the health data sets and the individuals whose experience is captured therein. But it seems that something of this nature will be necessary to knit together a system that is not organizationally and institutionally integrated.

A third point I want to make is that it is very, very important for us to move beyond race and ethnicity as proxies for what is affecting people's health at the community level. That is really how we use it. Most of the interest in race and ethnicity is not biological. Some of it may be racism or institutional disadvantagement of individuals because of race, but mostly we use it as a proxy for the community in which the person lives, works and is sick or healthy.

We need to get beyond that in order to better address public health issues, and also to move beyond the limitations that are coming with the new race and ethnicity coding in the census and in our own data sets.

So I will under strike the suggestions of an incremental approach to things. Given the system, given our political organization in the United States, that is clearly the way things work, but the vision needs to be there operating continuously in the background, so that the increments are all going to add up in the right direction to the fullest extent possible.

Then a final point. I'll come back to the discussion that Paul Buescher brought out about the linkage of data sets. Let me put it in a larger context.

At this point in time, given the choice between development of data systems and the ingenuity of analytic staff, I would probably choose to have the ingenuity working for me. I am more comfortable with that, and what I really think is important is to not restrain that ingenuity.

With vision, it will be sometime coming, and we need to continue to prove what is possible out of health data sets in the meantime, to establish the value, to do a better job of addressing policy and practice questions as they come up.

So I think Paul's point about not restricting linkage can be generalized to simply not to overlook the need to continue to build and support local and state capacity and to allow people to find ways to address public health questions with the systems that we currently have.

I think with that, I will stop.

DR. LUMPKIN: We have a third panelist who will be here sometime in the near future. But why don't we proceed to questions?

DR. BLAIR: Could you elaborate a little bit more -- you indicated we needed to move beyond racial and ethnicity. Are you indicating that collecting data by racial and ethnicity doesn't have utility, or are you winding up saying that we need to do more with that data or collect more data? I didn't quite understand.

DR. BOOKER: I'm not sure that I have an answer, or would take sides on that, in terms of its current utility or its future utility. The complication of the changes and the elaborations and coding of the data present some huge problems, or huge challenges to using those data.

At the same time, it is I think fairly widely recognized that capturing race on a birth certificate or on a death certificate is at best a proxy for a lot of things that have happened in that individual's life or will happen to them in their future life.

Overlooking issues of family income, education, all the social determinants of health shortchanges our ability to detect and deal with how those factors influence the health outcomes of the populations.

In a state like North Carolina, where that varies so widely, the relative riches of the Triangle and Charlotte and other places to many parts of the state, where the economy and the public health problems still look a lot like the 1950s, not taking that into account means we miss a lot of opportunities in public health.

DR. HARDING: We have had a couple of themes going through a number of the discussions today. One was about providers and how to encourage or incentivize providers to do the right thing, to get up to date on their computer skills, to be willing to supply the right kinds of data that is needed and so forth.

Mark, you mentioned that a good many of the people working in your shop were a little bit less than excited about computer issues. How could we incentivize them to get excited about this? Somebody mentioned the doctors, that you have to give good feedback to them, something they can use and feel like it is really affecting their practice and their care of patients. What could be done to help you with your employees to get them up to speed, so to speak?

MR. KOLMAN: I agree with what was stated for physicians. My understanding is that people are in generally not going to adopt something unless it has got some practical application.

DR. HARDING: In my hospital, the pediatric residents have an electronic medical record. What it means to them is, they spend their lunches inputting that thing. That is what it means to them. So they aren't real excited about that.

MR. KOLMAN: It is a disincentive.

DR. HARDING: Right, but I just wonder if you had any thoughts about that.

MR. KOLMAN: I don't the applications that might be useful for some of my staff that have difficulty getting towards that. Even if they were to see part of this vision that putting out the NHII and espouse that, they go, yes, I see where we are headed at some point in the future, and I want to get on the bandwagon. But unless they have something that is going to make their lives easier, it is going to be difficult --

DR. HARDING: Their lives as employees of the Person County --

MR. KOLMAN: Right. A lot of people bring in skills because they have application at home, on the Internet or whatever, so some of those skills are applicable at work. Other than increased communication, access to data and information, I don't see that much -- I just don't know what the applications are yet. Somebody can dream up some of those applications to get to the physicians -- it is really all providers across the board, not unique to public health by any means.

DR. LUMPKIN: Marc, if you can help me, 35,000 people in the county, how big is your health department?

MR. KOLMAN: It is 55 to 60 employees.

DR. LUMPKIN: Really? And you provide clinical services?

MR. KOLMAN: We do, though we are not a primary care provider. Do you want a synopsis of how we got to be 55 or 60 employees?

DR. LUMPKIN: Yes, because that is pretty large for a county that size.

MR. KOLMAN: We have got a home health and hospice agency of about a dozen. We've got environmental health, which is septic and well permitting, food and lodging. I think we are seven staff, four solely administrative over the whole department. We do a couple of case management care coordination type services for children and pregnant women. Then it is child health, maternity, family planning, breast and cervical cancer prevention, TB, communicable disease control.

DR. LUMPKIN: Do you also do WIC?

MR. KOLMAN: Yes, and WIC. That is how we get to 55 or 60.

DR. LUMPKIN: That makes it. We actually have a few like that in Illinois, so it is a very common situation.

John, you mentioned -- you made a statement early on, something to the effect that some of the things that are attractive to those on the clinical side are not attractive to health statistics in the vision. What would those be?

DR. BOOKER: I think what I had in mind was just that there is a more immediate sense of benefit and purpose and attractiveness to patient records systems and things from the practice side. Those things aren't quickly going to necessarily track over to the public health statistics side, although obviously there are many benefits through the potential connection of those things.

It is a little bit harder -- part of what I think I was thinking about is that I believe there are some fundamental kinds of issues that data systems won't solve for us. Think about it this way.

One of the groups we work with is Healthy Carolinians. Paul alluded to this earlier. We try to provide data to local groups down to the county level and in some cases even lower than that. But there is an inherent tension there about how low can we go, and in what small numbers should we be trying to provide health data.

One of the points would be simply that we are not -- I don't believe -- the prospect is for us to provide individual case level data down to counties, so we are always going to be using what might come out of the patient record system in the aggregate, in some form or fashion.

