Renaissance Hotel
Washington, D.C.
The National Committee on Vital and Health Statistics was convened on June 27-28, 2001 at the Renaissance Hotel in Washington, D.C. The meeting was open to the public. Present:
Mr. Scanlon reported that the National Academy of Sciences is starting work on a study for HHS of the adequacy of its race/ethnicity data. He noted that the Committee may want to interact formally with the Academy on this project. For the next stage of its survey integration initiative, the Department is focusing on long-term care data. In the dissemination area, HHS is developing a user-friendly one-stop portal to give users easy access to all of the public-access data in its possession. Dr. Braithwaite reported on several projects underway on standards, including development of NPRMs for claims attachment transaction standards and for updates to already-released standards.
Dr. Robinsue Frohboese of the Office of Civil Rights reported on the early stages of implementing the HHS privacy regulations. The Secretary opened a 30-day comment period after the initial release of the rules; some 11,000 comments were received and reviewed; and the Secretary and the President decided to go forward with the privacy rules on April 14. In addition to highlighting some workability issues that the Secretary will be reviewing, the comments revealed some misconceptions that the Department is acting quickly to address. It is doing so through a comprehensive set of guidance that will be published, mailed, and posted on the Web [www.hhs.gov\ocr\HIPAA] in the near future. During the two years that covered entities have to come into compliance with the rule, the Office of Civil Rights will be working on building its internal capacity and engaging in public education, outreach, and technical assistance. OCR is also developing the enforcement regulation and penalties. The Committee and OCR staff agreed to continue to work together on these efforts.
Dr. Clancy talked about racial and ethnic disparities in health in the U.S. as a quality of care issue. Getting at the disparities associated with socio-economic position, race and ethnicity, she said, is a critical quality improvement opportunity, but one that cannot be realized without the right data and strong leadership. Currently, performance reporting does not include SES or race/ethnicity. Because there is no clear description of what data currently exist in health plans or of how relevant data could/should be collected, AHRQ took part in an exploratory meeting with MCO leaders, employers, and others, with co-sponsorship from OMH and the Commonwealth Fund. The participants expressed both interest in better data and concerns about how to obtain them in view of the barriers. AHRQ, which last year started a new contract for an integrated delivery system (IDS) research network, has three task orders in progress in this area.
Dr. Carter-Pokras reported on efforts by HHS to address managed care organizations’ legality concerns around collecting race/ethnicity data, to assure providers and plans that it is not illegal at the federal level to collect this information if it is used to improve healthcare quality. OMH has funded the National Health Law Project to review state laws and regulations governing the collection and reporting of data by health plans and insurers, and a parallel project funded by the Commonwealth Fund to review federal laws and regulations for Medicare and SCHIP. She and Violet Woo reported on the findings of these and other projects.
Dr. Nerenz reported on a project on data for health plan report cards on quality of care for minority patients. It is funded by HRSA and the Commonwealth Fund. Generally, comparative health care quality report cards do not contain data on race/ethnicity; this leaves both consumers for whom this is an issue and organizations wishing to close the gap with no valid and focused source of information. The planning phase of the project has been completed, in which a broad-based group of stakeholders recommended a set of health care concepts, measures and clinical conditions along with policies and procedures to follow in collecting and reporting data for a report card. In phase 2, the project is working with a diverse group of nine plans to assess the feasibility of various methods of obtaining data on race/ethnicity and using them to generate stratified quality reports. Results are expected by the end of 2001. It remains to be seen whether the findings from this group with “unique capabilities” can be applied more broadly to health plans.
Asked to comment on the difficulty of getting data on socio-economic position, Dr. Clancy said several studies and surveys have shown that people have greater reluctance about revealing this information than they do about many other seemingly more personal matters. Education is somewhat easier to obtain. Dr. Carter-Pokras said the Census found the same thing.
Dr. Iezzoni explained how the functional status project originated in the Committee’s 1996 consideration about core data elements recommendations, which included a place-holder for functional status. In 1999, the Population Subcommittee decided to study this question and quickly concluded that it was not feasible, given resource limitations, to do anything meaningful on the complex question of measurement. This led to the Subcommittee’s subsequent decision to focus on the recently-revised International Classification of Functioning, Disability and Health (ICF) as a candidate for a code set for functional status information in clinical and possibly administrative records. The Subcommittee’s recommendations in its report on this subject have four key points:
In discussing the report, Committee members suggested that the Committee not recommend reporting functional status information on claims or encounters, even as a statement of principle, until the recommended research and testing are completed on the feasibility of using ICF in administrative records. Apart from that, they expressed support for the report and its recommendations. Following discussion and a variety of suggested solutions to the issue raised, the Committee agreed to revise recommendation 2 in accordance with this suggestion. A motion to approve the report as revised was passed unanimously.
Dr. Madans reviewed the decisions surrounding OMB Directive 15, the original directive on the collection of race/ethnicity information, and the decision to change the standard for the 2000 census. She then discussed the statistical consequences and complexities of that decision. She stressed that what is underway is a 10-20 year transition. For vital statistics, the major concern is that at least for 2000-2002, the numerators and denominators are incompatible: no multi-racial people will appear to be dying or having children, so a bridging file of some sort will be needed for this period. The concern for the future relates to mortality data, because of the likelihood that multiple race will not be reported on death certificate, though it is an option. She discussed various implementation issues, noting that NCHS plans to do research into best practices for data collection for various modes, e.g., household-based surveys, phone surveys, vital records.
Dr. Pinal reviewed the key changes that greatly affect data products, such as allowing people to select more than one race category, the creation of a new race category, Pacific Islander, and the multitude of options and combinations for multiracial categories. He showed Census findings of self-report on race/ethnicity in various categories in different localities. In general, no racial category had more than one percent that also reported themselves as multiple race, and 90 percent of those who report as multiple race report only two categories. There is considerable variation among states in race/ethnicity reporting and counts. Dr. Pinal showed some tables and graphs suggesting how the Bureau is interpreting and portraying the 2000 Census information.
Mr. Harrison stated that “the federal statistical system is in great jeopardy of becoming virtually very difficult and very meaningless for conducting the kind of analyses by race that you’ve conducted in the past.” He cited some of the statistics reported by Dr. Pinal in support of his statement, noting the wide ranges of possible numbers and pointing out that “these are the denominators you will be using through much of the health surveillance system.” The numbers are especially meaningless for American Indians, and in some areas for Asians and Blacks. He argued that what is needed is adequate bridge statistics to tell what the population distribution and characteristics would have looked like absent the revision to the new race/ethnicity standards. The Census Bureau will be conducting a bridge study, but he asserted they are also needed at the county level given the wide regional variations and the need for accurate local data. The bridges need to be to both the past and the future, to provide meaningful trend data and interpret future events. He noted the flaws in drawing conclusions about racial disparities in health because of the flaws in the race/ethnicity information. The underlying question is whether the statistic has any real-world meaning. Mr. Harrison called on the OMB, the Census Bureau and the statistical system to take seriously the issue of providing adequate bridge statistics.
