Minutes of the July 24, 2002 NCVHS Workgroup on National Health Information Infrastructure Hearing

07/24/2002

Department of Health and Human Services

National Committee on Vital and Health Statistics

WORKGROUP ON THE NATIONAL HEALTH INFORMATION INFRASTRUCTURE

July 24, 2002

Hotel Monaco
Chicago, IL

MEETING MINUTES


This hearing by the NCVHS Workgroup on the National Health Information Infrastructure was convened on July 24 in Chicago, Illinois. The meeting was open to the public. Present:

Committee members

  • John R. Lumpkin, M.D., M.P.H., Chair
  • Jeffrey Blair, M.B.A.
  • Simon P. Cohn, M.D., M.P.H., FACP
  • Clement Joseph McDonald, M.D.
  • Edward M. Shortliffe, M.D., Ph.D.
  • Kepa Zubeldia, M.D.

Staff and liaisons

  • Mary Jo Deering, OPHS, Lead Staff
  • Eduardo Ortiz, M.D., AHRQ
  • Steve Steindel, Ph.D., CDC
  • Michelle Williamson, NCHS
  • Marjorie Greenberg, National Center for Health Statistics (NCHS), CDC, Exec. Secretary
  • James Scanlon, ASPE, Executive Staff Director
  • J. Michael Fitzmaurice, Ph.D., Agency for Healthcare Research and Quality liaison

Others

  • Katherine Jones, NCHS
  • Debbie Jackson, NCHS
  • Jackie Adler, NCHS
  • Stan Edinger, AHRQ
  • John Lamoureux, Baltimore City Health Department
  • Michael Millenson, Markel Foundation
  • Meg Richards, Illinois Department of Public Health
  • Marc Overhage, Regenstrief Institute
  • Ron Nath, Apelon
  • Ted Eytan, Group Health Cooperative of Puget Sound
  • Warren Williams, CDC
  • Vik Kheterpal, GE Medical Systems
  • Jim Cimino, Columbia University
  • Barry Gordon, CA Cancer Registry
  • Philip Marshall, WellMed
  • Kathryn Bingman, Cerner
  • Sandy Fuller, AHIMA
  • Oscar Morales, EPA
  • Peter Waegemann, Medical Records Institute

EXECUTIVE SUMMARY

The NHII Workgroup is monitoring the implementation of its recommendations and determining where further action is needed to promote the NHII. This hearing is the first event sponsored by the Workgroup for that purpose.

POPULATION HEALTH: REGISTRIES

  • John Lamoureux, Baltimore City Health Department
  • Warren Williams, Division of Cancer Prevention and Control (DCPC), CDC
  • Barry Gordon Ph.D., California Department of Health Services

Mr. Lamoureux described the Baltimore Immunization Registry Program (BIRP) and its data collection and exchange, and the barriers to full implementation. Immunization registries are computerized tracking systems that collect and disclose information on a jurisdiction, and they contain decision-support modules. The BIRP started in 1995 and covers children under age six. Participation by providers was initially voluntary and is now required. Mr. Lamoureux cited a long list of areas needing improvement, including data quality, Internet accessibility, uniform functional standards, integration with other public health information systems, cross-jurisdictional connectivity, privacy, and steady funding.

Mr. Williams talked about population-based cancer registries, noting that data standards, collection, reporting procedures and use are fundamental to efforts to reduce cancer in the U.S. Registries exist in all states, DC, and the territories. With CDC?s support, a workgroup drafted an implementation guide for using the HL7 reporting approach as a springboard for further development. Ongoing work is focused on developing a comprehensive medical record vocabulary using the SNOMED vocabulary. Mr. Williams stressed the transitional nature of these and other projects, calling them ?beginning steps? toward the NHII. He said the registry community and industry standards groups need to be educated on the use and benefits of controlled vocabulary and coding schemes.

Dr. Gordon said that his state rapidly implemented statewide reporting and data collection for its cancer registry at least partly because of the provision of clear electronic standards and software. The cancer surveillance standards effort has been a success in his state and at the national level, which he attributed to a combination of commitment, tools, and courage. Remaining issues are harmonization problems, lack of well-structured reporting, and limited implementation of HL7 standards. After describing several HL7-related projects in California, he outlined the steps toward more consistent and widespread structured standards and reporting.

In the discussion period, members expressed interest in the role of resource limitations in registries? choices, the conditions that have facilitated or hindered the development of registries, the feasibility of an interchange solution that would permit differing implementations but the exchange of the same core data elements, ways of bringing registries closer to the NHII vision of enabling multiple uses by multiple users, and ways to get people to use standards. Dr. Deering asked these and other panels to send her their priorities for the federal role in their arena.

POPULATION HEALTH/NHII INTERFACE

  • Margaret Richards, Ph.D., Illinois Department of Public Health
  • Oscar Morales, EPA
  • Marc Overhage, M.D., Ph.D., Regenstrief Institute & Indiana University

Dr. Richards described three systems that exemplify the state?s population health dimension: the Birth-Related Data Project, the Illinois version of NEDSS (I-NEDSS), and Tracking Our Toddlers? Shots (TOTS, an immunization registry). The state decided to develop its own version of NEDSS because of issues of timing and adaptability. It is committed to interoperability with NEDSS. The overall purpose is to allow providers and labs to report cases or open cases for investigation simultaneously with local and state health departments.

Mr. Morales reported on EPA?s information exchange, data standards, and Web-based registry activities, and expressed the agency?s willingness to participate in work related to the NHII. EPA expects to participate in the environmental health tracking network being developed by CDC. It can share its experience in partnering with state agencies to build the National Environmental Information Exchange Network. EPA has a vigorous internal implementation program to get people to use the standards it has developed.

Dr. Overhage described the Indiana Network for Patient Care (INPC) in Indianapolis, a secure, operational electronic medical record network. Focusing on its population health uses, he noted that population health data are needed not just for public health uses but also by payors, accrediting bodies, and others. Among other things, INPC is the custodian of the participants? separate databases and helps develop agreements on how data are to be used. Building on data standards, notably HL7 and DICOMM, has made this possible. In the population health domain, it serves as an immunization registry, electronically reports reportable conditions, and does ?some crude surveillance.? He concluded that 1) INPC lends credence to the NHII idea that ?data can be repurposed and reused,? 2) standards are critical, and 3) a little work can go a long way with such systems.

In the discussion period, Workgroup members and staff raised questions about interoperability between Illinois and other states, the factors that enabled the success of the Indianapolis venture and the extent to which it is transferable, and privacy protections in that project.