One of the issues that we face is trying to define the sorts of things that make sense at what level or what unit, that can best be applied to problems. Population health by definition is an aggregate phenomenon. We are not dealing with one person at a time. So there is some level of aggregation that makes sense for different questions.

In looking at those issues, we also move on to gathering local specific data about those communities and conditions believed to be connected to health status and other issues in those communities.

I'm not doing a very good job of explaining this, but simply saying that systems that have all patient data do not represent a total resource to all public health questions. I have worked extensively with Indian Health Service and Native Health Service. Native Health Service in Alaska had a cancer registry 25 years before the state had a cancer registry. Every encounter is captured from the data system. Aside from the data system being sort of cumbersome, data are in there. Yet we still worked with those native health organizations to gather local specific data amongst those populations to provide them with additional information from which to plan and make decisions about health care in those regions.

So I guess my point is two things. One, the public health side may not -- there may be a lag before the public health side is in a position to take advantage of a lot of the development that is going to happen in patient care and the data that are produced, and also there will be some dimensions in public health that won't ever be part, in my estimation, of those patient record systems.

DR. SONDIK: One of the points that came up this morning in terms of the importance of selling this vision, the overall vision or the narrower health statistics vision, is in having examples, in which you can show that this makes a difference, that something would improve by having this system.

So Mark, I am really taken by a couple of things. Let me just ask you a series of questions. Do any of your employees -- yo have actually got quite a number, but what percentage have computers at home, do you think? Do you have any idea?

MR. KOLMAN: I can tell some people that have them at home, because they are pretty proficient. Or they are proficient at home because they come in and show some proficiency. They pick up software programs. I figure they have got some other exposure.

DR. SONDIK: I was always intrigued with this experiment that Ford did -- and I don't know what the results of it have been -- in which they gave all of their employees computers. I would love to know how --

MR. KOLMAN: We'll take them.

DR. SONDIK: There are plenty of computers around. As you know, as people decide they need 700 megahertz for whatever reason, and they give up their 300 megahertz or whatever -- there are plenty of computers around at slower speeds than that that are really good computers. It seems to me these are things that we could harness for some good and some training.

Do you use a non-computer based geographical information system? In other words, if events happen, do you see any need for something like that? Do you think that could help?

MR. KOLMAN: We are almost in the final stages of setting up GIS in our county, with the primary application towards tax records and things like that. Whether we are going to end up using it for health related information, I would think ideally, yes. Whether we are going to get -- let me talk about the last county I was in, that was perhaps a better example. We dealt a lot with drugs and sex, basically, STD's, teen pregnancy, stuff like that.

I had a lot of ideas about the application of GIS, but when I talked to my staff, they knew where the pockets were in a small county. They knew where to go. So in that county anyway, I don't know that it would be all that applicable. Probably the time and sophistication in order to get that, purchase the technology and get training in how to use it and stuff, it would be questionable.

I know it has got some very clear applications in some urban areas and things like that. Whether it would in my current county, I'm not that sure, either. We have right now a lot of concern about cancer rates. Our caner rates really aren't very high compared to anywhere else, but there is this perception that they are. Some have started doing a little bit of work with John's staff, doing some further analysis on that. Would a GIS mapping system to mark all the cancers in the county be useful, it might be useful now.

DR. SONDIK: What about in making contacts with docs? Suppose there were -- you had an e-mail system or a fax system and you were able to do that? That is all part of this.

MR. KOLMAN: The technology I think would be great. I think it would save people a lot of time.

One of the barriers that we run into is really political. We have got a relationship between our maternity staff and the OB staff. They don't trust phone calls, faxes, anything. In some docs' offices we have much closer, better working relationship. I would assume that e-mail and instant messaging and stuff would help quite a bit.

We also have -- out of close to 60 staff, we operate 40-plus PCs. We are operating still a handful of 486s. So we have got a whole smattering across a realm of things. So we pretty much use whatever we can get hold of.

DR. SONDIK: What kind of linkage do you have with the schools?

MR. KOLMAN: Technology linkage?

DR. SONDIK: No, it can be any kind of linkage, personal linkage.

MR. KOLMAN: This varies from county to county, depending on the relationship between the department and the school administration, basically. In our county, we have I would say a pretty limited relationship. We will get called on some particular concerns. There is a particular case, there is a kid with HIV who is getting involved in fights. I was called on Friday, in fact, to figure out what to do.

We don't have that much -- we don't have a very comprehensive school health program. We have one nurse employed by the entire school system, which is 3500 to 5,000, something in that range.

They have data in files that we have already identified that would be useful for a lot of health issues, but it is completely --

DR. LUMPKIN: My question -- because I really want to tease out -- this is the first time we have had a state and a local department sitting at the same table next to each other, so I can cause trouble, particularly since we're not in Illinois, so there are fewer ramifications for me causing trouble.

Marc, what are your leading health issues, and how do you know that?

MR. KOLMAN: I know that -- well, I can tell you what they are. The chronic diseases are cerebral vascular, cardiovascular disease. I know that because -- or I think I know that because we have just gone through a health assessment process. We have regurgitated data that the state has provided to us as part of a community health assessment process. That is required over two years.

I was not very convinced of that until I went through this health assessment process over the last several months.

DR. LUMPKIN: What did you think it was, if it wasn't --

MR. KOLMAN: What do I think it might be otherwise? Good question. I guess the outcome makes sense, given the community and population, economic status.

DR. BLAIR: Maybe your intuition might have a lot of validity. What intuitively did you feel it was, and what was it about the information that you read that you considered had some credibility to cause you to adjust your thoughts?

MR. KOLMAN: I guess my intuition would go based on economic status, racial background, just looking at the community and thinking about the kind of issues that we deal with in the health department. We have a somewhat elderly population, sedentary, old agrarian, formerly agrarian community.

Things that confirmed it is looking at disease trends across different age groups and populations within the county. So the statistics basically were confirming that. So I look at leading causes of death and what types of morbidity, reportable diseases were showing up in Person County that confirmed what you would look at through your intuitive approach. The last county I was in, the same kind of thing. You can see what you deal with in this community. The statistics actually confirm that.

DR. LUMPKIN: You said that the population is sedentary. How do you know that?

MR. KOLMAN: Intuition.

DR. LUMPKIN: The reason why I am asking these questions is, some of the issues that we are addressing is the availability of county specific data, because behavioral risk factor does not provide county specific data for Person County. So issues related to, if chronic diseases are the number one problem in your county, part of the vision that we need to have has to be to give you the tools to say that, number one, if it is a sedentary lifestyle, and two you introduce a program to try to change that, how do you know if the program worked, because there is no data in that regard. So I think that is an issue.

