The NCVHS Subcommittee on Populations was convened on Thursday, September 17, 1998 in the Hubert H. Humphrey Building in Washington, D.C. The meeting was open to the public. Present:
(ACTIONS: See section below on "Work Plan and Charge." Actions are indicated in bold face.)
Dr. Iezzoni explained that the purpose of this meeting is to inform the Subcommittee about the background, status, and future of survey integration. In the afternoon, the Subcommittee discussed their findings and what actions they wished to take in response.
Dr. Wunderlich set the context for the discussion by describing the 1992 NAS study assessing survey integration efforts. She noted that the study's findings are still relevant. After examining existing surveys, data needs, and the proposed NCHS plan for a proposed National Health Care survey, the NAS committee recommended a design framework for a coordinated and integrated system of health care data collection activities involving several organizational entities; and, the development of a phased multi-year strategy for implementing the framework for a national health care data system. It also recommended that NCHS establish a research agenda. Dr. Wunderlich reviewed both the thinking and the recommendations of the NAS Committee, in regard to survey integration itself, the advisory structure, and research agenda.
Ed Hunter of NCHS then briefed the group on the Department's activities in this area since then. The NAS Committee report and other committees have noted the fragmentation of HHS data and health care reform provided another stimulus in prompting the Department to look at data gaps. After analyzing the various "stovepipe" systems, a workgroup led by Ed Hunter and Ross Arnett developed an integration plan within the Department and got OMB clearance in April, 1995. The Secretary created the Data Council at about this time, with the charter to coordinate across the Department; to look at survey plans as they are developed within the agencies and to stimulate action within agencies; to look at budget priorities across the Department; and to advise on what the agencies should include in their proposals. The Data Council has a work group on survey integration, which Mr. Hunter and Mr. Arnett co- chair. As senior advisor to the Secretary, Dr. Sondik also supports data planning and a broad view of how agency efforts can be integrated.
Implementation of integration began in 1995 with the sample and questionnaire linkage between the NCHS National Health Interview Survey (NHIS) and the AHCPR Medical Expenditure Panel Survey (MEPS). Mr. Hunter described this linkage, which involved sample frame, timing, and question content. Other successful integration efforts include combining health insurance survey of employers and the internal (NCHS) integration of NHIS and NHANES.
A current focus of survey integration is provider surveys, where the same problems of overlap, differing operationalizations, apply. The provider problem remains to be solved: there is not a consensus on priority data needs in this area, as well as identifying the conceptual model of health care delivery to be used. This requires setting analytic priorities. A work group in the Department has been working on that for a year or more, and it has evolved into a public-private collaboration. Other current steps include broader inter-Departmental review of employer surveys and linkage of NHANES with USDA’s nutritional survey. Some steps outlined in the initial integration plan have not been accomplished, including potential integration of the MEPS and the Medicare Current Beneficiary Survey. NCHS is currently pilot testing an example of an integrated approach to State level data collecting SLAITS. Discussions evolved to the next steps in the data integration process an on issues relating to confidentiality and how NHIS-MEPS may limit data release.
The HCFA presentations focused on several examples of survey integration. Mr. Reilly explained the Consumer Assessment of Health Plans (CAHPS), which uses the HIS as a sampling frame. The current targets are Medicare enrollees, although commercial and Medicaid enrollees are also targeted for the future. Dr. Goldstein and Mr. Sekscenski reported on applications of CAHPS to managed care and fee-for-service, respectively. The latter survey, being piloted in five states, is 18-24 months behind the former. Finally, Dr. Hardy furnished additional details about using the Health Interview Survey as a sampling frame to link other surveys: the National Survey of Family Growth (NSFG), the Medical Expenditure Survey (MEPS) and NHANES. She also briefly discussed the benefits of linkage from the agency perspective, and mentioned several methodological issues.
The discussion period focused on the applications and limitations of CAHPS for obtaining data on gender, race and ethnicity.