PERSONAL HEALTH: PERSONAL HEALTH RECORDS

  • Philip Marshall, M.D., MPH, WellMed, Inc.
  • Vik Kheterpal, M.D., Medicalogic/GE
  • Peter Waegemann, Medical Records Institute
  • James Cimino, M.D., Columbia University

Dr. Marshall described personal health records (PHRs) and said that several standards are needed, notably, minimum data sets to facilitate data sharing, interoperability between electronic systems, patient control and privacy standards, and consumer terminologies. Minimum data sets provide the foundation of portability, and Dr. Marshall said that about 50 data elements could comprise this set. He concluded with six recommendations for NCVHS.

Dr. Kheterpal commented on the fact that the health field lags behind other sectors in its adoption of digitization. GE Medical Systems? experience in this area is based on the use of a PHR from a provider-populated model by more than 1,000 patients over two years. Different uses of the PHR, leading to different standards issues, include the uni-directional dissemination of information, patient alerts and reminders, and two-way communication with the provider. Major challenges for PHR creators are the lack of a common lexicon, the need for a mature ROI model for adopting entities, and the dangers of ?scope creep and immature project phasing.? To accelerate such initiatives, he concluded with suggestions for federal help.

Mr. Waegemann differentiated among five different types of PHRs. He called attention to confidentiality concerns because of inadequate privacy training for database operators. After noting the multiplicity of PHR initiatives around the world, he concluded that a major issue in the future is going to be reconciling the options in a way that benefits both consumers and providers.

Dr. Cimino said he defines health information to include both patient-specific information and more generic health information, because many messages include elements of both. He described two Columbia University projects, the Patient Clinical Information System (PatCIS) and the Myocardial Infarction Health Education Aimed at Rapid Therapy (MI-HEART). From both projects, the researchers learned that patients can handle the technology; they want their results; they want to collaborate with their physicians; and the EMR and the Web together can be an effective educational medium. The researchers ran into standards issues related to browsers, terminology, identifiers, and access to high-quality information. He concluded by recommending several ?next steps.?

In the discussion period, the group talked about the need and potential for standardizing patient-centered terminologies, the feasibility of voluntary adoption of a patient identifier methodology, and interpretations of the PHR using the NHII notion of a connected infrastructure with patient ability to view clinical data.

INTEGRATION OF E-MAIL WITH CLINICAL AND PUBLIC HEALTH INFORMATION SYSTEMS

  • Ted Eytan, M.D., Group Health Cooperative of Puget Sound
  • Ron Nath, M.D., Apelon, Inc.
  • Kathryn Bingman, Cerner Corporation

Dr. Eytan said that Group Health is implementing a comprehensive clinical information system with accessible PHRs, to be ready in 2003. It is already heavily engaged in secure messaging, its privacy-protecting version of e-mail. He described these systems and practices, saying that one of the standards issues it has encountered is authentication. Health systems wanting to transform health care can assist by critically assessing standards under development and applying them as demonstrations, as Group Health does. A current barrier to the development of standards in the online patient arena is the lack of adoption of online services across the majority of health care providers and plans. He urged HHS to show leadership and promote the use of these capabilities in the field.

Dr. Nath described his company?s thought processes and decisions as it created a consumer health vocabulary. They decided on an eight-grade reading level, and among other topics incorporated alternative health terms because of consumers? interest. He noted consumers? increasing sophistication in using clinical medical terms; these and lay terms form a continuum, with a large area of professional-lay overlap in the middle. Thus his company retained many clinical terms. They found that not only are consumer terms imprecise, but many do not map to professional terms, notably in the alternative health domain. Dr. Nath noted a few applications of the consumer health vocabulary to e-mail messaging.

Ms. Bingman described Cerner online demonstration projects in Winona, MN and Eastern Maine. She asserted that in a ?true personal health record,? information is owned and maintained by the consumer and entered by both consumer and health care provider. She discussed the need for standards for the PHR, identifiers, communications, self-care management, and security, adding that standards alone will not suffice because they do not insure the sharing of information across organizations. Achieving the public health benefits require critical mass in communities or across the nation. She identified proprietary interests on the part of technology suppliers and the absence of a business case for standardization as barriers to standards development.

In the discussion period, Workgroup members and staff expressed interest in the projects? participation rates and the impact on them of the in-person sign-up, the return on investment, and priorities for the federal role and for standards.


DETAILED SUMMARY

WELCOME AND INTRODUCTIONS

Dr. Lumpkin called the meeting to order and asked all present to introduce themselves. Introducing the day?s hearing, he said that the NHII Workgroup is monitoring the implementation of its NHII recommendations and determining where further action is needed to promote them and to develop the concepts in the report. This is the first event sponsored by the Workgroup for that purpose. He welcomed the panelists.

POPULATION HEALTH: REGISTRIES

John Lamoureux, Baltimore City Health Department

Mr. Lamoureux described immunization registries and how they collect data and exchange information, as well as the barriers to full implementation. At present, 78 percent of the nation?s two-year-olds have been vaccinated. CDC is aiming for 90 percent or more. The obstacles include the population?s mobility, changes in health insurance status, and parents? assumption that their doctors are keeping track of their children?s vaccination status.

Immunization registries are a tool for sustaining or increasing coverage rates. They are computerized tracking systems that collect and disclose information on a jurisdiction - which Mr. Lamoureux said is not as simple as it sounds. Registries also contain decision-support modules. He noted the individual and public health benefits of these systems.

The Baltimore Immunization Registry Program (BIRP), started in early 1995 with foundation funding; it now uses Title 5 funds. It focuses on children under age six. Pediatric participation was voluntary until 1998, but this resulted in only 18 percent participation. A marketing campaign that began in mid-1998, combined with penalties for those failing to report in a timely manner, raised the percentage to 85. Physicians who are overdue with reports can be taken to civil court. Health care providers, WIC clinics, daycare centers, Head Start Programs, and schools report electronically or with paper logs. About 60 percent of reports come by paper, but the higher-volume clinics report electronically. The Maryland State Health Department is currently planning its immunization registry, and when it is online, BIRP will be able to tap into other data sources such as MCOs and billing clearinghouses.

Turning to areas that need improvement, Mr. Lamoureux cited data quality, software upgrades, Internet accessibility, increasing the scale and scope of BIRP reporting (they have been directed to include blood lead testing and asked to add other conditions), and restoring linkages to state birth and child mortality reports.