Another piece of this --

DR. BLAIR: Can I just add one piece here, too?


DR. BLAIR: I just wonder -- because I am a great believer in intuition -- when you read that data, did you feel as if in some cases they weren't asking the right questions? Did you feel as if it is possible your intuition still might have been right, but the data that was returned to you just didn't flesh that out?

MR. KOLMAN: I think the question goes down to two things. One is level of aggregation that you have been talking about, and the second is what John was talking about related to the determinants of health. So we have stuff broken down by racial and ethnic and age.

The other things, such as cardiovascular disease related to fitness issues or your economic background or were you abused as a child, or who knows how many other things you are going to link in there, if we are truly to get a sense of what the determinants of a community's health status are, you have got to get below the number of factors that we are looking at.

So yes, I believe that what people are dying from are what are indicated on our mortality stats. What causes that, I don't know that. There are I'm sure ways to get at deeper levels than what we have gotten so far. But I am trying to look at other people who can do that, can help think about that. Again, it goes to a resource issue.

In a small community, we don't have that. I don't have an epidemiologist, we don't have people who do policy research in Person County.

DR. SONDIK: Is there a county medical society? What I was going to ask you is, if we have this vision, whichever one it is, there would be a lot more communication and information shared by health providers, all health providers including the schools for that matter. Does the physician community have a good sense of what the problems are, or do they wait to see what comes in the door next, if you will?

MR. KOLMAN: I honestly would say that I don't know our physician community all that well, which should tell you something. We don't have a lot of feedback. We're looked at as -- I listed the programs, services that we provide. We are looked at as a health services provider, we are not looked at nearly as much as the public health authority that has understanding of the community's health statistics and determinants of health as I might think we do, and perhaps in other state scenarios, that happens.

But we are largely seen as providers.

I think the interest -- I don't think I am speaking out of turn -- for local physicians are largely business interests. So they look at carving a niche. We are an hour from here, not far from Duke and Durham, so there is a lot of overlap and competition, if you will, to keep people local, with local providers. A lot of physicians practice like one day a week in Roxborough, our county seat, and have satellite clinics for Durham providers.

Would they change their practices if they had a different picture of the community's health statistics? I don't know.

DR. SONDIK: Do you get into third parties? Do they talk to you about what the profile of problems is, or profile of the population, that sort of thing? Who do they talk to?

MR. KOLMAN: I get a sense they have much better statistics than we do.

DR. SONDIK: Where do they get them?

MR. KOLMAN: I guess they are a member of a pool, I would think. Or decisions on like who to enroll, they enroll Person County employees, and how much do they charge for getting the risk stats.

DR. SONDIK: The sense I have is, they infer a lot of that. There are different standards of evidence, if you will. I think they use a looser standard of evidence that is based on testing out and seeing whether or not it actually turns out to be the case. In other words, we'll make a business decision, and if we invest, if it doesn't pay off, then of course we're do something else. We don't seem to do that in public health.

In general, the rural does a lot more on the basis of some evidence and inference, and then feeding back in a loop.

I do think though that what you both had to say is very important for us in terms of getting a handle on what the problems are, how better statistics and information would be used not only by the health sector, but health information that is used by the other sectors as well. I think it is very important as was said earlier today that we include in this the other -- we think about this as part of not only a better information set for health, but better information that can be used by other sectors as well, and try to intersect that with that as much as possible.

It is natural when you start thinking of things geographically; then things fit together quite well. But if you don't think of it that way, then we tend to isolate things. So when we do our various surveys, for example, whether they are national surveys on health or BRFSS and so forth, we tend to isolate the questions and talk about health issues. There may be other things that we should be talking about as well. That would give us a better clue as to what other factors may be involved in either health behavior or cause for that matter.

DR. HARDING: I was going to make an observation on the inference and intuition front, and how maybe what we do -- when we talk about the health statistics vision in particular, we may be conveying the impression that what we are trying to do is replace intuition and inference with all the data you could ever need to make a decision, which you would never have. So what we are really talking about doing is better informing the intuition or the inference.

It may be with hard health statistics and facts, it may be -- I would guess that much of your intuition about your county came not from health data, but from census data. If your inference is based on age of the population and formal occupation, a lot of things in health are the age of the housing stock to look for lead kids, or income or other things. Instead of making it sound like we are being machine driven, go for the hand count instead of the machine count. We are really trying to inform better the collective inference and intuition.

When you talked about not communicating well -- this is fairly common -- not communicating well across public and private, what you really want to do is tap into the things they have that are driving their intuition and inference, and add them to the things that you have that are driving yours, and perhaps there is something to be gained by that.

DR. LUMPKIN: Although I think we have to be careful about that, because again, measurement is very important. When you ask physicians how often do they counsel patients on smoking, they give 80 percent, and when you actually look at the records, it is 40 percent. So that is one of the important parts of measuring.

Marc, I'm going to ask you a question, and then I'll ask John a similar question. If you were asked, because Christmas is coming up, that there were certain things about your county that you don't know about that you would like to know, what would be on that list?

MR. KOLMAN: I had a conversation about that just this morning, in fact. I had mentioned this one issue since we released our community health assessment to the community, we have the sense that cancer rates are out the roof in Person County. I think that is justification, because we have the number one relay for life in the country, so that is justification for why do we have the biggest relay for life per capita.

DR. SONDIK: What is relay for life?

MR. KOLMAN: It raises money for cancer, the American Cancer Society. They walk around the track for 24 hours, whatever. So people are saying, why do we have the biggest relay for life money raiser per capita. So I luckily got the most recently health statistics and said, we don't have our cancer rate. So I have gotten all these calls going, we need to find out what the truth is on cancer in Person County.

So I need to get some more detailed statistics and figure out how to present it in a way that convinces the community of what they will buy as a reality of what cancer incidence is in the community. So just reiterating statistics out of there somehow has not done that job yet.

Another cut I would really like is environmental factors, so between the manufacturing that happens in the county and we actually have a big issue with septic systems and are they impacting the public's health. I don't really know how to get at those two questions.

DR. LUMPKIN: And connected to that, if you were given the choice, because you had one extra dollar to spend, would you spend it on collecting more data or hiring somebody to do more analysis?