Issues of survey integration are not as critical for the BRFS at this time; it is already consistent in various ways with several other surveys. It is not mounted nationally but collected by states, and meeting state needs is an important feature of the survey. The core of the survey is data that are comparable across states.
Ms. Cynamon described SLAITS as one of the most integrated surveys, at both the sampling frame and questionnaire level. It is based on and integrated into the National Immunization Survey, which provides a huge sampling frame. Using money from the evaluation fund from the Department and CDC, SLAITS developed questionnaires to be administered, using a health module largely based on the HIS.
Speaking from the state perspective, Dr. Friedman expressed strong support for the SLAITS and stressed the value of viewing it and BRFS as complementary systems. Dr. Amaro expressed hope that SLAITS could fill some of the holes in collecting health data on racial and ethnic minority populations.
At Dr. Iezzoni's request, a panel consisting of Dr. Friedman, Dr. Amaro, and Dr. Newacheck discussed the information presented in the morning session and their recommendations for the Committee's response.
Dr. Iezzoni proposed that the Subcommittee develop specific, concrete suggestions for next steps. This topic will not be included in the work plan until options are considered.
The group discussed the shortages in the availability of analytic staff and how (or whether) to call it to the Secretary's attention. They agreed to separate budgetary issues from work-force preparedness. Dr. Friedman, Dr. Newacheck, and Dr. Amaro will work on a letter on the former. The group will continue to look for solutions to the staffing issues, possibly through collaborations with the private sector. Dr. Iezzoni asked for volunteers to take the lead in this area, and Ms. Coltin offered to talk to people on the industry side.
In regard to confidentiality issues related to NHIS design, Dr. Mor offered to write a memo to Kathleen Frawley and the Privacy/Confidentiality Subcommittee suggesting that the issue of privacy be placed on the agenda for attention by the entire Committee.
Dr. Newacheck was asked to track the issue of public release of information from Medicare CAHPS and Medicaid CAHPS.
Follow-up on the islands and territories meeting: a report and recommendations will be developed; a time frame was not determined.
Population-based data: The work group still exists and has staff assigned. Members agreed to identify a work-plan based on the interests of individual members.
Medicaid managed care: Dr. Iezzoni asked Ms. Coltin, Dr. Mor, Dr. Starfield, and Dr. Newacheck to suggest what they see as the end product, and how to proceed to get there.
Post-acute care: Several members offered overviews of the types of work needed to understand the issues in this area. Dr. Iezzoni suggested aiming to produce a document about data policy by the end of 1999, to serve as a conceptual framework for post-acute care ranging across the continuum of types of patients and settings of care. The possibility of engaging a contractor to assist with this project was entertained, along with possible sources of supplemental funding. Dr. Iezzoni, Ms. Coltin, Dr. Mor, and Ms. Rimes will develop a proposal for the October 30 Subcommittee meeting, outlining how they will begin to address the issue and identifying dates for hearings for next winter and spring.
Healthy People 2010: The Subcommittee agreed to prepare a draft letter with comments on Healthy People 2010, for presentation at the November full Committee meeting. Dr. Starfield will focus on socioeconomic factors, and Dr. Amaro will focus on racial and ethnic disparities, and they will both give input to Dr. Carter-Pokras, who will help prepare the comments. Dr. Rouse will flag mental health/substance abuse issues. Ms. Ward was asked to attend the December 2-3 Seattle meeting on Healthy People 2010 data.
Core data elements: The Subcommittee will work on a plan for integrating new work on needed core data elements into the standards development process.
Members discussed how to balance the Committee's agree-upon role as an advisor to NCHS with the need to keep a broad purview in regard to population-based data. The implications of this for the Subcommittee's charge and work plan will be watched in the future, and Dr. Friedman and Dr. Newacheck will work on a recommendation for how to approach it. The group agreed that at a future meeting they should consider adding to the charge the task of monitoring data on disabled people.
Members were asked to get their comments on the charge to Ms. Rimes.