The Committee reviewed two draft letters to the Secretary prepared by the Subcommittee on Standards and Security: a transmittal letter concerning DSMO-recommended changes to standards, and a letter presenting a series of NCVHS recommendations based on recent Subcommittee hearings on implementation of HIPAA standards and industry readiness. The Committee discussed both letters and recommended a few changes. Motions were passed on day two approving both letters, as revised.
This Task Force is an HHS project by four collaborating agencies: AHRQ, the Center for Medicare and Medicaid Services, FDA, and CDC. Its goal is to help provide “opportunities to learn from errors.” While other initiatives are working on other forms of reporting for accountability, which the public is demanding, this one is for voluntary reporting to promote learning.
The PSTF is looking at existing reporting systems, which are incompatible and incomplete, and considering what the participating agencies can do to make the interface for reporting more user-friendly and reporter-friendly so the information can be sent to a single place and then sent to the appropriate federal agency. It is also looking at how to integrate the data to create learning opportunities and result in changed practices. Attention would be given to making the data available to researchers and policy makers, something that is not happening with existing systems. A reporting summit was held at the end of April at which stakeholders were told of the task force’s plans and invited to give input about what would be useful to them. The summit showed the desire for leadership and openness to the efforts of DHHS.
In the discussion period, Committee members urged that attention be given to adverse events as well as errors, and that the task force think broadly about engaging all HHS agencies and encompassing collaboration around clinical data management, reporting, and electronic medical records interfaces. The importance of going beyond effectiveness measures also was stressed.
The NQF was conceived in 1998 by the Presidential Advisory Commission on Consumer Protection and Quality in the Health Care Industry. It has been operational since early 2000. Its primary purpose is to increase the provision of superior or high-quality care. Its strategy is to standardize health care quality measures and reporting, to encourage public disclosure of this information, and to encourage that health care purchasing decisions be made on the basis of quality of care data. The Forum, which is “an organization of organizations,” has broad stakeholder representation from both public and private entities on its board of directors. Dr. Kizer described its composition, governance, and activities. He stressed that the Forum does not intend to create new measures; it sees its niche as reviewing those that exist, evaluating their evidentiary base and other factors, and assessing whether they meet national priorities for improving health care. The Forum is also involved in helping establish consensus in the latter area. Much emphasis is being given to the power of broad-based buy-in and market forces to bring about change with respect to quality of care.
Dr. Corrigan said the IOM recently released a report envisioning the National Health Care Quality Report, which was recommended by the aforementioned President’s Advisory Commission. There are ten recommendations, falling into four areas: the framework and categories; measure criteria and guidelines; data sources, collection and analysis; and audience-centered reporting criteria. She described the elements of the conceptual framework and said the committee adopted criteria for the selection of measurements, both individually and overall. It recommends that a broad-based independent committee or advisory group help assess and guide improvements in the report over time. The report will have several versions, tailored to different audiences. IOM plans to start an information technology and quality project by the end of the year, looking at what constitutes the NHII and in particular the resource requirements for it.
Dr. Reilly stressed that the quality report is a Department-wide activity. He began by discussing the measures that populate the framework outlined by Dr. Corrigan. HHS is now developing and identifying measures to populate the matrix (the framework), with input from the private and public sectors about possible measures. It is also setting priorities for what conditions to focus on, and identifying data sources. Over time, it will also identify and determine how to fill data gaps. For the reports, there will be multiple products to meet multiple needs, including a written report to Congress, web-based reporting, press materials, and more, with queries to stakeholders as to what would be useful.
Dr. Lumpkin and the Committee affirmed that they would like to explore further ways that NCVHS can contribute to this effort; it will develop a specific proposal on this.
NEDSS is an interconnected electronic information network of systems that is not limited to data about diseases but can support other conditions and syndromes. Surveillance is just one of many information system functions needed for public health; others include notification, communications, and knowledge management. In the current situation, individual legacy systems that were developed as a result of categorical funding by Congress go from local to state to CDC. These systems do not talk to each other and are a great burden to CDC’s local and state partners. To remedy this situation, the long-term objectives of NEDSS are to have ongoing, automatic capture and analysis of data, to use data that are already electronic, to have systems designed on relevant data sources, not diseases, and to integrate public health and health care systems. NEDSS includes a standards-based system architecture. All fifty states have received funding to assess and plan for NEDSS-compatible systems. CDC is working on many fronts with an array of public and private partners at federal, state and local levels to achieve these goals.
The Public Health Conceptual Data Model provides the framework for what categories of data are needed for public health, especially surveillance. Dr. Koo showed and discussed a schematic of the model. The purpose of the model is to reduce development efforts for computerized systems and enhance public health’s data exchange capabilities. It has been helpful in representing public health data needs to the Standards Development Organizations (SDOs); the model is being harmonized with the Health Level 7 Reference Information Model. The data model is being disseminated so users can provide feedback and use it. Next, in addition to continuing to work with states, SDOs, and other stakeholders, CDC will prototype NEDSS specifications and standards in new system development at the state level and begin integration testing. Dr. Koo noted the close relationship between this project (NEDSS) and the NCVHS projects on the EMR, the NHII, and 21st century health statistics.
In the discussion period, Committee members talked about the incentives and barriers to state buy-in to the new CDC systems, and what can be done to maximize the momentum toward interoperability.
The meeting concluded with brief subcommittee and workgroup reports and a review of future agendas.
After all present introduced themselves and he reviewed the agenda, Dr. Lumpkin asked Mr. Scanlon to brief the Committee for the Department.
Mr. Scanlon reported on a study the National Academy of Sciences is preparing to conduct on the adequacy of race and ethnicity information in health systems administered by HHS and its partners. The study was mandated by Congress (Public Health Service Act Amendment) and is funded by HHS. It is a comprehensive study of the data collections, systems and practices required under any HHS program relating to race or ethnicity data. A report to HHS and several Congressional committees based on this study is due within a year of the law’s enactment.
The wide scope of the study includes assessing the data needed to enforce existing protections for equal access to healthcare, and examining the effectiveness of existing systems and practices for that purpose. ASPE is putting together a funding consortium that is expected to begin work with a 12-member expert panel at the Academy soon. The panel will probably meet 5-6 times. HHS is proposing a workshop to deal with issues in the private sector. Mr. Scanlon said the project probably would not look at SES factors, although they are mentioned in the legislation. He noted that the Committee may want to interact formally with the Academy on this project.
In another area, the Data Council is looking at the next stage of data and survey integration, continuing a process that began several years ago. The current focus is long-term care data, and an interagency workgroup has been put together to devise a coordinated plan for addressing these data needs. It will brief the Committee on its progress. In response to questions, Mr. Scanlon said surveys of adult long-term care residents under age 65 are being considered, and HHS is looking at the data gathered in certification surveys as a good source of information. Finally, he reported that the Department is working to integrate its data dissemination through creation of a Web-based “one-stop portal” to all HHS publicly available statistics. A new user-friendly search tool will be tested in a month or so.