MR. KOLMAN: Probably more analysis right now.

DR. BOOKER: So would I.

DR. DEERING: I had a question. It is not data and not technology. I come from the NHII side of things. It is a question of self identification.

Part of what underlies the NHII vision at least is a certain concept of the role of the public health, and that public health is not just, as you have described yourself, a provider of services, nor is it just shoveling data vertically and horizontally, but that it is a much more proactive link to -- that you are not just providing services and you are not just moving data, but that you are proactively connecting with the individuals in your community and the providers in your community.

Clearly, we have heard that that really isn't usually the case in public health. I guess I would like you to say, number one, did you see that in our report? Did it hit you that we were making some assumptions about the way public health is viewed and viewed itself, that may not ring true and if so, what do we do? Or am I reading more into it?

MR. KOLMAN: I agree with your assessment on where public health could lie. In thinking about the NHII report, I don't recall getting that impression. I'm not saying it is not there, but it is not what I recall.

In some way, to enhance that image -- again, I of course talked to the North Carolina experience, and public health departments in other places are very different than ours. But what I see as a pretty key role in the community.

DR. LUMPKIN: John, I wanted to ask you, you already told us one of the things you'd like to have that you don't have, which is a trauma registry. What else would you say would be something that you would see as being important for you to know about North Carolina that you don't know?

DR. BOOKER: Well, as Marc was talking, I was sitting here thinking. What if he had the handle on how many people had been seen by physicians or other providers in Person County for diabetes, for management of heart disease, for all those kinds of things?

Obviously, one big next step would be access to the medical care sector, to the private and managed care organizations' data about patient care. Chris earlier I think made a good pitch for years of potential life lost, but we would also like to know what happens along the way. We know quite a bit about what happens to people at birth, and we are continuing the effort in North Carolina to try to through linkage compile a longitudinal data set for children up to entering school, up to and including entering school.

Many things happen that all don't result in death or hospitalizations, so there is a big void of information out there. That would be, given the choice, probably high on our list of things to add.

So the interest in the medical record and the -- I don't want it to sound as if I'm saying that is of no interest to public health, but the energy for that is going to come out of the health care sector. We are secondary. It is important that we be kept in mind, and that access to the information in those systems be made available, but the barrier there is not the technology. That was the other point that I wanted to make.

Beyond that, I think simply more resources at the local level, because it may be intuition or it may just be keen observation. One of the practices we have in North Carolina is to run all of the data through local people, with more or less success, primarily because people are aware of that vague and empty look that you can get from local people when you present them with a lot of data about what killed folks last year. It is not a very revealing piece of information unless you had many more things to it.

DR. SONDIK: I'm glad you mentioned the process side, if you will, because I don't think it has been getting enough attention. I don't think it has been getting enough attention in general.

One of the hopes for HIPAA is that the standardization that will come about through the HIPAA process would enable the transaction information to be more useful to public health. That has been a kind of a theme in the department. Certainly it is a theme in NCHS, although I'm not sure we are making great strides toward that. But we have made great strides in actually sitting on the committees that work on standardization, which I think is a very important step.

My question is, if you could get your hands on some of this transaction data, whether it is from HMOs or from Blues or wherever it is you're getting it, for that matter, if you could get it from HCFA, which I guess you can get, which you find it useful? And what would it take for you to make use of it? Would you do it in part of it? Would you question how good it is? Or say we'll accept it.

One of the questions that always comes up is just how good is this data, and I'm not going to use it until I know how good it is. On the other hand, we're not going to know how good it is until we start trying to use it.

DR. BOOKER: In some cases, we know how good data are or are not, and we still use them, because they are what we have. I think that is always a consideration.

Obviously, the first thing that we could do with that information is to fill in the picture descriptively, since we know a lot about Medicaid clients, we know a lot about CHP enrollees. In some states, state employee benefit plans and others help to try to put together the picture, but there are parts missing.

Beyond that, we get back to the issue of children moving in and out of Medicaid, moving in and out of CHP, moving into managed care, moving to other places, and the necessity to take full advantage of the information to be able to follow people over the life course and through the events in their lives. So that is I think a large goal and probably a long term one.

DR. LUMPKIN: The third member of our panel is here, and so now that she has had a few minutes to settle in, if you would introduce yourself. We're going out live on the Internet, so if you would introduce yourself into the microphone and speak directly into it, it doesn't pick up well from a distance. And, welcome.

DR. UPSHAW: Thank you. My name is Vaughn Upshaw. I am at the University of North Carolina School of Public Health. I work in two capacities, one as a faculty member in the Department of Health Policy and Administration. I also coordinate the leadership program in public health, where we operate DRPH for mid-senior level people who want to return for a terminal degree in public health with an emphasis on practice and leadership.

Previous speakers have already done my work for me, but given that I put some time in putting a presentation together, I'll give it to you anyway, and I will encourage them to speak up when I get it wrong here.

I wear a hat that also I think is not represented here. I am currently president of the National Association of Local Boards of Health, so I represent a group of citizen policy makers at the local level, and certainly bring an interest and concern about health data that is available for good policy making, and would be happy to talk to that at the conclusion of this, if you have questions.

I did have an opportunity to read both the vision and the infrastructure report, and I have been coached to use the language I think you are familiar with here, so I will stick with that, unless there are any questions about which report I am referencing.

First I will speak to the vision report. One of the things that was most important for me as I read this was that it really did strongly support a coherent framework. Having worked across this state and with many other states, particularly with local public health agencies, I recognize that the lack of coherent understanding of what data do we need and what should we be looking at to diagnose the health of a population is a problem. If this document begins an effort to get a coherent framework in place, I strongly support that.

It is important for localities to think outside of traditional geopolitical boundaries. Counties are not sufficient for many localities in making determinations about health statistics. Populations do not use county boundaries in seeking care. Many rural areas actually are tracking people across county boundaries, whereas in urban areas such as New Jersey, a county may represent many cities, and those municipalities want unique data that they can use for their own purposes.

So as you think about a framework, I think it needs to recognize the complexity of these multiple geographic jurisdictions and how people might pull data that would be relevant for them, is an important consideration.