Dr. Iezzoni welcomed the participants to the meeting, the purpose of which is to discuss survey integration. Brief self-introductions were made.
Introducing the first two presentations, Dr. Friedman traced the beginning of public discussion of survey integration at DHHS to the 1992 NAS Report that Dr. Wunderlich will discuss. Since then, there have been few written reports on survey integration. Following her presentation, Ed Hunter, co-author of one article on the subject, will provide an update of subsequent efforts.
After six years, much of the NAS report's discussions, conclusions, and recommendations remain relevant. In re-reading it, Dr. Wunderlich found it a good starting point to consider what still needs to be done to integrate health surveys. When the report was released, its subject matter, findings, and recommendations reflected the context of health care systems of that time and the efforts at developing a comprehensive health insurance plan. Since then, the issues in health care appear to become ever more complex, and the need to slow down the escalating costs of health care even more urgent. This is a subject of continuing debate as the landscape changes with the emergence of new health care providers, new health care settings, new financing arrangements, and new technologies.
Recognizing the need to keep pace with the changing landscape, NCHS developed plans to restructure their system for surveys of providers and service setting data. An integrated national health care survey was developed that built on the existing provider surveys which were merged and expanded over time and geographically linked to the National Health Interview Survey (HIS). Toward that end, NCHS asked the National Research Council of the National Academy of Science (NAS) to undertake two major tasks: (1) to identify the principle current and future needs for health care data by public and private policy makers, health care providers, health service researchers and others; and (2) to determine the extent to which a restructured survey could meet those identified needs.
The report reviewed main features of the proposed survey from two perspectives: (1) the extent to which it would enable NCHS to meet changing data needs associated with health care delivery and policy issues identified by the Committee; and (2) technical features of the integration plan in its developmental stage at that time. However, the report went beyond a simple review of the NCHS plan and suggested a broader strategy for the survey that the Committee believed would more nearly meet the needs of the health care system.
It became clear that no single survey would meet all of the criteria, address all the technical problems, or meet all the users' needs for data. Such a coordinated and integrated system of health care data collection requires the involvement of numerous organizational units. In the committee's judgment, the NCHS plan did not provide the full capacity to address the important questions, but a good step would be to link the survey geographically with the NHIS primary sampling unit (PSU). That, in itself, would not result in the meaningful integrated survey suggested by NCHS, so the committee recommended a design framework and a phased multi-year strategy for implementing the framework for a national health care data system. It also recommended that NCHS establish a research agenda to examine the issues and test the design framework for feasibility, and make other necessary changes.
The panel recognized the great expense in gathering information on medical expenditures, and the need to get current data quickly and cost-effectively. They recommended selecting a sample of respondents from the Health Interview Survey from whom data would be collected on health care use and expenditures--not a new idea, but one that was based on developmental work that had been done a few years earlier. They have been pleased that NCHS and AHCPR have moved in that direction in recent years.
The committee's report noted several important issues not directly addressed in its charge. Some were broadly related to NCHS and went beyond the health care survey; others went beyond the structural issues of collaboration and coordination of data gathering and analysis within the Department. Dr. Wunderlich identified four general areas that must be addressed in developing and implementing an integrated data system: (1) an advisory structure for a health care data system; (2) improving Departmental coordination and collaboration; (3) enhancing the analytical capabilities of NCHS; and (4) augmenting resource requirements.
Believing that both the concept and operations of this kind of data system should undergo external reviews, the Committee recommended:
Mr. Hunter explained that since his appearance before the full Committee about two years ago with Ross Arnett (AHCPR) to discuss the survey integration plan announced as part of the Vice President's government re-invention effort, there has since been little public discussion; but that he and Ross are left to "carry the torch."
As background, he referred to an editorial he co-authored with Mr. Arnett in the American Statistical Association's magazine. He noted that the HHS Survey Integration Plan is still an active document the Department supports as an overall approach to planning surveys. After the failure of the national health care reform effort, the Department and their other partners recognized major data gaps in their ability to look at the impact of health reform, without baseline data for tracking change. He emphasized the need for research to support policy decisions needed throughout the administration and Congress.