Dr. Braithwaite reported that action on the final security, provider identifier, and employer identifier rules are due out “sometime this year.” NPRMs are in the works for claims attachment transaction standards and for updates to already-released standards. The Secretary will issue an NPRM withdrawing the NDC standard for all transactions except retail pharmacy ones. Industry has nearly completed its recommended changes to the implementation guides for the adopted standards, which will be followed by the HHS rulemaking process to adopt the changes.
Dr. Frohboese reminded the Committee that the privacy rule, which was published in December 2000, was originally scheduled to go into effect in late February 2001. As a result of the Congressional Review Act, which allows Congress to review major regulations for 60 days prior to their effective date, the rule was delayed until mid-April. During the interim, the Secretary invited comment from consumers and covered entities, especially about changes that were made in the final rule and its workability in the healthcare context. About 11,000 comments were received during the 30-day comment period, and they were reviewed using the same procedures and many of the same people as in the previous comment period. Subsequently the Secretary and president decided to go forward with the rule as written on April 14. The Secretary is considering whether any changes are needed to enhance workability and prevent unintended consequences.
The comments revealed some misconceptions that the Department is acting quickly to address. It is doing so through a comprehensive set of guidance that will be published, mailed, and posted on the Web [www.hhs.gov\ocr\HIPAA] in the near future.
During the two years that covered entities have to come into compliance with the rule, the Office of Civil Rights will build its internal capacity and engage in public education, outreach, and technical assistance. The TA will be tailored to each type of covered entity. Meanwhile, the General Counsel’s office is developing an enforcement regulation, which will give consumers and covered entities details on the enforcement process. Dr. Frohboese stressed, however, that the Department wants to focus on voluntary compliance wherever possible.
In response to questions, Louis Altarescu of the General Counsel’s office said the enforcement regulation and penalties, as well as the complaint process, will apply initially to all the regulations, including both privacy and security. This may be modified over time to build in greater flexibility. As regulations are added or modified, the enforcement regulation will be modified as needed.
Mr. Blair pointed out that NCVHS has been active in advising and holding hearings on these issues. Committee members and the OCR representatives discussed ways the two bodies can collaborate in the future, for example by using NCVHS to host forums on privacy regulation implementation, with advice from OCR about whom to invite. One such hearing is taking place in August, convened by the Subcommittee on Privacy. On another topic, Dr. Frohboese and Mr. Scanlon described activities in the Department itself to come into compliance with the rules. OCR will be providing technical assistance to federal agencies as well.
Dr. Clancy, who directs the Center for Outcomes and Effectiveness Research at AHRQ, said her talk would set the stage for thinking about disparities in healthcare as a quality problem. The growing interest in reporting on clinical performance and the possibility this raises of creating perverse incentives for providers to avoid sicker patients make this an important time to examine disparities as a quality of care issue. In addition, more and more publicly funded beneficiaries are being enrolled in managed care arrangements. Hundreds of studies, she said, show that African-Americans are less likely to receive evidence-based treatment.
Patient characteristics are among several non-clinical determinants of health outcomes¾a “vigorous field of inquiry in health services research,” although the pathways of these determinants are as yet unclear. Getting at the disparities associated with socio-economic position, race and ethnicity is a critical quality improvement opportunity, but one that cannot be realized without the right data and strong leadership. AHRQ recommends that performance measures should be stratified, that populationwide measures should be adjusted (as other countries have done), and that both SES and race/ethnicity should be accounted for. The question is how to get the data to do this. Currently, performance reporting does not include this information.
Besides the absence of data and leadership, other challenges include concerns about privacy, data quality, and possible misuse of data, as well as resistance and inertia by health care organizations. There are controversies around questions of what proportion of observed disparities can be attributed to health care, the boundaries of accountability, and whether we know enough to intervene.
Because there is no clear description of what data currently exist in health plans or of how relevant data could/should be collected, AHRQ took part in an exploratory meeting with MCO leaders, employers, and others, with co-sponsorship from OMH and the Commonwealth Fund. NCVHS member Kathy Coltin participated. Discussions pointed to the variability in current capacity in this area and the question of whether the initiative to measure performance should extend beyond MCOs to include other kinds of health care organizations. There were indications of both employer and purchaser interest in the information, and interest in developing a business case for collecting the requisite data. The question of how to collect the data accurately and efficiently remains a big challenge. In addition, there are concerns about the perceived misuse of data, about privacy of the information, and about potential legal barriers. Finally, data on income and education are very difficult to obtain.
AHRQ, which last year started a new contract for an integrated delivery system (IDS) research network, has three task orders in progress in this area. One is to Denver health to examine the capacity to conduct studies on the impact of race/ ethnicity on access, use, and outcomes of care. Denver will look at the Intermountain Healthcare System and the New York Presbyterian Hospital System as well as their own. They will describe the data, analyze their utility, identify strategies to enhance capacity, report on what data sets could be used to conduct this research, and present the results to chief clinical officers of these organizations. The other task orders are to United Health Care and RAND to examine variations in cardiac performance measures associated with race/ethnicity and SES.
Finally, Dr. Clancy said AHRQ plans to publish a report on disparities in healthcare, and it will include the Committee among those it asks for suggestions for measures.
Dr. Carter-Pokras noted that the Minority Health and Health Disparities Research Act of 2000 calls for a National Academy of Sciences study of HHS data systems and their ability to look at race/ethnic disparities in healthcare, access and quality. ASPE is taking the lead on this project, for which the procurement process has begun. Managed care organizations’ concerns in this area relate to cost and confidentiality as well as their perceptions of legal barriers to data collection and awareness of anti-discrimination obligations. (She noted that in California there is a petition effort to put a “racial privacy initiative” on the ballot next year that would prohibit the state from classifying individuals by race, ethnicity, color, or national origin in public education, contracts and employment.) In addition, there are many collection and data quality issues, such as those concerning who identifies a person’s race/ethnicity, changes over time in preference and categories, and the need for more detailed information.
HHS is working hard to address legality concerns in this area, to assure providers and plans that it is not illegal at the federal level to collect this information if it is used to improve healthcare quality. OMH has funded the National Health Law Project to review state laws and regulations governing the collection and reporting of data by health plans and insurers, and a parallel project funded by the Commonwealth Fund to review federal laws and regulations for Medicare and SCHIP.
Violet Woo presented the findings from the first of these projects (which will also be published in a report). So far, data have been collected from databases and in telephone conversations. Next, there will be visits to selective states for more in-depth analysis. Four states (CA, MD, NH and NJ) have specific legislation or regulations prohibiting either insurers or MCOs from asking about race/ethnic identity on application forms. 41 of 51 jurisdictions provide some level of protection against discrimination on the basis of race or national origin for health insurance consumers or managed care enrollees. Ms. Woo noted that decisions about the collection of race/ethnicity data for Medicaid are left to the states. Most report it to the federal government, but it is not required. 47 states collect race/ethnicity data from SCHIP applicants.