I have said I come from health policy and administration, and I'll put an emphasis in my next point on administration. As we talk about health data and increasingly are looking for evidence of things that we do making a difference, we have got to have a way to link financial and health statistics. I heard you just asking John about reimbursement data; unfortunately a lot of public health activities don't have a direct reimbursement stream, but that doesn't mean that they don't cost us. So finding ways that we can look simultaneously at the cost of doing business to effect population health and at the same time look at the outcome data on what health status measures are, I think we can begin to build some mechanism for local policy makers and administrators to say, if we do this and commit these resources, here is a fair expectation of what we might yield as a result.

Using web-based systems I think is a critically important part of whatever we do next. Having things in a central clearinghouse where individuals, businesses, providers can draw down from a web-based data source I think is important.

I will emphasize here, and you will hear me bring it up again the importance of looking at existing tools that local health agencies currently are using. You talked with Dr. Booker about HETUS data; JCHO has its own sets requirements.

The National Association of County and City Health Officials, also known as NACHOO, has been working in cooperation with a number of groups under contract with the Centers for Disease Control to put together a new strategic health planning tool called a MAPP. It is Mobilizing for Action Through Partnership and Planning. As I looked at the infrastructure report and particularly the community component of the infrastructure report, much of that is already captured in the MAPP tool. So I would strongly encourage you to look at existing products that are out there, and build in recognition for what you are trying to achieve through those mechanisms that are already in place or soon to be in place.

Regardless of the tools, local folks need guidance in analysis and interpretation. We can collect data. We know how to do that. We don't really know how to analyze and interpret that data very well.

I go back to my mention of the web-based tool. If there were standardized formats or suggested reports, if you could pull up a model of here is how data would be analyzed and here is an interpretive report, an individual at a local agency could then draw their own data in and create a mock against that model, would go a long way toward helping facilitate that data that is collected actually being moved into an arena where it could be considered for policy.

I also very much supported the end of the vision document's mention of needing broad collaboration and support. It is clear that making health related decisions extends well beyond just health providers. Back to my hat with the National Association of Local Boards of Health, we are talking about business people, educators, retirees, county commissioners, any number of folks who are outside the health care provider arena who are making big decisions about what kinds of things are going to be done or not done in terms of funding, manpower, programs and the like.

Just conceptually, I needed to put my hands on at the local level, who is asking for data, who is providing data. I have this little bubble chart here just as a frame of reference. Locally, you are having to deal with individuals who want information about their own health history. They may be providing that. They may also want access to it. You have providers who are in and out. You have community-based organizations, the media, government agencies, community at large, businesses and education, all of which are looking for and providing information. So how a local health officer or a local health system collectively manage those numbers of groups and their unique requests and needs I think is a critical part of what kind of infrastructure you put together, and can people at the local level access it in ways that make sense for them.

I will now turn a little bit more to the infrastructure report. There is a mention in the report that medical care systems need to be able to talk to one another. I would want to emphasize that it needs to go beyond medical care systems to include more broadly public health and health and human service systems. As you begin looking at the MAPP document and others that we know about, determinants of health, you have got to take into account income disparities, racial and ethnicity, educational achievement. Many other things beyond just health data contribute to health, and so how the system recognizes and draws in that data I believe is important.

I had mentioned the need to build an evidence base at the local level. There is a tension between having sufficient numbers to get accurate statistical data upon which you can draw conclusions that you have got to be able to act. You cannot stop your programs because you don't have big numbers.

So measuring health at the local level and building that evidence base is going to require some creativity about tools and approaches, small area analyses, running multiple year averages, helping localities figure out how can they make a strong evidence based decision using the kind of data they have available. I don't know how you do that, but I know that it is going to be critically important as the mantra for evidence based decision making becomes more prevalent and county commissions, boards of health and others are going to be asking their local programs, where is the evidence that this works.

I had already talked about having model reports and using existing tools, so I won't dwell on those further.

At the local level, we are talking about collection, management, analysis, interpretation and dissemination of data. Locally, there needs to be a common language and framework. So back to the vision statement, how you play out what that framework looks like, and what the different responsibilities are for the partners locally in providing as well as accessing and using data, I think we have got to start with a clear vision of where we want to end up, before we start collecting things just because we can. It is important to know why we are doing it.

Convenience and consistency are going to be important for locals. They have already got a lot on their plates and adding more in terms of collection without building it into existing programs that they have already got to collect data for, or making these things convenient for local folks to use, it won't help if it is just one more layer.

I think the only other thing that I haven't spoken to on this particular slide is just reiterating the need to protect privacy and confidentiality, particularly as you become more local and you are talking about smaller numbers, where you cannot easily aggregate and make people anonymous. It is an issue nevertheless, back to being able to track and identify people so that you can improve their health status and make sure that we are managing the population, we have got to have access to that. So important considerations, requirements for training, encryption, all sorts of other things, have got to go into play there.

Work force training and education in data management. Basic training for local public health people as well as others in the health care arena and health and human services what is a ratio, what is an incidence, what is a prevalence. Very simple statistical and epidemiological procedures they can run locally and understand what they mean will help tremendously in how people begin to analyze and interpret.

Just a tidbit. If you are not already aware of the public health work force, in the neighborhood of 75 to 80 percent of people employed in local public health agencies do not have a background in public health. So the kinds of things that we are talking about in terms of biostatistics and epidemiology often are not a part of peoples' training and background.

Software systems and hardware that can interface with one another. I work with counties where we typically have HSIS, we have health data in one system and then we have county finance handling all the financial data, and never the twain shall meet. I think any reference that you make back to how do you build that evidence base back -- I know you guys have the health side, but that finance side at some point does need to come into the mix. If we continue to leave it out, we are going to continue to find ourselves in an upward spiral without knowing why.

There are lots of programs that require data by federal and state governments. I would just encourage that as you create a large system, you think about how each of those programs have their needs met within it, rather than having to have parallel silos.

I spoke earlier about flexibility and defining boundaries, being able to think about regions, about parts of counties, municipalities unto themselves, small area analysis, other appropriate political tools, and helping staff learn how to do those things.

I have a strong bias toward quantitative approaches when looking at local health data, because so many things are contextually bound. It is important to be able to draw down secondary and large data sets in order to provide a foundation for what you're doing, but if that is not coupled with a clear understanding of where you are working and what the culture and dynamics and priorities are there, the interpretation will fall flat. So how we at the national level and at state levels encourage and facilitate that combination of methodologies in order to make the most meaning out of the data that we have I think is critically important.