Looking for ways to integrate information through the prisms of hundreds of data systems in the Department, they found most of them in "stovepipe" form, and many of them unable to answer the key questions of the policy makers. They found areas where overlap could be reduced or eliminated, linkages accomplished, and redundant efforts consolidated in order to meet high priority needs. After clearing the plan within the Department, they got a senior level buy-in at OMB, in April, 1995. Strategically, they looked for opportunities where surveys could be modified, enhanced, and linked in ways that would meet multiple needs--a major challenge for a set of agencies that have very specific missions, whose budgets go through very defined channels, and whose tradeoffs within each organization require serious consideration.
The Secretary meanwhile formed the Data Council with the charter to coordinate across the Department; to look at survey plans as they are developed within the agencies and to stimulate action within agencies; to look at budget priorities across the Department; and to advise on what the agencies should include in their proposals. The concern in multipurpose efforts is that one agency may capture only a very small part of the benefit while addressing the interests of the Department as a whole. The question, then, is what the Department can do to assure that the right investments are being made.
Since Mr. Hunter and Mr. Arnett co-chaired a survey integration work group that is part of the Data Council, they created a process to develop an advanced plan where agencies are asked to identify their survey and major data systems when their budgets are submitted. Also, the Secretary has assigned Ed Sondik, the Director of NCHS, to a dual role as Senior Advisor to the Secretary. In that role, Dr. Sondik has the opportunity to look across the Department at data planning issues and to stimulate discussion of opportunities for change.
A major change in the 1995 plan was the linkage between the Health Interview Survey, the Medical Expenditure Panel Survey (MEPS), and the National Medical Expenditure Survey (NMES), which was about to be launched as a free-standing, one-time survey. A concern was that the most recent data source from a major expenditure survey was a 1987 survey. Because of resource limitations, the MEPS was planned as a point-in-time cross-sectional study with the option of a one-year follow-up. Since an ongoing basis for looking at medical expenditures was needed, the MEPS was turned into an annual, ongoing survey where people are enrolled in the panel and followed for two years rather than one. The major change was that the sample for MEPS would be drawn from respondents to the HIS rather than from an independent frame.
In addition to sample linkage, there was an attempt to integrate question content for uniformity to allow a crosswalk between those who were interviewed in the HIS and responses to questions later posed in the MEPS. They intended to follow through into other surveys, e.g., the Health and Nutrition Examination Survey (NHANES).
With the major focus on questions at the sample level of MEPS/HIS, they quickly turned around the operations of both surveys and accomplished the link for the 1996 survey-- miraculously, the two agencies, long at odds over such things, worked out a joint plan and made it operational. Now, they are taking a serious look at just how successful it has been from the cost standpoint.
Releasing data to the public is problematic because the ante has been raised on disclosure risk as more information on the same individuals is available. Aside from operational efficiencies, it is important to look also at the analytic payoff from combined data sets that include expenditure data on MEPS and all the non-expenditure data from the HIS.
Another change is that there used to be two separate employer health insurance surveys in the Department, but in 1994, NCHS with AHCPR and HCFA conducted a National Employer Health Insurance Survey with the intent to repeat it periodically. At that time, they were preparing the expenditure survey to include a follow-up of persons interviewed, going to their employers to get more information on the health insurance plans. Although the surveys were similar in content, the operational approach was different. The challenge then was to merge these surveys operationally to permit a crosswalk. The merged surveys are now the insurance component of the MEPS, which includes one frame that was derived from the household respondent and one, done at the state level, from a frame of employers from the Census Bureau.
Another area of change is the National Health and Nutrition Examination Survey (NHANES). The original intent was to use the Health Interview Survey as a sampling platform to launch a variety of household surveys, of which NHANES is one, which will be launched in 1999,using the same sampling areas as the HIS, resources permitting.