Noting the Healthy People 2010 objective of having 95 percent of the nation?s under-six-year-olds in immunization registries, he said that because of the disparate systems developed by states and the political and financial obstacles to developing a national immunization registry, the best that can be hoped for in the next decade or so is ?to quilt together individual state registries.? A large challenge to states in sharing information across jurisdictional lines is to recruit and retain participating health care providers. This requires ensuring that they get something back from the registry. Other challenges are ensuring uniform functional standards, integration of immunization registries with other public health information systems, cross-jurisdictional connectivity, privacy/ confidentiality/security, and steady funding. Mr. Lamoureux urged NCVHS to encourage CDC-NIP-sponsored work toward meeting these challenges. He noted that CIRSET has been working on HL7 standards. The All Kids Count Centers for Innovation in Health Information Systems also is working toward connecting public health systems. The latter organization is also talking about creating a pool of shareware.

Baltimore?s marketing strategy for the BIRP began with defining the target. It proved very challenging to compile a list of doctors. They developed and distributed marketing materials to health care providers and parents. They invited doctors to report to them in the medium of their choice, and offered to send people to practice sites to capture vaccine histories for children under three. They also offer reminder and recall support services. Mr. Lamoureux stressed the importance of delivering on such promises consistently. Another key ingredient is person-to-person follow-up.

Warren Williams, Division of Cancer Prevention and Control (DCPC), CDC

Mr. Williams talked about population-based cancer registries. He noted that data standards, collection, reporting procedures and use are fundamental to the organizations working to reduce cancer in the U.S. The information in cancer registries is used to implement cancer prevention and control, to identify when and where to enhance cancer screening efforts, to increase access to care and evaluate quality of care, and to develop effective research, among other uses.

The registries exist in every state, the District of Columbia, and territories. Cancer registry data typically are collected from aspects of the medical record, then reported to a central or regional registry, and finally to a national source. CDC?s National Program of Cancer Registries and NCI?s Surveillance, Epidemiology, and End Results (SEER) program work together on these efforts. Some direct reporting occurs from laboratories, physician offices and nursing homes. Uniform data reporting standards exist for registries, and most organizations involved in this effort collaborate under the auspices of NAACCR, an umbrella organization. NAACCR?s committees refine and publish data standards and procedures for reporting, with the format revised annually.

CDC has recognized the need to enhance data collection practices in response to a constantly evolving electronic environment and demands for ever-higher-quality data. Professionals interested in applying industry standards to improve cancer data reporting convened in October 1998 to examine the issues. With CDC?s help, a workgroup was formed to explore developing an alternative reporting protocol to the NAACCR flat file format. The group drafted an implementation guide for using the HL7 reporting approach as a springboard for further development.

In addition, ongoing work is focused on developing a comprehensive medical record vocabulary using the SNOMED vocabulary. Maps are used to relate this to ICD codes. More work on implementing the vocabulary is still needed, along with a transitional approach to data-gathering. DCPC has formalized new partnerships in the last four years for key industry standards groups to help promote the development and use of data standards. Another with the Regenstrief Institute will help promote, re-mine and implement the LOINC code set. Mr. Williams stressed that it is important for industry groups to appreciate and work with the reporting needs of cancer registries, which the cooperative agreements help accomplish. Achieving the NHII vision requires synergy among many organizations, to strengthen the reporting of clinical data.

As an example, he described a three-year, state registry-pathology lab project in two states to investigate the use of the HL7 paradigm and encourage standardization of electronic reporting for pathology data. One purpose is to give feedback to the College of American Pathologists on its standardized reporting protocols for surgical pathologists aimed at improving the completeness, timeliness and quality of cancer data. The registry community and industry standards groups need to be educated on the use and benefits of controlled vocabulary and coding schemes. Mr. Williams stressed the transitional nature of the projects, calling them ?beginning steps? toward the NHII. The metamorphosis of a large-scale population-based surveillance effort to enable real-time electronic reporting and take advantage of developing technologies and standards will be gradual Finally, he outlined several engineering-related principles and practices that are needed to support the NHII.

Barry Gordon Ph.D., California Department of Health Services

Dr. Gordon said that several people in cancer registries joined the standards venture early, when the possibility of electronic reporting was first recognized. California produced software to encourage physicians to use a standard system. The state implemented statewide reporting and data collection rapidly at least partly because of the provision of clear electronic standards and software.

Describing the conflicts between SEER and the College of American Pathologists on how to code treatment and staging, he noted that those who study trends and those with a clinical view will inevitably come into conflict over coding systems. Recognition of the potential gains in efficiency from an agreement helped the standards work, along with motivation from vendors and states. The second such effort in California involves standards for reporting potential cases by pathology labs.

Dr. Gordon said the cancer surveillance standards effort has been a success in his state and at the national level. Nationally, a neutral organization hosts the standards, all the U.S. and Canada are covered, and all key players are participating. Some data harmonization has been achieved, and there are common tools. And states are now measured on their ability to meet data standards and quality by CDC and NAACCR. He credited a combination of commitment, tools, and courage in achieving these ends. The distributable edits technology is one of the tools that strongly encourages standards.

Other areas need more work because of harmonization problems, lack of well-structured reporting, and limited implementation of HL7 standards. Dr. Gordon described several HL7-related projects in California, noting the lack of a clear business case motivating organizations to use a new messaging format for reporting and the lack of economic or regulatory motivation to switch to HL7. One project concerns integrating in-hospital messaging into cancer reporting; another is on electronic pathology reporting by path labs. Impediments include hesitation to report electronically, confusion about HIPAA regulations, and lack of agreement on which secure transfer protocols and authentication schemes to use. Finally, the Reporting Pathology Protocols project is promising because of the breadth of the participation in it. All of these efforts are focused on not just implementing but evaluating the new standards and techniques.

After mentioning other systems that registries could benefit from interfacing, including hospital discharges and state vital status records, Dr. Gordon outlined the steps toward more consistent and widespread structured standards and reporting:

  • Public health and clinical groups need to value connectivity, include it in their high-level agendas, and have the courage to measure compliance.
  • More ?glue? projects are needed to create interfaces and prove their worth.
  • Pilot projects need enough funding to include SDO representatives and others key to the interface.
  • New proposed coding and message structures must be implemented in real messages and environments before completion.
  • Standards need to be accompanied by portable edits and other implementation tools.

Discussion

Dr. Shortliffe asked Mr. Lamoureux how much of his eclectic approach to reporting in Baltimore is a function of not having the resources to create a unified electronic infrastructure. He also expressed interest in what conditions contributed to the unusual degree of support for standards in these two areas of public health, immunization and cancer, and also why cancer is so far ahead of immunization in its development. Mr. Lamoureux said that his department concluded when they began the registries that managed care organizations? electronic databases were not usable for this purpose. He added that the State of Maryland intends to establish reporting standards for its registry.