Dr. Carter-Pokras then reported on research by Dr. Ruth Shire reviewing federal policies and practices in this area. It found that health-related data collection by race, ethnicity and primary language is legal and fully authorized under Title VI of the Civil Rights Act of 1964. No federal statutes prohibit collection of this information, and in fact three health services program statutes and two health services regulations require it. Dr. Carter-Pokras then briefly reported on other projects addressing data quality concerns, and a project by Dr. Tom LaViest of Johns Hopkins University, on measuring racial and ethnic discrimination in healthcare settings. She noted that Dr. Braithwaite has been key to an effort by CDC and ASPE to develop a business case for including race/ethnicity data in encounter data.
Dr. Nerenz, Director of Health Care Studies at the Michigan State’s Institute for Managed Care, reported on a project on data for health plan report cards on quality of care for minority patients. The project is funded by HRSA and the Commonwealth Fund. Generally, comparative health care quality report cards do not contain data on race/ethnicity; this leaves both consumers for whom this is an issue and organizations wishing to close the gap with no valid and focused source of information. The planning phase of the project has been completed, in which a set of health care concepts, measures and clinical conditions were recommended along with policies and procedures to follow in collecting and reporting data for a report card. Three healthcare organizations collaborated in this phase, and input was sought from a technical advisory panel and from community leaders in the African-American and Hispanic communities. These groups generated three broad policy recommendations:
In phase 2, the project is working with a diverse group of nine plans to assess the feasibility of various methods of obtaining data on race/ethnicity and using them to generate stratified quality reports. A recent HRSA contract is making it possible to involve six additional plans. Results are expected by the end of 2001. The key questions concern the ability to get the data and link it to other datasets, and what variations they will reveal, both within plans on quality of care for different groups and across plans on quality of care for specific groups. Three targeted surveys are being administered to determine consumer assessment of their health plans. Two variables that were of special importance to community groups are patient involvement in decision making and adequacy of information for self-management. Patient groups were surveyed on aspects of care for asthma, diabetes, and prostate cancer. Cultural and linguistic competence of programs is also being assessed using objective criteria as well as member satisfaction.
None of these plans normally has race/ethnicity data on their members, and they are testing one or two of five different methods for obtaining the information. The questions being looked at include whether the plans can and should collect the data, when and how they should collect the data and from whom, and how the data should be used.
Asked whether the project found that plans that were unable to participate experienced common barriers in trying to do the things the nine participating plans are doing, Dr. Nerenz said one-quarter to one-third of the plans contacted are participating. It remains to be seen whether the findings from this group with “unique capabilities” can be applied more broadly to health plans.
Asked to comment on the difficulty of getting data on socio-economic position, Dr. Clancy said several studies and surveys have shown that people have greater reluctance about revealing this information than they do about many other seemingly more personal matters. Education is somewhat easier to obtain. Dr. Carter-Pokras said the Census found the same thing.
Regarding plans’ commitments regarding their use of the resultant data, Dr. Nerenz said the only commitment is to the research project; any other uses of the data will be voluntary.
Mr. Rothstein observed that in society today there are “diametrically opposed political forces” concerning race/ethnicity data: those who argue that race is nothing but a social construct with no scientific value, and it therefore shouldn’t be used, and those who say information on race/ethnicity is needed to achieve equity in society. He asked whether the research plans are taking these currents into account. Dr. Carter-Pokras granted that the first point of view is correct but she argued that race/ethnicity data are still needed to detect and correct disparities. She added that politically, common ground must be found between the two sides. Dr. Clancy observed that framing the challenge as an issue of fairness may be unifying, along with the goal of improving quality of care for everyone.
Dr. Lumpkin welcomed Dr. Lisa Iezzoni, who recently completed her term as an NCVHS member, and thanked her for her continued hard work on the report she had returned to present. He noted that the subject has been important to the Committee for a long time.
By way of background, Dr. Iezzoni said that in its core data elements report in the mid-1990s, the Committee recommended a place-holder for functional status and further work by HHS to determine what to fill it with. In 1999 the Subcommittee on Populations decided to study this question. It began with a broad survey of functional status measurement, and quickly concluded that it was not feasible given resource limitations to do anything meaningful on the complex question of measurement. Because HIPAA had created attention to and work on code sets, the Subcommittee decided it was appropriate to focus on a classification developed by the World Health Organization, the International Classification of Impairments, Disabilities and Handicaps (ICIDH). WHO was completing a 10-year revision, which was approved in May 2001 as the International Classification of Functioning, Disability and Health (ICF). The Subcommittee held three hearings where testimony was provided about the utility of functional status reporting in administrative records, the appropriateness of the ICF for this purpose, and what steps would be needed to use the ICF for functional status reporting.
Dr. Iezzoni acknowledged the work of Subcommittee staff, consultants, and the report’s author in this effort. She said the Subcommittee developed recommendations with four key points:
Noting that members had had a chance to review the report in advance, Dr. Iezzoni asked the Committee to approve it.
Dr. Cohn expressed support for most of the recommendations, but suggested that the Committee not recommend reporting functional status information on claims or encounters, even as a statement of principle, until the recommended research and testing are completed regarding the feasibility of using ICF in administrative records. Following discussion and a variety of suggested solutions, the Committee agreed to revise recommendation 2 in accordance with Dr. Cohn’s suggestion.
It was moved and seconded that the report, as revised, be approved. The motion passed unanimously, with thanks to Dr. Iezzoni and Ms. Kanaan. The final report will be posted on the NCVHS website.
Dr. Madans reviewed the decisions surrounding OMB Directive 15, the original directive on the collection of race/ethnicity information starting in 1977. Statistical agencies agreed in 1977 that they needed a consistent way to monitor progress toward equality. The approach started with four basic categories for collection and reporting, but agencies were free to add more detail and collapse back. Most of the issues that arose in the recent review of the Directive were also present in 1977, such as problems with the way measurements are done. The Directive represented a compromise that was “a little good for a lot of things but not best for any one thing.” Dr. Madans used data on racial differences in infant mortality to illustrate why such measurements are still needed.
The new OMB Guidance issued in 1997 reflected an awareness of demographic changes in the country and growing resistance to answering the single race/ethnicity question. There was a desire to make a change before the percent reporting multiple race got any larger. Since 1982, the National Health Interview Survey had been allowing respondents to report more than one race and asking the follow-up question, “If you had to choose, which race/ethnicity category do you relate most closely to?” Dr. Madans stressed that what is underway is a 10-20 year transition. She reviewed the categories, noting that there are now five rather than four, which causes some problems for bridging although large numbers of people are not involved. The main change is in allowing people to identify more than one race. This is especially a challenge for those agencies interested in the characteristics of the people in the various categories. Bridging is fairly easy, using retabulations. The NCHS population-based surveys are compliant and Dr. Madans said they are in “pretty good shape.” The question is whether OMB will allow NCHS to retain the follow-up question (see above), which makes the bridging possible and is a useful analytic variable in its own right.