They took a longer term agenda to look at the possibility of linking at the person level. NHANES is a very different kind of sampling and operational process: whereas the HIS interviews go into 200 to 300 communities, relatively few people per community are involved because the data are gathered by a single collector going door to door. With NHANES, the plan is to go into a few communities, perhaps 15, and interview as many people as the HIS, using a mobile clinic for greater efficiency. Real issues have to be worked out to find out whether the NHANES person-level approach is feasible. However, the questionnaires in the two surveys are being linked so that the questions asked in the household portion of the NHANES are drawn from the HIS, allowing analytic coordination.
Another relevant area of the IOM report is provider surveys, where the same practical issues occur, e.g., the instances where Department people call on providers for similar or overlapping information, with different forms, different contractors, and with different analytic possibilities.
In the push toward data consolidation, there will be differences in what needs to be asked of different providers, e.g., nursing homes. AHCPR has a nursing home component to the expenditure survey, while NCHS gets periodical information about nursing home staffing patterns and patient loads on a broad sample, but not much detail on expenditures. These two efforts still need to be put together. Similar issues exist in the different surveys of hospitals and emergency rooms, e.g., the broadly based NCHS emergency room survey and the SAMHSA household survey on drug abuse.
The provider problem remains to be solved. This requires setting analytic priorities. A work group in the Department has been working on that for a year or more. They have surveyed a wide number of people within and outside the federal agencies to talk about their data gaps and needs. Now they are trying to pull all those responses together and to specify policy questions and research questions to drive their data systems planning. For that, outside help was sought from people in the major professional associations who conduct surveys in these same areas. Thus, it has become a public/private working and information exchange group. This experience has demonstrated the importance of the topologies and descriptive variables used by different organizations and agencies.
The initial focus on operational issues led to a search for mechanisms for getting information in an integrated way on health care costs, access, insurance coverage, and health status--data not being obtained through existing surveys such as SLAITS and the BRFS (both discussed below).
The possibility of using the HIS as a sampling platform for the SAMHSA household survey presents the complication of expanding the household survey to include state level issues. Another issue is the President's mandate to include a survey on teenage smoking.
Mr. Hunter closed his report with examples of things that have yet to be done:
In response to a question about the broad provider survey, Mr. Hunter replied that no thought had yet been given to building a longitudinal component for, or linking it with, the provider aspect of the MEPS. To a question about the makeup of the public/private surveying effort, he noted a recent meeting hosted by the American Hospital Association, with participation by survey team members from AMA, AAHP, the National Association of Health Data Organizations, Medical Group Management Association, Health Care Financial Management Association, CBO, and state legislatures. He explained that there is no formal announcement of meetings, but an informal word-of-mouth invitation.
Asked to comment on potential advantages/disadvantages of survey integration in practice, Mr. Hunter said they had not really taken advantage of the analytic side, having been so busy getting the surveys fielded; nor have they looked closely at what could be done to analyze the two data sets.
As a public user of these data sets, Dr. Newacheck did not see any demonstrated advantages in terms of analytic capacity, and did not see that it really serves the user community very well. Mr. Hunter agreed about the disadvantage in linking interagency data for public release, a highly problematic area, especially given the complication of disclosure avoidance. Both AHCPR and NCHS are talking about establishing public research data centers modeled on the Census Bureau that use linked files in identifiable but not identified form. Cost, he said, is more of an issue than non-disclosure.
Dr. Wunderlich was asked why it was decided to use a group other than this Committee; she explained that a joint study between the National Research Council and the Institute of Medicine felt that a major review across Departments that would result in recommendations to the Secretary could best be achieved by an independent outside group supported by staff that was independent of all the agencies involved.