On the second question, Dr. Gordon said the cancer field has helped, with both NCI and the College of Surgeons having interests in standards. He noted that the costs of standards and automation are shared; they are not borne only by the state. And states can make it costly for providers not to report electronically, as California does. Dr. Lumpkin observed that CDC has heavily funded state-level immunization registries, but there are complexities that slow the development. Dr. Shortliffe also noted the differences in the types of providers that report to the two types of registries.

Dr. Zubeldia, noting the impossibility of finding a single solution for all users, asked the speakers? views on the feasibility of an interchange solution that would permit differing implementations but the exchange of the same core data elements. He further asked about the communication standards and vehicles for such an arrangement. Dr. Gordon agreed with the implicit analysis. He asserted that the exchange must be Internet-based, adding that contents standards are dealt with well, but there are still several wrapper standards, and that is where leadership is needed.

Mr. Lamoureux reiterated his ?bias toward city-based immunization registries.? With immunizations, he said, the standards and tools have been established; it is now a matter of the states supporting them. He added that Baltimore would ?surrender a lot of the customization? if there were a valid and robust system to replace its existing one.

Dr. Williams stressed the importance of the interface component of the systems. Asked why this is not happening, given that the implementation guides exist, he said it is happening in pilot efforts but it takes time and resources for the field to focus on this.

Noting the three overlapping dimensions of the NHII model, Dr. Lumpkin asked if thought has been given to giving parents access to immunization registry reports for use in their personal health care decisions. Mr. Lamoureux said the BIRP information is available on paper, but not electronically; the same is true for providers, although they will have electronic access over the Internet by the end of the year.

Dr. Lumpkin noted the absence of individual clinicians and information system vendors from the cancer registry initiatives, and noted the possibility of having cancer registry reports come automatically from clinical information systems, obviating double data-entry.

Noting the technical progress being made on standards, Dr. Fitzmaurice asked what can be done to get people to use the standards. Dr. Gordon said he wonders the same thing; he speculates that the resistance is cost-based. He added that regulatory standards will help move the process. After Committee members and panelists commented on the many facets and types of standardization, Dr. Lumpkin asked the panelists to later submit their views on whether NEDSS answers the issues of interconnectivity between the various registries.

Dr. Deering observed that the NHII concept is to maximize the linkages among related content areas and expand the concept of who uses the data. She asked panelists to send her their recommendations on the priority role for the federal government in moving toward these goals. She later made the same request of the other panels.

POPULATION HEALTH/NHII INTERFACE

Margaret Richards, Ph.D., Illinois Department of Public Health

Dr. Richards described three systems that exemplify the state?s population health dimension: its Birth-Related Data Project, the Illinois version of NEDSS (I-NEDSS), and Tracking Our Toddlers? Shots (TOTS, an immunization registry). Only the third is fully operational; the others are under development.

She showed an illustration of the technical infrastructure, with actual or potential connections to 94 local health jurisdictions and some 220 hospitals. The state is helping local health departments get broadband high-speed Internet connections, as needed. A scholarship program and other initiatives are helping set up a hospital health alert network. A critical access hospital telehealth and telemedicine program provides various services to hospitals; 20 are now online. They are moving toward having all public health partners have access through an Internet Web portal.

The Birth-Related Data Project aims to develop one master electronic record that can be filled out at the hospital and passed on to the seven state programs that use birth-related data. Programs will see only the data they are interested in.

Immunization data can be entered into the TOTS registry over a touchtone telephone (also used for requests for records) or electronically. The goal is 90 percent coverage; the state is now at 58 percent participation.

One of the objectives of I-NEDSS development is an electronic version of the morbidity card now used for communicable disease reporting. The state decided to develop its own version of NEDSS because of issues of timing and adaptability. It is committed to interoperability with NEDSS. Illinois? system will mirror its paper system. The overall purpose is to allow providers and labs to report cases or open cases for investigation simultaneously with The overall purpose is to allow providers and labs to report cases or open cases for investigation simultaneously with local and state health departments. A salmonella module will be piloted in ten local health departments in October. The most important features of I-NEDSS are its historical approach and its central repository, part of the effort to avoid duplication of cases. New cases can be opened only after a search determines that the patient is not already in the database. Dr. Richards described other features of the system, such as its ?my-case? folders.

Turning to interoperability, she said the ability to communicate across jurisdictions and send the information to CDC is critical. Local jurisdictions or disease reporting systems that choose not to participate in I-NEDSS could be the problem, but the hope is that everyone will choose to participate.

Oscar Morales, EPA

Mr. Morales said he was invited because of interest in EPA?s information exchange, data standards, and Web-based registry activities. He expressed the agency?s willingness to participate in work related to the NHII, and noted that it faces many of the same issues. Its primary information exchange partners are states and tribes, with whom it has been developing a network. CDC and ATSDR are the lead agencies in administering money in the 2002 budget to build an environmental health tracking network. EPA hopes to participate, using the Environmental Health Tracking Network to help lay the foundation. A MOU for such a partnership is now under development, and EPA and CDC are exploring the best ways to share information.

Mr. Morales said EPA can share with the network its experience in partnering with state agencies to build the National Environmental Information Exchange Network. The network will gradually replace the traditional approach to information exchange, in which states fed data directly into multiple EPA systems. States and EPA have been at work for three years on the foundation for the Network, which is two years old. The main partner was the State/EPA Information Management Work Group. Data standards, defined by the Environmental Data Standards Council, are provided as a basic element for the data exchange and integration activities. Eight standards have been promulgated and three others are under development. Mr. Morales described the process for creating standards, which takes about six months. They are mandatory for the owners of EPA?s databases and voluntary externally.

The barriers in this arena, he noted, are the same as in others: internal legacy data, compounded by complex and fragmented systems trying to use them. EPA has implemented a vigorous implementation program internally to get people to use the standards. To facilitate the use of the standards, the agency also developed an Environmental Data Registry on the Web, the authoritative source of references on environmental information. The EDR, an implementation of the ISO metadata registry standard 11179, was designed as a result of the efforts by Cdr. Bob Mays and NCVHS. In addition, the Environmental Information Management System, run by the R&D office, provides scientific documentation accessible with a standard Web browser. Finally, EPA uses grants to foster improvements in information exchange and integration, support targeted research, and strengthen the role of small businesses in federally funded R&D projects. EPA is now negotiating trading partner agreements with states.

Marc Overhage, M.D., Ph.D., Regenstrief Institute & Indiana University

Dr. Overhage said a microcosm of the NHII is being created in Indianapolis, called the Indiana Network for Patient Care (INPC), a secure, operational electronic medical record network. This presentation focuses on the implications for population health. He asserted that ?population health data is just individual health data for lots of individuals,? explaining that there are many overlaps between those uses of data and health care providers? uses. Moreover, population health data are needed not just for public health uses but by payors, accrediting bodies, policy makers, and others. He described the sharing of health care data for public health and other uses as ?health care?s last mile? and said many issues must be overcome in order to do so.