Comparative data. Another tool that is out there. It has received some criticism, but I think it is a first step. The community health status indicators project, public health foundation and Health Resources and Services Administration have collaborated, and there are available online county by county reports. A county can pull up their own health data, compare that with 2010 objectives, compare that to national and state rates. Then there are some manufactured data for peer counties across the country, so you can see where you are compared to others who have similar size, resources, et cetera.

I think comparative data is important. The more that this process can provide that, the better. Sample reports, how do people take the information and put that into meaningful presentations, and then how do they use that information to begin making arguments for how this is effective, both in terms of health outcomes as well as resource use, very important at the local level.

Disseminating information, another tool that is out there that is not being widely used by local health departments because it is costly, but I think has a lot of potential value. The health outcomes through health care forum, outcomes toolkit is a web-based software system where they have a central server. You can draw down from that server lots of secondary data, but through allowing users on board in a local community, the hospital can put their data in, local providers can put their data in, the health department can put its data in, and then anyone can go in and actually see the data that is there, can generate reports. You can decide who has access to manipulate the data, and it becomes a public forum for anyone who is interested. So if the fire department needs to write a grant and they want to get some basic information on injury, they can go pull that down and they can get ER visits, they can get other records, and it becomes a web-based vehicle for everybody at the local level to participate in and draw down from. I think tools such as that are a good model for how this process might want to evolve.

DR. WEINZIMIER: It's called outcomes toolkit?

DR. UPSHAW: Outcomes toolkit, health care forum. It is the American Hospital Association, the health care forum. Lumin and Associates and some other folks helped to develop this. It has only been out for about a year. I have seen it piloted. They are looking at -- I'm not sure what the final arrangement is going to be, but back to my discussion earlier about the MAPP tool with NACHOO, NACHOO and Health Forum had actually talked about how they would consolidate the strategic planning tool at the local level with an online database support system.

I don't know how far they have gotten with that. It requires a lot of manpower and resources to figure it out, but you all have right now the power to move things in that direction. So technologically it is there. I think part of it is will and just directing ourselves to saying yes, that is a reasonable way to be looking at how to make things available locally that everyone can access.

Dissemination is going to be an issue regardless. Part of it is something that we are actively taking on and trying to put out there, or making it accessible say in a web-based platform where then individuals who have an interest in that data can actually go in and draw it down in the way that they need it. So dissemination is one way of getting out what we think people need to know, but also making sure that things are available for others when they decide they need to know it.

In summary, I think we have got a great opportunity to be working both from the bottom up and the top down. The infrastructure does need to build on evidence base at multiple levels, and I emphasize at the local level, because if we are not -- as my final statement says, if we are not getting it to work locally, it is not working.

I think we have got many tools that are out there that can be drawn into this. Reducing the duplication among them would make a great step forward for locals.

I think I'll stop with that and entertain any other questions, and invite my colleagues to fix anything I might have gotten wrong.

DR. LUMPKIN: We have time for a few more questions, and then we have someone from the public who wants to testify also. So questions?

I've got one. I always seem to have one. Am I to infer from what you are saying -- and it depends upon -- and I'm not sure how to resolve this issue. You have identified a problem which I think we have also identified, which is the difficulties that many local jurisdictions, particularly the smaller ones, have in analyzing data. Is this an issue that is best resolved through more training or is this something that ought to be designed into the information systems so that they have the capability of providing the analytical tools, and obviously the hypertext, the ability to allow people who have less training to actually learn while they're using it.

DR. UPSHAW: Yes. My bias is toward the latter. I think it is too important. You can't have people using only predesigned tools. As they become more sophisticated they will start asking more difficult questions that those predesigned tools will not necessarily be able to anticipate or accommodate.

But just getting people up to the first rung of the ladder and having some pre-set way back to the health care forum and the outcomes toolkit, where you can essentially lay out a forum where individuals could go in and insert their local information, get the calculations and then see a model report for how that would actually go in and be analyzed, interpreted, printed out and disseminated.

I think the more it is lockstep with certainly some decision tree choice options in there, but dummy proof I guess is what I'm looking for. Speaking for myself, does this number go in the numerator or the denominator. Making those kinds of decisions, I think the more assistance that is out there, the better. Local health departments by and large do not have the resources to hire biostatisticians and epidemiologists. So if they don't have that, they have a health educator or a nurse that is interested in trying to generate some of this, the more it can be created in such a way that helps them get at least a basic analysis done, the better.

DR. LUMPKIN: Who do you think, in your vision of developing these tools, should develop them? Should it be the federal government? If so, what portion? Should it be state government?

DR. UPSHAW: Well, I guess there is some shared responsibility here. I have been a part of a lot of things that are collaborative, so I don't have a real strict line about where it falls completely. I think states really are the primary arbiter and responsible parties for health statistics, but uncrossing at the federal level with large federal programs and the national centers wanting to take the data that is available at the local and state level and make sense of it, they have got a stake, too.

So I think I would look for some shared leadership here. That doesn't help, does it?

DR. LUMPKIN: I'll ask my question differently, because obviously I think we agree this needs to be a collaborative project. Who should pay for it?

DR. UPSHAW: I think again, I would be looking at a couple of different things. I think I would begin with looking at going to some foundations for pilot funds. I think you have got a good opportunity with something like the Health Care Forum and the outcomes toolkit as the model you can look at, and begin doing some initial work, to see what would make sense within a state or a handful of states that are further along. I know for example North Carolina have fairly sophisticated data systems that might be ready for creating a federal link. Other states are not quite to that point yet.

But as you begin to develop models, then there is some incentive for states, if you begin demonstrating that they work, to put money into doing those things that will yield options for them, so incentives for legislators to see that this will actually product some value for them.

Right now, the voter registration process is one that is not perfect when it comes to funding, so a lot of the machinery is pretty antiquated that is being used right now. Now that we see the problem we're in, everybody wants the best machinery.

DR. LUMPKIN: I thought hand cut was the best.

DR. UPSHAW: Well, either way. There are still counties in North Carolina that, if you want to know who was affected by the most recent outbreak of pertussis, you go in and they have them on index cards in a little recipe file. So hand counting is still being done.

But I think there are initial steps that I would take with some external support to begin building some value. Kennedy-Frist. We've got some other mechanisms in place right now where the resources are there.

DR. WEINZIMIER: What about working through universities, putting on your hat as a university person? What if the National Center for Health Statistics or other federal agencies work through your school to build a curriculum or something like that to help local health departments?