The provider-based surveys are all in compliance with the new standard, although very few people check more than one race. Problems stem from the fact that NCHS has no control over primary data collection, which is done in the hospital. Similar problems exist with vital statistics. So people who are counted in the Census as multiracial are not showing up as such in hospital data or vital statistics. This is compounded by the fact that revised certificates will not be adopted by all states until 2003. The major concern is that for vital statistics, at least for 2000-2002, the numerators and denominators are incompatible: no multi-racial people will appear to be dying or having children, so a bridging file of some sort will be needed for this period.
The concern for the future relates to mortality data, because of the likelihood that multiple race will not be reported on death certificates, though it is an option; or it may be reported but ascribed incorrectly. Efforts will be made to assess this problem using a “mega-match” of census and mortality data. Dr. Madans stated that for 2000 rates, NCHS will base its reports on projected 1990 numbers, and it may do that for the preliminary and final reports. Another report will be done based on 2000 estimates once they “figure out how to do that.”
She discussed various implementation issues, noting that NCHS plans to do research into best practices for data collection for various modes, e.g., household-based surveys, phone surveys, vital records. One issue is what to put in the public use file, as that affects what can be bridged. The content of tables is another challenge, along with how to combine the data to establish new trend lines into the future. Trend analysis is especially important, because of the Healthy People process and its programmatic implications. A related question is what additional information should be collected in the future to track the goal of eliminating health disparities. In closing, Dr. Madans described a few of the Center’s research projects. Examples are the factors affecting what people select as their primary racial identification, and, in vital statistics, trend analysis on births to parents of different races.
Dr. Pinal said the “modern chapter” of collecting race/ethnicity at the federal level began on October 30, 1997, when OMB issued new federal regulations, now called revised standards for classifying race/ethnicity in federal data. He reviewed the changes in the new standards. One key change was to sequence Hispanic origin ahead of race in an attempt to reduce the non-response rate on that question. There were changes in terminology as well. The key changes that greatly affect data products were allowing people to select more than one race category and the creation of a new race category, Pacific Islander, splitting it off from the former Asian/Pacific Islander category. Among the race categories, “some other race” was added, along with “two or more races.” There are additional variables. The total possible combinations of the 15 check-off boxes for race gives more than 32,000 possible combinations.
Turning to the multiracial category, Dr. Pinal said everyone wants to know the extent of the reporting of two or more races. In the dress rehearsal, 5.4 percent of the Sacramento, CA population and less than 1 percent of a SC county chose this category. Nine-tenths of the people who selected “two or more races” selected only two. Data from the American Communities Survey and the 2000 Census show the same pattern. Hispanics are especially likely to select two race categories. In general, no racial category had more than one percent that also reported themselves as multiple race. There is considerable variation among states in race/ethnicity reporting and counts.
Dr. Pinal showed some tables and graphs suggesting how the Bureau is interpreting and portraying the 2000 Census information. He noted that 40 percent of American Indians also selected another race category (compared to 5 percent of Blacks and 14 percent of Asians). He also showed that the new categories have dramatically affected the counts for the smaller population groups; for example, the Asian population grew by either 48 or 72 percent, and the Pacific Islander by either 9 or 140 percent, depending on how the numbers are interpreted.
Mr. Harrison stated that “the federal statistical system is in great jeopardy of becoming virtually very difficult and very meaningless for conducting the kind of analyses by race that you’ve conducted in the past.” He cited some of the statistics reported by Dr. del Pinal in support of his statement, noting the wide ranges of possible numbers and pointing out that “these are the denominators you will be using through much of the health surveillance system.” The numbers are especially meaningless for American Indians, and in some areas for Asians and Blacks. He argued that what is needed is adequate bridge statistics to tell what the population distribution and characteristics would have looked like absent the revision to the new race/ethnicity standards. The Census Bureau will be conducting a bridge study, but he asserted they are also needed at the county level given the wide regional variations and the need for accurate local data.
Another problem is how to construct meaningful trend statistics when observed changes may be a function of nothing more than changes in the reporting systems. This cannot be teased out without bridge statistics.
In addition to the “bridge to the past,” another issue is the bridge to the future ¾ making sense of future events based on current statistics. Mr. Harrison cautioned against mistaken assumptions based on Census and other statistics¾for example, assumptions about the multiracial population, when NCHS research suggests that only half of the population with racially mixed parents are reported as multiracial in the 2000 Census. He noted the flaws in drawing conclusions about racial disparities in health because of the flaws in the race/ethnicity information. The underlying question is whether the statistic has any real-world meaning. He noted that the Committee raised concerns about such issues in its letter to Katherine Wallman of OMB when the revised directive was under consideration. Mr. Harrison said much time must be spent on how to present meaningful statistics on the group that reported more than one race, “a group that nobody can define independently.”
Another problem is the inconsistency of reporting more than one race: 50 percent of those interviewed in the dress rehearsal reported differently in the 2000 census, and there is also evidence that people are not reporting consistently to school systems, employers, hospitals, and so on. This makes “horrendous numerator/denominator matching problems” very likely.
Again, Mr. Harrison called on the OMB, the Census Bureau and the statistical system to take seriously the issue of providing adequate bridge statistics. He suggested that OMB’s “political sensitivities” overcame its professional judgment.
Dr. Lumpkin observed that we are in the midst of dramatic social change with respect to the meaning of race/ethnicity and people’s ability to define themselves, even as understanding is growing that race is merely a social construction with no scientific validity. He thanked the speakers for their thought-provoking presentations and expressed hope that the Populations Subcommittee would study and propose recommendations on this issue, in view of the importance of eliminating health disparities.
Dr. Cohn, who chairs the Subcommittee on Standards and Security, read a draft transmittal letter to the Secretary, recommending that he accept all the changes to the HIPAA standards that have been recommended by the DSMOs, and that the process be expedited. The letter is to accompany the 67 changes, mostly technical, recommended by the DSMOs. Dr. Cohn accepted a friendly amendment to add the names of the DSMO representatives who worked on this in the body of the letter, in appreciation of their efforts. He explained that the implementation guides will be modified to accommodate the recommended changes. The Committee, which proposed no substantive revisions to the letter, will vote on it in day two of this meeting.
The Committee then reviewed an early draft of a letter to the Secretary presenting a series of recommendations based on recent Subcommittee hearings on implementation of HIPAA standards and industry readiness. Dr. Cohn read the key portions of the draft, and Committee members offered comments and suggestions for the Subcommittee to consider at its afternoon meeting. No fundamental disagreements with the document were voiced. The group discussed the pros and cons of suggesting a measure that would introduce more flexibility into enforcement by allowing submission of a plan to come into compliance within 12 months.
The Committee then recessed into subcommittees for afternoon working sessions.
Dr. Meyer reported on an HHS project by four collaborating agencies: AHRQ, the Center for Medicare and Medicaid Services, FDA, and CDC. The Patient Safety Task Force is chaired by Dr. John Eisenberg. Its goal is to help provide “opportunities to learn from errors.” Dr. Meyer noted that besides all the errors that cause harm to patients, there are many more “no-harm events” and “near misses” from which providers could learn.