Mr. Hunter was asked if there were limitations of national health data systems regarding health status of various ethnic and racial groups in the survey integration process. He acknowledged the need to improve race and ethnicity data and shared the concern about the impact on their ability to estimate small population groups. One option they considered would have curtailed the sample size of the HIS. Another option was not to do the HIS any more and to invest those dollars in a more detailed expenditure survey with a smaller sample that could get more detailed information; however, that would have entailed losses they considered unacceptable in terms of what it would do to population groups. He mentioned that the Data Council has a work group on racial and ethnic data. Dr. Carter-Pokras added that the working group's accomplishments include an inclusion policy for racial and ethnic data that was accepted by the Secretary and circulated last fall to all the agencies. They are now in process of updating the 1995 inventory and HHS people-oriented data systems to find out how well the agencies are implementing the new policy. Many of the agencies have similar working groups. Dr. Iezzoni suggested that she give a fuller briefing to the Subcommittee later at one of their breakout sessions.
The CAHPS effort began in AHCPR with the idea of developing a standardized set of questions that can be asked across populations: commercial, Medicaid, and Medicare. HCFA has been working with AHCPR in developing the three-part Medicare version: (1) survey of Medicare beneficiaries currently enrolled in managed care plans, which is ongoing; (2) survey of disenrollees from managed care plans; and (3) survey of beneficiaries in the fee-for-service sector. The latter two activities, required under the Balanced Budget Act, are still under development. HCFA uses the CAHPS data also to provide information on health plan quality improvement, which is shared with the health plans.
Another purpose served by CAHPS data is in the regulatory function of HCFA-- monitoring health plans. Finally, CAHPS data will be used in reporting for the Government Performance and Results Act.
Much of their effort is spent on the managed care survey which they are getting into the field (see below).
The managed care survey at HCFA (the Medicare Satisfaction Survey) includes core questions supplemented with items relating to ease of getting special equipment and therapy, ways of getting home health services for chronic conditions, and activities of daily living. That survey now contains about 85 questions. Major components are overall ratings of (1) the experience with the health plan; (2) health care received; (3) doctors and other health professionals in the plan. The ratings are on a scale from zero to ten, with zero representing the worst.
For reporting survey information, the team developed a format called CAHPS composites, which combines three or four questions relevant to the same area. The first round went out in February 1998, and the second round is now in the field. The first round covered 600 non-institutionalized beneficiaries per plan randomly selected from a sample of over 136,000 Medicare beneficiaries. From this first round, they got a 74 percent response rate. The survey will hereafter be conducted annually.
Dr. Goldstein briefly described the steps involved in conducting the survey: pre- notification letter to the beneficiaries in the sample; postcard that served as thank you to respondents and reminder to others; second mail survey; telephone follow-up to non- respondents, or for those without telephone an overnight Fed Ex package, with a third survey. Her slides illustrated their two basic formats for presenting the information: bar graphs or a rating system using stars. The bar graphs were found to be more understandable. This year this information will be shown on their Website (Medicare.gov) as Medicare Compare, and will present data from both CAHPS and HEDIS. They also will provide extensive reports to health plans about survey findings to assist in quality improvement efforts.
The two closely-paced rounds were necessary, she explained, to meet the deadline for a fall report to consumers. Each fall they distribute a handbooks, "Medicare and You" that provides comparative information from the CAHPS data.
In the second round of Medicare CAHPS, the part now based on a survey by NCQA and AHCPR will be merged with the CAHPS survey. Also changed is the eligibility period-- from 12 months to 6. In developing a common instrument, AHCPR and NCQA decided to use the strongest items from the original CAHPS survey and the member satisfaction survey, so that the new survey would reflect input from all users. One issue they had to confront, especially in the HEDIS member satisfaction survey, were trend issues--what would be critical information to collect in future years? It was also necessary to keep the survey as short as possible. As of January 1, 1999 they expect everyone to be using this instrument, which will enable them to compare information across populations and plans.
For the second round they included plans that had contracts in effect as of January, 1997. An evaluation of the CAHPS survey and reporting formats are being done in Kansas City. Before turning over the presentation to Mr. Sekscenski, she briefly mentioned the disenrollment survey now in inter-agency development that is designed to ascertain reasons why beneficiaries disenroll from plans.