DR. UPSHAW: It doesn't get around the question of who is paying for it. Universities run on money, so it would have to be contractor grant supported to get that.

DR. DEERING: There is an SBIR grant that is used in another area, small business innovation, that was at least for a long time the only thing that was funding potentially commercial innovative technology applications.

It occurs to me, I'm not sure that has ever been applied to public health. I don't know if CDC -- do you have SBIR grants?


DR. DEERING: Are they funding things like this?

DR. SONDIK: There have been some. I know NIH is funding some.

DR. DEERING: NIH has had a lot, but I'm not sure how many of them have been at the local public health level.

DR. SONDIK: In the public health area, they have funded some.

DR. SONDIK: Actually, no one today has talked about the role -- unless I missed it, the role of the university, schools of public health, medical schools or whatever in the funding. What were your thoughts about that, about the role that they can play across the spectrum here? I think you have touched on many, many points where they could play a role.

DR. UPSHAW: I know I am being recorded, so I am trying to figure out how to be really careful about this. I think universities can do a lot more than what they have done with regard to helping translate and communicate how health data can be used, what are appropriate ways of collecting and managing data at the local level.

The tendency in large universities, research universities such as the one where I work is to do the large studies, and national data sets are the gold star or clinical trial data sets. In this forum, we really don't have that same attractiveness perhaps for researchers in many cases. There is not a clear agenda for what needs to happen. I think this process that you are currently a part of may begin to shape that agenda and with that will come more interest on the part of researchers for how would we use this.

But there is still a fair gap between the health services research agenda which I think is very well organized and well financed through AHCRQ and other organizations such as NIH and the like, versus population based data and population based evidence research. That, we are still trying to figure out what the boundaries are, what kinds of things would be of interest, how to make that fundable and meaningful.

I think the faculty are probably going to be receptive to that only when it is clear that there is money and there is a goal in mind.

DR. LUMPKIN: Bill, do you want to jump in at this point?

DR. KALSBECK: My name is Bill Kalsbeck. I'm from the School of Public Health at UNC-Chapel Hill. I'm not sure in what capacity I'm here. I just am an interested public person, I guess.

I think the points have been made. The only thing that I might mildly disagree with is, I believe that in spite of the fact that there are some perhaps constraints in trying to do big-time research with local data sets, there are nonetheless as many interesting research issues, as least in my area, which is in the methods area to study, whether you are talking about a locally gathered data set of some Hispanic women in a migrant farmworker study or some big national survey. There are measurement issues that one confronts there.

The same kinds of things that come up there should be of concern to anyone who is using those data as well. So in that sense, I think the issues are there. The question is, and here is where I agree, I think sometimes issues like that might arise in more of a state or local context, where we say it would be nice if we could work on that, but it is just one of those things, we have to get on with what we have to do, and that is collect some substantive data, and we don't spend the time on the methods issue.

What I think we need are mechanisms whereby if someone who is doing work on data relevant to the state of North Carolina, and they say they recognize this as an important issue, there needs to be some mechanism by which a researcher can say, hey, we need to take a look at this and study it. Right now, there may be some ad hoc mechanisms, but I think something that is a bit more systematic needs to be in place.

DR. SONDIK: To pick up on where Mary Jo was going, I think biomedical research, one continues to learn and learn and learn. There are products that come out of that. It seems to me if this worked right, there would be products that would come out of this. But that means we need something in place to be able to deal with those products, market them, put them in place, which is a major role I think for the private sector.

We heard some of that this morning at least in one area. But I think it is an area I am not supposed to be testifying in. Let me ask you, do you think this is a significant role for the private sector, in other words, a partnership between the research community, the applications public health community, as a middle role between those two?

DR. KALSBECK: I guess my central point, and this is admittedly from a methods perspective, there are issues that are of relevance to what I do and to what other methodologists are interested in that one faces, whether you are dealing with a major national data system or a local study. They arise in any context. Whether it is public or private, it matters not.

I think what is often lacking is the wherewithal. It often boils down to resources, money, to be able to fund things like this, some mechanism whereby if something comes up in the process of collecting substantive information, one wants to have the ability to study these things more systematically.

It may be a partnership in which -- as what we are trying to do now with a center grant that we have been funded by Ed's agency is to try to identify issues that are of relevance to people who are doing research, and to provide a means by which we can study them. They need not be limited to the public sector. Private could be there.

What we are trying in the activity of our center currently is to forge a partnership between in our case, people who are doing research in the substantive area and methods people, so that we can in partnership in that form learn some things to facilitate the research that they are doing, and to address important issues as we see them.

DR. LUMPKIN: Thank you. Any other questions? I'd like to thank the panel very much for joining us, and Bill also for sneaking in and taking the time to come over from Chapel Hill.

This will be our last hearing this year. We have another one in January in D.C. I think after this discussion we probably want to put NACHOO on the list for sure. We will be having a hearing in January. We also will have a full committee next week. I don't think that we are scheduled to have any breakout time, but we may want to look at lunch one of the days to round out our thoughts on the attendees in January.

DR. BLAIR: The breakout time is seven in the evening.

DR. LUMPKIN: Oh, seven in the evening at your place?

DR. DEERING: I had another question. We had talked about reserving an all-day session for our two work groups. I have on my calendar December 19, and I just wanted to confirm, had we locked that in, that our two groups are reserving the full day for that?


DR. DEERING: Is Kepa still here? Because I think he said that didn't ring a bell with him. So I wonder if we need to reaffirm that.

DR. LUMPKIN: Yes. So I'd like to wish everyone --

DR. KALSBECK: I just have a clarification. I apologize if I am breaching protocol here. I came with something to -- a statement to make. Was this a private session that I interrupted?

DR. LUMPKIN: Oh, I'm sorry. No, no, no. I thought you had integrated your --

DR. KALSBECK: I come to support my colleagues here. However, I did have a couple of things that I would like to say, and if I could, a statement to read.

DR. LUMPKIN: Please.

DR. KALSBECK: What I would like to do is to describe two initiatives which I believe are needed to help shape the future of the nation's data system and to make it one that will more fully meet important information needs, particularly those having to do with health outcomes.

My experience in recent years has been that I have heard a lot of discussion about the need for information on health outcomes, and I have heard various discussions about things that we need to do to facilitate our abilities to study health outcomes. But I think there are a couple of things that are lacking that this group might wish to consider.