The public is demanding information on safety events: a Kaiser Family Foundation pool found that 73 percent of its sample supported requiring health care providers to report medical errors. Some existing systems work, such as the National Nosocomial Infection Surveillance System (NNIS), which combines reporting with a system to develop programs and interventions to reduce risk in this area. Nevertheless, reporting is controversial, and there is a tradeoff, or need for balance, between reporting for accountability and reporting for change or improvement. The balance point is what David Marx calls a “just culture,” where there is both trust and motivation to report.
The Institute of Medicine report on medical error called for steps by the federal government to develop reporting systems. This is being done through contracts with the National Quality Forum, which has been asked to develop two lists: a small list of egregious and easily preventable (and also rare) errors in hospitals; and a list of patient safety practices following the systems-based approach recommended by IOM. AHRQ will soon release a commissioned report on the evidence base behind various patient safety innovations and practices. The Quality forum will identify proven measures that warrant implementation.
The Patient Safety Task Force (PSTF) looked at the existing reporting systems in both the private sector and federal programs, and commissioned a report on existing state-based systems. Several such systems exist or are under development, but generally they are very uncoordinated and are also what Dr. Meyer called “a data graveyard,” ignoring the responsibility to analyze the data and feed it back to those doing the reporting so they can improve their practices. So the quality improvement loop is not closed. The systems are redundant; taken together, they result in a breakdown of the patient safety effort related to reporting, because the systems are not adding value to the reporting function and therefore people are not using them. A single event often requires multiple reports to different agencies and departments, using different forms.
The PSTF is looking at the existing reporting systems and considering what it can do “on the front end” among the participating agencies to make the interface for reporting more user-friendly and reporter-friendly so the information can be sent to a single place and then sent to the appropriate federal agency. Then, on the “back end,” it is looking at how to integrate the data to create learning opportunities. Attention would be given to making the data available to researchers and policy makers, something that is not happening with existing systems. This effort is designed to get the four collaborating agencies’ reporting systems to work together to improve the quality of care.
Noting the difficulty researchers have today in gathering data on quality of care, and the very real human consequences of that barrier as a result of undetected unsafe practices, Dr. Meyer said the ultimate goal is to produce a common user interface. There would also be a federal data pool of de-identified data that would be useful for researchers and others, with integrated data sets.
A reporting summit was held at the end of April at which stakeholders were told of the task force’s plans and invited to give input about what would be useful to them. The summit showed the desire for leadership and openness to the efforts of DHHS. Burden was highlighted as a big challenge, and there were calls from numerous stakeholders for work on language and vocabulary, and also for opportunities for public input into the process. Dr. Meyer stressed that this system is designed for improvement and learning, not accountability, and participation will not be mandated. The task force is also working with international collaborators, including Australia, New Zealand and the UK.
An implementation planning/feasibility study is planned and a pilot project is being considered, as is future work on vocabulary and data standard issues. Partnership building is also underway. Updates are available on the website, www.ahrq.gov.
Dr. Starfield asked if the task force was including adverse events in its scope, noting that these are different from accidents and errors. Dr. Meyer said the word "hazard” may be a non-restrictive frame for these issues, as hazards can have a range of causes. Also, the notion of reducing risks to patients could pull in a range of concepts. He noted that there are boundary questions to be addressed about what is regarded as a quality issue.
Dr. Shortliffe proposed that this collaborative effort, as exciting as it is, might be too narrowly defined given the need for a “shared big-picture vision” across all HHS agencies and encompassing collaboration around clinical data management, reporting, electronic medical records interfaces, and more. Dr. Meyer said he had articulated “one of the key discussions” within the task force about the larger implications of the project and the need to interface with many other activities in the public and private sectors. The conclusion after much discussion is that the task force needs to “keep this effort small,” while keeping an eye on the big picture and doing lots of outreach to stakeholders to keep them apprised of what is happening and to hear their views. A starting point for the thinking is the IT notion of open architecture. One activity with potentially broad applications relates to vocabulary. Dr. Shortliffe urged that technical issues be approached with great care to ensure that vendors and developers of data systems can develop simplified systems that work for multiple reporting applications.
The NQF was conceived by the Presidential Advisory Commission on Consumer Protection and Quality in the Health Care Industry, which in 1998 recommended a patients’ bill of rights and creation of this forum. A planning committee was convened that conceptualized a governance structure and operational mode, leading to the formation of a 501.c.3 public benefit corporation in mid-1999. Dr. Kizer is president and CEO of NQF, which became operational in February 2000. The primary purpose of the Forum is to increase the provision of high-quality care. Its strategy is to standardize health care quality measures and reporting, to encourage public disclosure of this information, and to encourage that health care purchasing decisions be made on the basis of quality of care data. The data must be uniform to be meaningful.
The Forum, which is “an organization of organizations,” has broad stakeholder representation from both public and private entities on its board of directors. CMS and AHRQ are both members, along with state agency representatives, private purchasers and others. The bylaws require that the majority of the board represent consumer (e.g., AARP) and purchaser (e.g., General Motors) organizations. NQF meets the OMB definition of a voluntary consensus-setting body, thus permitting federal employees to sit on the board. The Forum convened a strategic framework board to lay out a national strategy for health care quality measurement and reporting. Its report is going through the formal consensus process. The prominence of the Forum’s board members makes the measures it endorses likely to be operationalized. There are member councils for consumers, purchasers, providers and health plans, and research and quality improvement organizations, and each is represented on the board. The board has liaison seats for several organizations, including JCAHO, NCQA and IOM.
Dr. Kizer stressed that the Forum does not intend to create new measures; it sees its niche as reviewing those that exist, evaluating their evidentiary base and other factors, and assessing whether they meet national priorities for improving health care. The Forum is also involved in helping establish consensus in the latter area. Dr. Kizer described a few Forum projects. Examples are standardizing performance measures for acute care hospitals, compiling a compendium of best practices for patient safety, looking at quality metrics for minority populations, and convening a summit (February 2002) on information management technology and possible ways to “break the logjam.”
Mr. Blair asked how the Forum was relating to other groups addressing the same issues, such as standards development organizations. Dr. Kizer said the Forum can help bring all the parties together, including consumers and providers, to reach consensus, leading to implementation by private and public sector entities.
Mr. Scanlon raised the question of voluntariness and whether it is possible, as has been suggested, for federal agencies to circumvent the rulemaking process in adopting standards approved by the Forum. Dr. Kizer said “some of this remains to be worked out,” but much depends on buy-in and support for the Forum’s policies. He noted that with buy-in there is the potential for government agencies to operationalize practices immediately, along with the private sector. He agreed that this represents a market force rather than a requirement. Dr. Fitzmaurice likened the adoption of a quality measurement set to adoption of additional codes in a codeset, which does not require a new regulation. Dr. Kizer stressed the potential power of a voluntary effort.
The IOM recently released a report envisioning the National Health Care Quality Report, which was recommended by the aforementioned President’s Advisory Commission. The Health Care Research and Quality Act of 1999 included a provision directing AHRQ to submit an annual report to Congress looking at national trends in health care quality. AHRQ asked the IOM to set up a committee to advise on that report. The committee was chaired by William Roper, and it released its report a few months ago. The two audiences for the report are the policy and legislative community and the American public at large. There is a strong feeling that the public needs to be brought along about quality of care and the need to improve, to have the political leverage to take strong action.