Most of the city?s medical-surgical hospitals participate in the INPC, plus other providers including a large managed care system, the homeless care network, public school clinics, and state and county public health departments and their clinics. There is a secure data network as well as a database of databases - i.e., INPC is the custodian of the participants? separate databases and helps develop agreements on how the data are to be used. Data from other sources are added to those repositories, using INPC as the primary data entry point. Consistent data formats and coding systems are used. A lot of clinical data are automatically collected as a byproduct of care. The data shared by all institutions include ED and outpatient visits, hospital discharges, and inpatient lab results. Many more elements are being added, making it a rich data set.

Building on data standards, notably HL7 and DICOMM, has made this possible. They initially tried to have institutions? laboratories map their data into the LOINC standard, but that effort failed for multiple reasons. The current mapping approach, which analyzes and maps HL7 messages and then does a largely manual error-check, has been very successful and is increasingly efficient.

In clinical care, the data are used to produce reports such as emergency care abstracts for physicians. In the population health domain, this serves as an immunization registry, electronically reports reportable conditions, and does ?some crude surveillance.? Using data from a shigella outbreak in 2000, he showed how the IMPC surveillance system was able to identify the outbreak eight and a half days sooner than the health department. The data are also used for health services research.

In conclusion, he observed that 1) INPC lends credence to the NHII idea that ?data can be repurposed and reused,? 2) standards are critical, and 3) a little work can go a long way with such systems. Authentication continues to be a big issue for his project, which relies on individual institutions to authenticate providers to access the data. Coverage also continues to be a major issue, notably for ambulatory visits. There also have been challenges in getting the public health department to use the INPC as a primary data source.

Discussion

Asked about interoperability across state lines, Dr. Richards said there have been discussions about regionalizing the Illinois effort. She noted that the majority of states are, like Illinois, developing their own NEDSS systems. Dr. Lumpkin added that the time pressure under current grants to get a system up and running has created a disincentive for taking the time needed to negotiate with other states on building a standardized system.

Dr. Shortliffe asked Dr. Overhage about the transferability of what has been created in Indianapolis, given the unusually high degree of cooperation and buy-in by providers there. Dr. Overhage credited both the leadership of visionary health care providers who look to the public good and the presence of a single medical school rather than competing ones as major factors in the success, together with support from AHRQ which ?lowered the entry barrier.? Finally, he noted that this project ?moves slowly? in order to build trust; for example, they avoided collecting STD data for five years until people became comfortable with the system.

Asked about privacy protections, he said they have been very careful about privacy and use a private data network, multiple agreements, patient consent forms, IRBs, and other measures. Dr. McDonald said the clinical part is ?very clean? in terms of HIPAA requirements.

Dr. Lumpkin asked Mr. Morales, as EPA develops its information exchange network, to be attentive to the recommendations in the NHII report about affording multiple uses of data for multiple users.

In response to questions from Mr. Blair, Dr. Overhage said his project works closely with state and local health departments on their NEDSS development. He commented on XML and SNOMED, and said they have not yet ?directly tackled? personal health uses of the system.

In addition to asking the other panelists to send her their priorities for the federal role, Dr. Deering asked Mr. Morales to supply cost information on EPA?s system.

PERSONAL HEALTH: PERSONAL HEALTH RECORDS

Philip Marshall, M.D., MPH, WellMed, Inc.

Dr. Marshall defined a personal health record (PHR) as a system that enables patients to view or manage their own health information. In a recent study, a majority of patients showed interest in having one. Many organizations are starting to implement them and provide applications to consumers - e.g., drugstores, PBMs, health plans, doctors, and employers. Many vendors are supplying these organizations with the tools, while others are building them themselves.

Several standards are needed: notably, minimum data sets to facilitate data sharing, interoperability between electronic systems, patient control and privacy standards, and consumer terminologies to facilitate ease of use. For example, WellNet has created the Consumer Health Terminology (CHT) thesaurus as its infrastructure. Dr. Marshall noted that in this area, the marketplace has come a long way in a few years. The systems that have sustained involvement in this area have paid attention to privacy and security, coding systems, and interoperability standards.

The goal of PHRs is to enable patients to aggregate, integrate and share data across multiple providers and multiple systems. This can result in greater cost management and health self-management. What is needed for this to happen is a way to encourage the systems and organizations that hold patient health data to share basic elements electronically and securely with other systems chosen by the patient. Achieving this goal requires defining a minimum data set and a simple way for these data to be shared between systems, with privacy safeguards.

Minimum data sets provide the foundation of portability. Dr. Marshall said that about 50 data elements could comprise this set - e.g., insurance information, contact information, current conditions, current medications, test results, allergies, immunization history, surgical history, health risks, and family history. The question is how patients can get access to the data. This involves questions such as whether the systems holding data have the wherewithal to do the user interface, whether patients should have access to all systems, the mechanisms for access, and what to download to (Smart cards? Web-based applications?).

In any case, system interoperability is needed in order to ?go in and grab the data,? and that leads to the need to standardize electronic interfaces. He noted that a simple XML interface would be easy for almost any system to implement. The next challenge is authentication, at both system and individual levels, and he posed some questions in that area for consideration. Finally, patient control, privacy and authorization standards are needed. Here, Dr. Marshall stressed the growing importance of standards whereby patients can give authorization and permission for who has access to what data. He noted that the sensitivity of data is highly contextual; i.e., what is sensitive to one person may not be sensitive to another. There also must be audit trails.

He concluded with several recommendations, including these:

  1. Establish a minimum data set and encourage its use.
  2. In lieu of a national identifier, establish minimum MPI requirements for matching data to an individual.
  3. Support accreditation of PHR systems (URAC is one possibility).

Vik Kheterpal, M.D., Medicalogic/GE

Dr. Kheterpal?s company supplies clinical information systems to provider organizations. He noted that health care lags behind most other industries in terms of digitization. He compared it to banking and suggested ?the Quicken of health care? as a way of thinking about success. He asserted that populating one?s own record is an ?extremely onerous task.? A translation of technical content must be offered. He noted the dangers of partial information, e.g., a partial medication list.

GE Medical Systems? experience is based on the use of a PHR from a provider-populated model by more than 1,000 patients over two years. Dr. Kheterpal showed and described a screen shot of a health summary. Consumers are looking for access to their health records in a variety of modes and for a variety of purposes. Different purposes drive different usability and standardization issues. Different uses are the uni-directional dissemination of information, patient alerts and reminders, and two-way communication with the provider.