The first of these is to establish a separate broadly focused and semi-independent health information planning board to provide overall focus, direction and coordination to the national health data system, patterned after the Australian Institute of Health and Welfare, the Canadian Institute for Health Information and comparable panels in other countries. This board would serve a central role in forging an ongoing partnership of health providers and insurers, as well as public and private users and producers of health data.

I understand that there are various groups that have been established at the departmental level and so forth to look into some specific issues, but I am looking at a group that is quite overarching.

The board's main function would be to orchestrate and to broker the production and dissemination of population-based health statistics in the United States. Some of its major functions might therefore be to first identify and organize information needs at all levels of the health care system, establish a working collective partnership with all producer agencies that contribute to the health data system, third, negotiate a clear set of information gathering roles for each member of the partnership, fourth, develop and implement a long term plan to promote a collection of standardized high quality health data, fifth, develop and implement innovative new ways to gather needed health data, sixth, promote the privacy and confidentiality of all publicly acquired data and explore ways to deal with accompanying threat, seventh, support the development of new methods to effectively use collective data to answer important health related questions being asked by the health community at all levels, eighth, promote access to all existing health data by public and private researchers and policy makers and finally, explore new outlets and uses for health information based on current information technology.

A second initiative I propose would be pursued to better enable us to better understand health outcomes, in particular those things which affect these outcomes. The possibility of study of these types of interrelationships are great, covering the full span of a person's lifetime and crossing the full spectrum of a health care system, for example, in considering the impact of exposure to second-hand smoke as a child on respiratory ailments in adulthood, the role of stress during middle age as a determinant of cardiovascular disease in later life, features of the health care system that affect the health outcomes of its recipients, the long term effects of medical treatments, to name just a couple of examples.

To be specific, I propose the creation of a comprehensive national longitudinal health study to better understand the behavioral, treatment and policy effects on whole life health outcomes. Developing the capacity to do such a study might be accomplished, largely by applying existing health data gathering apparatuses to a general population sample based on a rolling cohort of births.

A birth quote-quote in this context might be interpreted initially as participation in a population-based household sample, and then subsequently to natural birth and immigration. The sample of births would be followed continuously until death quote-unquote, occurs naturally, or by out-migration to another country.

After gathering baseline data, this rolling cohort sample would be followed up to obtain a full range of needed data tied to health related attitudes, behavior and provider of care. Besides continuously monitoring the health experience of cohort members and families, linkable health services data would also be continuously gathered from the health insurers and providers.

An underlying purpose of this ambitious cradle to grave sample derived portrait of the American public would be to expand the budding synergy among existing health surveys into a broad and significantly more important source of health information. Each member of this reoriented confederation of the nation's most important health surveys would become an important component of a more comprehensive design paradigm, thereby creating a larger whole whose information benefits are significantly more than the sum utility of these surveys by themselves.

A key to the success of this longitudinal cohort study would be the creation of a reasonably non-threatening division of information gathering roles among those agencies with a vested interest in these roles. At the least, each agency partner in this venture must be allowed to claim major ownership to an information niche.

For example, one might align CDC and NCHS with vital statistics, morbidity and general health status. AHRQ with health services delivery issues, CDC Atlanta with health surveillance and promotion, HCFA and SAMHSA with aspects of health policy analysis, HRSA with the capacity to provide health care, and NIH with treatment and clinical efficacy.

General responsibility for developing survey instrumentation, data analysis and information dissemination for surveys addressing each information focus would be the responsibility of an agency partner, while developing the design and gathering the data in this longitudinal study would fall under the preview of the health information planning board described earlier.

By dividing responsibility in this way, an agency with an information focus would not be consumed by data production, but could focus its efforts on developing suitable data measures, converting data to information and disseminating the results.

Beyond the health data of particular relevance to its own mission, each agency would benefit by having ready access to lifetime person level data that would enable them to explain health outcomes or evaluate efforts designed to improve these outcomes.

For example, AHRQ analysts would eventually have a complete picture of health care utilization and outcomes to enable them to assess the relative efficacy of various service delivery models. To be sure, a longitudinal health survey like this one would be without precedence in terms of size, scope and complexity. Coordination, cooperation and careful planning by all agency partners therefore would be the key, with the health information planning board playing an all-important facilitative role.

For instance, to field a suitable general population sample to follow might involve first using a household sample of a wave of a large area sample, such as the current population survey or NHIS, provide an initial cohort of individuals and baseline data. Second, periodically supplementing the baseline cohort with samples of natural births and international immigrants in the sampled areas, and then third, periodically recontacting cohort members to collect data on their intervening health experience and any interactions with the health care system.

Other challenging but solvable problems would be faced and fielded in producing findings from the national longitudinal study. At the planning stage, the content of existing health surveys would need to be re-evaluated and perhaps re-prioritized around established information needs. Some survey measures would need to be developed or improved to meet new or emerging information needs, for example, the quality of life, gauging the quality of provider care and so forth.

Creative technological solutions would also be needed to preserve the privacy and confidentiality of all gathered data. To minimize cohort decay, followup non-response and the Hawthorne-like effects of respondents' health experiences, new ways to easily, efficiently and innocuously extract data would need to be developed. With technology evolving at the present rate, this is hardly a pipe dream, I would claim.

To avoid measurement softness in portions of the respondent data record, uniform standards and definitions would need to be established and enforced for all study questionnaires. Finally, the sample would need to be large enough to facilitate the production of acceptable estimates for states, sub-state health planning regions and other small areas, including specific racial and ethnicity groups of interest. Thus, new or improved statistical methods to produce small area estimates would be needed.

Continuing growth in health care costs and the complexity of the health care delivery process and in the public's focus on matters of health makes having timely, high-quality and relevant health data an essential part of the nation's future information landscape. To be assured that these data are fully relevant, both of the initiatives I have proposed I believe need to be given serious consideration.

DR. LUMPKIN: Thank you. We will certainly include that within the deliberations of the work group on the 21st century vision.

At this time, I have to leave to catch an airplane, so we are going to adjourn. I'd like to thank everyone who participated and those who gave presentations and a statement. We will be preparing the NHII reports. We are again shooting for some time in the spring, and the 21st century report will be prepared sometime later in the year 2001, probably fall or winter. So we appreciate that.

(Whereupon, the meeting was adjourned at 3:50 p.m.)