There are ten recommendations, falling into four areas: the framework and categories; measure criteria and guidelines; data sources, collection and analysis; and audience-centered reporting criteria. The committee used a two-dimensional conceptual framework: first, the components of health care quality (safety, effectiveness, patient-centeredness and timeliness); and second, consumer perspectives on health care needs (staying healthy, getting better, living with illness, and coping with end of life). These two dimensions laid out on the X and Y axes of a matrix create “the framework.” The committee viewed equity as a cross-cutting issue.
The committee adopted criteria for the selection of measurements, both individually and overall. It recommends that a broad-based independent committee or advisory group help assess and guide improvements in the report over time. The IOM report tries to envision the quality report care several years down the road, but it will take time to get it right. The goal is to have not only individual measures of aspects of medical care but a comprehensive set of measures that cross all clinical areas, cover all subpopulations, and pick up all the domains including preventive care. The committee felt that the key to being able to do what is needed in the long run is to use this tool to drive investment in automated clinical information systems. In the meantime, it will be necessary to draw on survey data and other data sources.
The committee recommended that the report have several versions, tailored to different audiences. The challenges that lie ahead include populating the framework matrix with a sufficient number of measures, establishing a comprehensive quality data set that is accessible to the public and to researchers, keeping the report focus narrow (3-5 findings), and identifying summary measures as needed.
On other IOM initiatives, Dr. Corrigan reported that it is planning to start an information technology and quality project by the end of the year, looking at what constitutes the NHII and in particular the resource requirements for it. The Institute will also be looking at next steps to follow up on the Crossing the Quality Chasm report and the report on medical error. An effort is also underway to take a close look at the Congressionally-mandated study on quality assurance and improvement efforts in Medicare, Medicaid, SCHIP, DoD, and the VA. A report is due on the last project in September 2002.
Dr. Reilly stressed that the quality report is a Department-wide activity. He began by discussing the measures that populate the framework outlined by Dr. Corrigan. The report, he said, is a mechanism to track progress toward the aims laid out in the IOM Chasm report, using the elements in the matrix, which cover the breadth of health care. He reiterated that the equity criterion is cross-cutting. HHS is now developing and identifying measures to populate the matrix, with input from the private and public sectors about possible measures. Nearly 700 measures were recommended as candidates, which are being reviewed using the criteria laid out by IOM. Dr. Reilly expressed hope that NCVHS would help move this agenda forward.
The process is also setting priorities around conditions. The Chasm report suggested 15-20 priority conditions. Using the four priority areas used in Healthy People 2010 (infectious disease, heart disease and stroke, cancer, and diabetes), Dr. Reilly illustrated by describing what indices might be used to evaluate quality of care in various cells of the matrix, particularly with respect to effectiveness. (More cross-cutting issues may be developed for measuring patient-centeredness and safety.) He noted that a general principle of this project is to align it with other Departmental activities.
Turning to data sources, he said the first report will use a “mosaic” of sources, using various population-based data collection efforts, establishment and provider-based data, administrative and regulatory data, vital statistics and surveillance activities, and so on. Attention is being given to the gaps in current data systems and how existing data collection can be adapted to fill them.
As for the reports, he said there will be multiple products to meet multiple needs, including a written report to Congress, web-based reporting, press materials, and more, with queries to stakeholders as to what would be useful. An inventory is being done of existing reporting projects, with attention to other attempts to do such an inventory. The first report is due in September, 2003. To meet that deadline, final measure sets must be selected by the end of November, 2001. Eight months are being reserved for the clearance process. Related efforts include preparing the other reporting products, developing a research agenda, developing composite measures, expanding existing data sources, increasing state participation, and establishing an external advisory body through which to get input from external stakeholders. Dr. Reilly said he hoped for ongoing opportunities to work with NCVHS on this project.
Dr. Starfield hailed the project’s recognition that this effort must go beyond effectiveness measures, and she asked about the thinking on the other criteria. She noted that addressing co-morbidity is one facet of patient-centeredness and of the shift from a disease focus. Dr. Reilly said the work on patient safety is helping to conceptualize this area. Dr. Meyer observed that the Quality Report matrix provides a stark reality check on the current inability to measure many aspects of quality¾so it is powerful in what it can’t tell us as well as in what it does.
Dr. Lumpkin and the Committee affirmed that they would like to explore further ways that NCVHS can contribute to this effort; it will develop a specific proposal on this. Mr. Scanlon noted that among other things, NCVHS can hold public hearings, review and deliberate and give feedback, and be a sounding board. Dr. Reilly said it hopes the Committee will be a conduit for getting advice to HHS on this effort.
Ms. Coltin observed that the speakers identified two areas where they need advice: in filling out the framework with measures, and on the data needed to create those measures. She said this suggests a natural partnership because NCVHS through its Workgroup on Quality has been talking about these areas.
It was moved and seconded that the letter on the updates of the standards recommended by the DSMOs, presented to the Committee the previous day, be approved. The motion passed.
It was moved and seconded that the letter conveying NCVHS recommendations based on recent Subcommittee hearings on implementation of HIPAA standards and industry readiness, also discussed the previous day by the full Committee and the Subcommittee, be approved. Dr. Cohn reviewed the changes from the preliminary draft. Members and staff suggested a few additional changes. The motion passed.
Dr. Mays then presented a letter drafted by the Subcommittee on Populations concerning the inclusion of race/ethnicity data in SCHIP regulations. A motion approving the letter was passed unanimously.
A motion approving revisions to the functions and process of the Executive Subcommittee was passed unanimously.
A motion approving a revised charge for the Workgroup on Health Statistics for the 21st Century was passed unanimously.
Dr. Koo began by saying that NEDSS is not a monolithic system but an interconnected system of systems. It is not limited to data about diseases but can support other conditions and syndromes. Public health differs from medicine and health care in that it focuses on the health of the population, emphasizes prevention, covers the full causal pathway to disease, and is carried out in a governmental context. She noted that while clinicians try to determine what kind of disease a patient has, public health “tries to determine what kind of person has a given disease, so we can try to identify risk factors and prevent disease.” For this effort, public health surveillance is a systematic and ongoing diagnostic tool that links to public health practice. It is the cornerstone for an approach that starts with a problem/potential problem, looks for the cause and risk factors, devises an intervention, implements it, and then evaluates its effectiveness.
Surveillance is just one of many information system functions needed for public health; others include notification, communications, and knowledge management. In the current situation or “mess” with legacy systems, individual systems that go from local to state to CDC were developed in large part as a result of categorical funding by Congress. These systems do not talk to each other and are a great burden to CDC’s local and state partners. Most are DOS-based, with the data, which are often incomplete and not timely, originating on paper forms.