He shared several lessons learned. First, stratification of the patient population and personalization are key factors in success and usability. Second, the biggest struggle stems from the fact that ?every clinician in a health care system believes they control the record.? Third, the enrollment process requires informed consent from patients who want to participate after having been advised of the risks and that this does not replace clinical interactions. Fourth, there are provider work flow issues to be worked out. Finally, the ROI model is ?dubious at best? in the industry today; the upshot is that ?consumers are way ahead of providers? in this area.

Dr. Kheterpal reiterated the importance of stratifying the target population of PHRs with respect to a host of demographic, medical, and other variables. He also noted that a major challenge is taking patients? personal health information and cross-referencing it to reference information.

Major challenges for creators are the lack of a common lexicon, the need for a mature ROI model for adopting entities, and the dangers of ?scope creep and immature project phasing.? To accelerate such initiatives, the field could use help with the following:

  • the development of a lexicon standard, possibly a minimum data set and standards as recommended by Dr. Marshall
  • ratification of a recommendation on security and enrollment processes
  • research funding for implementation and execution approaches
  • a regulatory mandate for a basic universal registry

Peter Waegemann, Medical Records Institute

Mr. Waegemann identified these types of PHRs: offline personal records using purchased software; Web-based commercial/organizational PHRs (with various payment schemes and providers); functional or purpose-based records; provider-based personal records; and partial PHRs.

13 million people in the U.S. have put their health information on a Web site, but only 10 million are in active use. Some PHR companies went under in the spate of dot-com failures. The main increase in PHR use is among the elites of developing countries who travel abroad for treatment of serious health conditions. Three international companies are working on an infrastructure based on a commercial PHR.

ASTM has worked on a standard that was piloted in May and is becoming a national standard, requiring several things of sponsors such as authentication and privacy guidelines. Mr. Waegemann said companies have not been requiring confidentiality training for database operators, as hospitals would. He has found that physicians do not trust the information patients have in their PHRs, and it will take time for them to get used to the idea of patients having complete health records. He added that in his experience, patients are not interested in a merging of the provider-based record and the PHR. This creates the problem of developing two different records.

In conclusion, Mr. Waegemann said that the question for the future is how the initiatives in different sectors will merge into one system for the benefit of both patients and providers.

James Cimino, M.D., Columbia University

Dr. Cimino said he defines health information to include both patient-specific information and more generic health information because many messages include elements of both. He described two Columbia University projects, the Patient Clinical Information System (PatCIS) and the Myocardial Infarction Health Education Aimed at Rapid Therapy (MI-HEART).

The PatCIS project, funded by the National Library of Medicine, started six years ago. The idea was to give patients access to their records and ?see what happens? - i.e., find out what was going on in their heads as they used it. The architecture was a Web-based front end to the New York Presbyterian?s clinical data repository for patients, allowing patients to view their data just as their clinicians would, with the addition of educational resources, guidelines, and ?info buttons.? Users found the interface functional. The components were data entry (by patients, with alerts to their clinicians), data review, advice, and education. Dr. Cimino described some of patients? semantic issues, such as misinterpreting ?negative? X-ray findings and the word ?limits? in ?within normal limits.? The info button application takes the text of reports, explains it, and gives links to other resources. The advice component brings together patient data and medical knowledge to give patients customized messages. The education component has links to outside resources.

The evaluation of the project used online questionnaires, usage reviews, and telephone and in-person interviews with patients and physicians. Thirteen patients were followed for 36 months. The investigators found that technical factors were not a problem for patients. The overwhelming majority of usage (three-quarters) was for data review, and three-quarters of that was for laboratory results. Patients generally discussed the results with their clinicians to improve their understanding of their condition, thus improving doctor-patient communication. There was little interest in the educational materials or guidelines.

The MI-HEART project was sponsored by NLM and the NHLBI as part of the National Heart Attack Alert Program. The objective of this program was to reduce the time from onset of MI symptoms to calling 911, using the Web as an educational medium. It also explored the use of clinical data to customize messages to patients. This was a controlled trial involving three patient groups: a control group that used paper-based materials, a non-tailored group that got standard materials in a Web-based delivery mechanism, and a tailored group that received customized Web-based messages. The researchers looked at how well the patients understood their personal risk of heart attack, their knowledge of what to do in case of a heart attack, and self-efficacy. Patients could customize the font size. Dr. Cimino showed some examples of customized messages using the patient?s actual risk factors, medications, questionnaire response, and so on.

This study had more than 120 subjects. Data collection continues, although most of the study has ended. It found that both Web-based groups improved compared to the paper-based group, suggesting that Web-based education is a better medium. In addition, the tailored group showed sustained improvement compared to the other groups.

From both projects, the researchers learned that ?patients can handle the technology?; they want their results; they want to collaborate with their physicians; and the EMR and the Web together can be an effective educational medium. They ran into some standards issues related to browsers, and Dr. Cimino urged the Committee to support standardization of browsers. In addition, he noted the need for patient identifiers and the issues around terminology. Access to high quality resources could use standardization as well, he said, along with the structure and access to guidelines. He noted the different levels of granularity used in the terminologies for patient data and for guidelines.

Turning to next steps, he said patients must be involved in setting the agenda, to determine what information they want and what they want to do with it. There are also infrastructure issues to be addressed, including standards, patient identification, confidentiality, and access. Education is another priority, so that more patients can both use the technology and understand the information. Finally, there is a need for a standardized format and content for data, educational materials, and health information resources.

Discussion

Asked his criteria for good browsers, Dr. Cimino said they need a standard language and a standard look and feel. A first step is simply to ?say what the standard is? - it doesn?t really matter what it is; people want rules to follow. As to who should advocate the standard, he suggested that NCVHS should, along with other standards groups.

Dr. Cohn raised the subject of terminology, noting that some speakers had called for an ?interface terminology? or ?patient-centered terminology.? He asked if this is an area for standardization. Dr. Marshall said there is an opportunity for standardization here. He described past efforts to create the Consumer Health Terminology thesaurus, using SNOMED. While this served as a ?differentiator? for his company, it did not stimulate a lucrative market for a vocabulary. He said his company would participate actively in a standardization effort, noting that ?there is a lot of latitude for misunderstanding by the patient? of the various PHRs options now being marketed. He added that only the terms that patients will run into need translating, e.g., ICD codes and lab results, and not ?all the granular concepts.?