Change is motivated by the need for more information on a broader range of issues; the increasing use of electronic information systems, with CDC’s partners building their own information systems; increasing efficiency; and new opportunities to enhance security and confidentiality. The long-term objectives of NEDSS are to have ongoing, automatic capture and analysis of data, to use data that are already electronic, to have systems designed on relevant data sources, not diseases, and to integrate public health and health care systems. Dr. Koo gave an example of how, once this is achieved, a physician using electronic medical records could easily get a differential diagnosis for a patient, be reminded of recommended actions relevant to public health, and ultimately automatically send notifiable disease test results (e.g., TB) to public health. The same system would allow public health to automatically track resistance to an antibiotic and notify healthcare providers and embark on an educational campaign.
To achieve these goals, CDC has developed a NEDSS system architecture with eight elements based on national or de facto industry standards, which can facilitate electronic interchange with the health care system. CDC has awarded NEDSS funding to all fifty states, to undertake assessment and planning for NEDSS-compatible systems. The CDC also is collaborating across its categorical programs, and implementing sophisticated security standards to protect sensitive data. Even before all the goals are achieved, CDC is working on capturing existing data electronically and learning how to use the information. Several pilot projects are underway with state and local partners, as well as with managed care.
The Public Health Conceptual Data Model provides the framework for what categories of data are needed for public health, especially surveillance. Dr. Koo showed a schematic of the model, which was developed with attention to the Australian and Canadian models and the HL7 reference information model (R.I.M.) as well as CDC systems. The purpose of the model is to reduce development efforts for computerized systems and to enhance public health’s data exchange capabilities. It has been helpful in representing public health data needs to the SDOs. CDC is working to harmonize the model with the evolving HL-7 R.I.M. The four subject areas in the model are parties, location, material, and health-related activity. After discussing a few details in each area, she discussed where the data model fits into the eight elements of the NEDSS State System architecture. Among other things, CDC, working with Computer Sciences Corporation, is developing a state “Base System” that States can choose to use as a platform for other modules. The data model can be mapped either to independent state systems or to programmatic CDC systems. The data model is being disseminated so users can provide feedback and use it. CDC has developed a logical model for the NEDSS-based system for people to use in designing their physical databases. It is working on message specifications for data interchange and vocabulary, and it hopes to disseminate the process for mapping systems to the model.
Some of the issues and barriers around NEDSS include privacy concerns, data ownership and access issues, and assignment of roles and responsibilities. Next, states will use the NEDSS specifications and standards in new system development. By October 2001, CDC will have awarded funds to approximately 34 states, many of whom have selected the base system option. CDC continues to work with SDOs and with the Public Health Data Standards Consortium. In closing, Dr. Koo noted the close relationship between this project and the NCVHS projects on the EMR, the NHII, and 21st century health statistics.
Dr. Mays asked Dr. Koo to comment on the uses of data for research purposes and on the ethical issues related to ownership of data for research and tracking. Dr. Koo said the “practice versus research” issue is a big one that has generated lots of discussion in the public health community. She noted that public health needs personally identifiable data to do its job (i.e., for public health practice not just research), and it is difficult to differentiate this in the public’s mind from research and why public health does research, as well. Dr. Koo and Committee members discussed the ethical issues of ownership. She pointed out that it is the state and local levels that need personally identifiable data for disease surveillance; they are used following confidentiality guidelines and shared voluntarily (without identifiers) with CDC. Dr. Mays mentioned the data ownership issues for people who have tests done in a public health setting. Dr. Lumpkin observed that data collection is dependent on state laws that define who owns the data, and the federal authority for collecting data is less clear in the public’s mind.
At Dr. Shortliffe’s request, Dr. Koo reviewed the evolution of the “silo” systems that are now the bane of comprehensive public health practice and integrated data systems. Congressional funding for different programs (TB, STDs, AIDS, etc.) led to the development of different data systems in different programs at federal, state and local levels. There have been efforts at coordination, infrastructure development, and integration since the 1980s, but they had not yet succeeded, and there were essentially no resources to work with until the Congressional funding for NEDSS. Dr. Lumpkin pointed out that ASTHO and NACCHO did some work with the Public Health Foundation and began to put pressure on the Assistant Secretary for Health and agency heads. This plus calls from CDC for integration created an environment where the issue finally got attention. Dr. Shortliffe commented on the challenge that lies ahead related to linking to the many different legacy systems in the states; Dr. Koo agreed that this is a huge challenge but added that when states upgrade their systems, CDC can offer them a modern upgrade. Dr. Lumpkin said the Conceptual Data Model is very important to the states in the upgrade process. Dr. Friedman said in the last few years states have come to accept the need to build systems that are internally integrated and interoperable with CDC systems. There is still a long way to go, however. In response to another question, Dr. Koo commented on the complexities of changing the systems regarding notifiable diseases.
Following on an earlier point, Mr. Scanlon cited the federal laws that authorize the activities that federal agencies engage in. He observed that HIPAA is another factor that has created momentum toward interoperability and an integrated infrastructure.
From the audience, Roy Rada from the University of Maryland expressed skepticism about whether states will voluntarily comply with CDC’s approach, and he asked what beside “good feelings” can be done to make it happen. Others commented that the market has not provided systems to meet these needs because the number of customers is limited, there is no money, and “the customers in general don’t get it.” Dr. Lumpkin reasserted that there is a sense of good will and of mutual benefit among the states. He thanked Dr. Koo for her presentation and said the Committee would remain interested in the progress of this initiative.
Dr. Lumpkin reported that the Executive Subcommittee would hold a day-long meeting in Chicago on August 14.
Dr. Cohn reported that the Subcommittee on Standards and Security would hold a hearing on August 20-21 on PMRI. Other hearings are tentatively planned for October, November, and December. The Subcommittee plans to follow implementation of HIPAA administrative transactions, work on PMRI, and start work on code sets in the next year.
Dr. Rothstein reported that the Privacy and Confidentiality Subcommittee will hold a hearing on August 21-23 on the topics of consent, minimum necessary, research, and marketing. Speakers will be invited, and there will be time for public testimony.
Dr. Mays reported that the Populations Subcommittee is interested in pursuing the topic addressed in this meeting, of implementation issues resulting from the new OMB race/ethnicity standard. It plans to meet with federal agency representatives on how they are implementing the standards, and then to hold a hearing to bring in other users to discuss the issues they are encountering in using the data. The Subcommittee also has discussed looking at variables other than race to understand disparities, such as SES. It has deferred decisions about long-term projects until new members are appointed.
For the Workgroup on 21st Century Health Statistics, Dr. Friedman reported that the group, working with its CDC and NCHS counterparts, has developed a schematic outline for its final report. It plans to send out a structure for recommendations in the summer to stakeholders to invite their submission of recommendations. A more complete draft should be ready for the September meeting.
Ms. Greenberg reported that the Workgroup on Quality plans to commission a report pulling together the findings from the several panels on quality it has organized over the last year or more.
September meeting: Subcommittee on Privacy and Confidentiality recommendations.
Members were asked to notify Debbie Jackson of any action items for the September or November meeting agendas.
With that, Dr. Lumpkin 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 9/26/01
________________________________________________________________
Chair Date