Dr. Cimino observed that patients and clinicians talk about the same concepts, but using different terms. The first step is to define standards for representational schemes in order to exchange these terminologies. Other steps are standards for how the terms are represented and standards for the content. Then it will not be difficult to map patients? terms to clinicians? terms.

Dr. Kheterpal agreed with both speakers, noting that the key is integrity in the clinical information being generated by the provider. Then this can be tailored variously to patients? different educational levels, learning styles and language preferences.

Dr. Lumpkin asked if the panelists thought there was a role for government in this area, and thus that NCVHS should get involved in its advisory capacity. Dr. Kheterpal commented that as rare as it is for a large corporation to invite regulation, in this case, a standard would be useful. He suggested that the government not establish it but enforce it once one starts to emerge. He noted the leverage inherent in Medicare payment requirements. Dr. Marshall said that incentives to get organizations who have minimum data to ?ante it up? to organizations working on patients? behalf could be very important, as well.

Asked about the feasibility of the voluntary establishment of criteria or a methodology for patient identifiers, Dr. Marshall said his organization uses a system of matches to identify individuals, and this could be workable in the absence of a single national identifier. Dr. Kheterpal said the pragmatic approach within health care organizations is to use a cross-referencing master patient index. What is lacking is incentives across health systems in the nation?s fragmented health care system. Dr. Cimino predicted that when patients start to drive the process and they see it as in their interests to do so, there will be voluntary efforts by patients to participate in having their own unique identifiers.

Dr. Zubeldia challenged the implicit worldview of ?islands with standard bridges that connect them.? He noted some of the pitfalls of patients having access to their clinical information and being able to pass it to other providers. He proposed, instead of data that migrate from the PHR to the provider record to the public health record, the possibility of a connected infrastructure with patient ability to view clinical data but without the ability to manipulate or move the data. He asserted that this is closer to the vision described in the NHII report. Dr. Kheterpal responded that this gets at questions of how to execute the PHR. He noted that legally, the patient owns his/her record. Providers have both trust and competition issues with allowing the portability of information, and no real commercial reason for doing so. The constraints stem from policy and trust issues, not technological barriers.

Dr. McDonald cautioned that XML is not simple when lots of people want to communicate; the same standardization problems arise. He asked why there is so much effort around this when his impression is that patients aren?t asking for it. Dr. Marshall responded that there is a good deal of evidence that patients are getting much greater access to health record information, not through ?personal health record portal play? but through ?other plays.? This suggests that patients do want it.

Mr. Waegemann likened current conditions to the invention of the printing press, when lay people could read the Bible for the first time: ?We are starting a process of patient empowerment? (which, he added, leads to better patient compliance and also to lower health care costs). Even though the commercial organizations are shrinking, patients undoubtedly want to see their lab results, medications, and so on. Returning to the earlier topic of patient identifiers, he asserted that for personal medical records, at least, these identifiers are not necessary.

Dr. Kheterpal said the issue is not that patients don?t want these online capabilities but that they are difficult to achieve. One of the main reasons why patients have not used online health records and health information more fully has to do with security concerns; the other is that they lack the technology.

INTEGRATION OF E-MAIL WITH CLINICAL AND PUBLIC HEALTH INFORMATION SYSTEMS

Ted Eytan, M.D., Group Health Cooperative of Puget Sound

Dr. Eytan said this year, Group Health primary care teams exchanged 17,500 secure messages with patients, over and above services such as pharmacy online, shared decision-making tools, and discussion groups. Group Health is implementing a comprehensive clinical information system with accessible PHRs, to be ready in 2003. Dr. Eytan noted that Group Health encompasses many stakeholders in the NHII taxonomy, including provider, consumer group, researcher, and health plan.

He showed a screenshot of the MyGroupHealth patient portal and said his remarks would focus on secure messaging. These services were developed by a team of information services professionals and clinicians (without benefit of standards, he noted) and piloted at the Redmond Medical Center from November 2000 to May 2001. Services are now available to all adult patients at Group Health medical centers, giving them access to all primary care health teams.

The team developed a lexicon for the services. Dr. Eytan said they use the term ?secure messaging? rather than ?e-mail? because of the security protections in the former. Characteristics include security, privacy, confidentiality, participation by the entire health care team, service standards for turnaround time, integration with the medical record, and population-level monitoring and improvement.

One of the standards issues was authentication, for which a community standard does not exist. Group Health requires patients to present physically to them or a notary public to authorize access to their information. Once authenticated, patients have choices about access to clinical data. Group Health has different processes for online activity for health care teams, nurses, patient care representatives, and business office managers. It has created reference guides on secure messaging, describing responsibilities and constraints.

The legal medical record is the paper chart, and clinicians decide which secure messages are placed in it. When Group Health has transitioned to its EMR, it plans to include secure messages as encounters, on a par with office visits and telephone encounters but capable of being filtered by staff users.

Group Health ?works vigorously to apply standards in the informatics realm,? Dr. Eytan said. It has defined its clinical information system as an automated tool for both patients and staff. Secure messaging is tied into the process of care and is a conduit to other services and information. It may also be the conduit for bringing structured data into the clinical record. It is Washington State law that patients can examine and amend their records. Group Health believes clinicians should be involved in patients putting data into the clinical data repository.

Dr. Eytan said that health systems wanting to transform health care can assist by critically assessing standards under development and applying them as demonstrations. Group Health selects technology that uses systems compliant with the latest standards. A current barrier to the development of standards in the online patient arena is the lack of adoption of online services across the majority of health care providers and plans. Group Health often tries to educate people about the value of the services. DHHS can help through its leadership - for example, by educating the public about the value of the capabilities he has described and promoting greater use of these services in the field. DHHS can also help by bringing all the stakeholders together to develop solutions. He concluded by showing a short video showing patients and doctors using the service.

Ron Nath, M.D., Apelon, Inc.

Dr. Nath said he would talk about consumer health vocabularies and their applications to e-mail. He described how his company created a consumer health vocabulary, saying it did so to meet a need, differentiate themselves from other terminology vendors, and add value to the standard vocabularies it provides to customers.

In developing the vocabularies, Apelon made decisions about multiple factors including the appropriate level of sophistication, the scope of terms, and what mappings to do. For content sources they used the UMLS, consumer-oriented medical Web sites, CPT codes, information on NLM Web site search, and reverse mappings. They decided to go with a ?general newsprint article? level of sophistication of terms, appropriate for an eighth-grade reading level. They included alternative health terms in the vocabulary because of consumer interest in it, as well as terms on anatomic locations, diseases, categories of drugs, procedures, and medical concepts. The aim was to have 10-20,000 terms, and they ended up with 15,000. (He compared this to WellMed?s 35,000 terms.)

He noted consumers? increasing sophistication in using clinical medical terms; these and lay terms form a continuum, with a large area of professional-lay overlap in the middle. Thus his company retained many clinical terms. For mapping they used SNOMED-RT, ICD and CPT, and they may map to the UMLS codings. They found that not only are consumer terms imprecise, but many do not map to professional terms, notably many in the alternative health domain (e.g., magnetism and crystal healing). They used in-house software tools and developed additional ones to mine Web sites. It took more than a year for nurse terminologists to map the consumer terms to SNOMED. Ultimately, only 10,000 of their 15,000 consumer health terms could be algorithmically mapped; 4,000 were manually mapped and 1,000 could not be mapped.

Apelon envisions people using the vocabulary to visit web portals and learn about their conditions. It can also be an interface terminology between clinicians and consumers. Apelon did not initially consider the applicability to e-mail and messaging, but Dr. Nath said he could see its usefulness, especially for automatically generated messages. He agreed with Dr. Deering that there are potential uses in checklists provided for structured messages between patients and clinicians. Regarding future directions, he joined other panelists in stressing the importance of standards, and said he would like to see a consumer vocabulary maintained by a public body such as the NLM, and perhaps its inclusion in the UMLS.

Kathryn Bingman, Cerner Corporation

Cerner provides information management systems designed to improve the delivery of health care. Cerner currently has 24 projects that connect consumers with their health care providers. Ms. Bingman described online projects in Winona, MN and Eastern Maine that focus on providing health care consumers with both access to their personal health information and an online connection with their health care providers. In the former, the large majority of physicians, the lab, pharmacies and the local hospital are connected to consumers. In Eastern Maine, personal health records have been deployed as a care coordination solution to support the citizens in making better health care decisions and in accessing necessary care and information. Eastern Maine is also partnering with statewide employers to further extend the depth of the program and enhance its rural health initiatives. Her comments on the challenges of consumer-driven health information management, and in particular on standards, are based in part on their experiences with these projects.

She distinguished the electronic medical record from a ?true personal health record? in that information in the latter is owned and maintained by the consumer. The ideal PHR should be populated by both consumer-entered information and clinical data collected across the continuum of care. It should be accessible by health care professionals at the appropriate time, be controlled by the consumer and be portable.

In order for personal health records to succeed, physicians and hospitals must save data electronically, within an electronic medical record, instead of on paper. A standard does not currently exist for a legally defined electronic medical record across all states, and this is a necessary component of a standardized PHR. Ms. Bingman noted that fulfillment of the HIPAA requirement for providers to give patients access to a designated record set in a standardized electronic format, coupled with consumer-entered information, would allow better management of personal health. Standardized PHRs would also speed submission of data to public health organizations. However, standards will not suffice because they do not ensure that the information is shared across organizations. The benefits for public health can only be attained when there is enough critical mass in a given community, or across the nation, to get meaningful information. This is particularly true when applying surveillance technologies that will capture the sometimes-subtle indicators that signal an biological attack or an outbreak.

Currently there are many different agencies, health systems, and providers that recommend self-care standards by disease state. National self-care guidelines would produce more predictable outcomes across the entire country if integrated into PHRs. Ms. Bingman noted that this would extend patient safety to the home. Currently there are a multitude of vocabularies used to codify clinical information. Cerner has incorporated the UMLS and uses mapping tools to promote interoperability. While this helps in the clinical arena there is still work that needs to be done to make the vocabularies understandable for the average consumer. All information accessed by consumers must be written at no more than an eighth grade reading level in order for it to be used by the general public. The multi-lingual needs of the United States population will have to be addressed as well.

Other issues of importance, she said, are security and identifiers for providers and patients. In Winona they initially required a personal digital certificate in addition to in-person presentation of an ID to gain initial entry to the system; however, consumers did not understand the additional level of security and found it cumbersome, so the digital certificate was removed. (She later explained that consumers had privacy concerns, but not security concerns, and they felt the digital certificate was not worth the extra trouble.) All communications are encrypted. Regarding identifiers, Ms. Bingman predicted that both physicians and consumers would have much greater confidence in the validity of health information if health identifiers were in place in combination with a national digital signature database. Finally, she recommended messaging standards for communicating clinical information. Among other things, messages containing person-specific health information should have a digital signature encrypting the message as only the intended person can read digitally signed messages.

Turning to the barriers of standards development, she identified proprietary interests on the part of technology suppliers and the absence of a business case for standardization. Incentives from the public and private sector should fund an industry initiative to develop standards. A neutral development environment should ensure common interaction. Cerner supports the HIPAA-mandated building blocks, which should be used as a starting point to standardize additional code sets and vocabularies. Solving the broader challenges around the cross-organizational sharing of clinical information will yield the benefit of having personal health records be truly all-inclusive and connected.

Discussion

At Dr. Lumpkin?s request, Ms. Bingman said she would supply the Committee with a write-up on the Winona project, including current status and next steps.

Asked about patient usage, Dr. Eytan said the in-person sign-up is a barrier that Group Health is looking at. About ten percent of eligible enrolled members use the PHR. Some 20 percent of these use the system each month, parallelling the use of health care. Ms. Bingman commented that they had the same barrier, but offered people many venues for signing up for the PHR. About ten percent of the community signed up, of whom about one-third use the service consistently. She noted that ?people only need it when they need it, not every day.?

Dr. Deering asked the panelists to share their findings on the return on investment for these programs. Dr. Eytan said his company?s position is that ?whatever we can do to keep patients healthy is worth the investment,? and that it is premature to do an ROI analysis when they haven?t yet reached critical mass. It is concentrating now on patient and provider satisfactions. Ms. Bingman said that her company has found that those who are willing to invest in the early stages of

?visionary projects? such as this ?are the ones that control the health economy.? She noted that some of the applications of electronic access, such as self-scheduling and ordering prescription refills, may contribute to clinical care and ROI but are extraneous to ?the true PHR.?

Dr. Deering made her customary request that the panelists send her their priorities for the federal role in this area. To this Mr. Blair added a request for information both on what standards they use and in what areas they failed to get a satisfactory business case because of lack of a national standard, suggesting what new standards could work as a catalyst.

Dr. McDonald said a trusted, encrypted identifier system of some sort would be of great help so that each health care system or research project does not have to identify participants when they log in. He added that the layers of passwords currently being used are burdensome.

Dr. Lumpkin thanked the panelists and said the day?s discussions had helped point the way forward for the NHII. He then adjourned the meeting.


I hereby certify that, to the best of my knowledge, the foregoing summary of minutes is accurate and complete.

/s/ John R. Lumpkin 10/15/02

_____________________________________________________________________

Chair Date