Speaking on the fee-for-service survey in Medicare, Mr. Sekscenski said that the timeline is about 18 to 24 months behind that of the managed care survey. A pilot test is in the field in five states (Arizona, Georgia, Pennsylvania, Washington, and Wisconsin), and responses are trickling in. In two of the states (Georgia and Pennsylvania) an overlap sample with the Health of Seniors Survey will provide information for case mix purposes for the CAHPS.
The pilot will (1) assess the effectiveness of survey instruments; (2) compare alternative items used in version A and Version B; (3) collect additional information on case mix, health status, and use of services for case mix adjustment beyond what we will receive form the Health of Seniors Survey; (4) and provide information for developing the sampling strategy for the Medicare fee for service CAHPS. There are two different instruments in the pilot: an 80- question survey in all five states that uses a 6-month recall period or a 12-month recall period to see which obtains the most information. They will probably go with the 6-month recall period in the final CAHPS that will be implemented in the fall of 2000. For the total size of the sample--around 3200 in the three non-overlapping states--they are aiming for 70 percent response rates in both of the pilot instrument surveys. In 1999 and 2000, they will be moving toward analysis of the information from the five-state pilot, and will be looking at reporting mechanisms similar to those for managed care.
Various approaches are being developed for providing information on health and fee- for-services plan within a geographical area to both intermediaries and beneficiaries. Decisions
about the type of reporting units to be used in the fee-for-service survey will be made with input from stakeholders and experts in the Medicare fee-for-service sector and the private sector.
Dr. Hardy furnished additional details about using the Health Interview Survey as a sampling frame to link other surveys: the National Survey of Family Growth (NSFG), the Medical Expenditure Survey (MEPS) and the National Health and Nutrition Examination Survey (NHANES). Because of the anonymous feature of the National Household Survey on Drug Abuse, they are not presently considering that link.
There are various levels for linkage, but the most common link is the respondent level, where they try to locate the same people who were in the original HIS or linked survey. Because the NHANES has a physical exam component, it is being linked at the primary sampling unit (PSU)--and though it will go into the same areas, will not get the same households or respondents.
The current HIS sample, which went into the field in 1995, will use data from the 2000 census to field a new design in 2004. The current design oversamples black and Hispanic populations, and is now in every state. It has more primary sampling units that are more spread out, which leads to some limitation in linking surveys. They are starting to gear up with the Census Bureau and other statistical agencies to plan the next design, which will be done with survey integration in mind.
The NSFG, a reproductive health survey targeted at women 15 to 44 years of age, was the first survey linked to the HIS in 1998, using the 1987 HIS sample; their cycle five sample taken in 1995 was based on one 1993 HIS sample.
Advantages of survey integration include cost avoidance in having to do the listing and screening; also NSFG targets women of a certain age group and oversample minority groups. This makes a linked analysis possible. Although they have not yet linked MEPS and HIS, the NSFG data file in 1995 includes about 75 items from the HIS; and they have worked through some of the confidentiality issues.
There is hope for improved analytic capabilities as well. Although confidentiality issues have to be sorted out, especially because MEPS has so much detail, there is the potential for disclosure that must be looked at very carefully. They are still trying to find ways that both the federal research community and those outside can take advantage of the link between the two surveys.
There are also disadvantages, e.g., size limitation and composition of the HIS sample. However, when more than the sample provided in a given year is needed, a couple of years can be linked, although this has some disadvantages. It is harder and more costly to trace respondents than was originally anticipated, particularly with the last round of the NSFG, because of the two- year time lapse. Moreover, the target group of NSFG is an age range that is more like to move often. Ultimately they were able to locate about 95 percent of the people, of which they got about 79 percent response rate, compared with the 83 percent response rate of the MEPS link to HIS.
Because of the physical exam component of NHANES, they are linking only at the PSU level. In 1999 they will be going to the 15 PSUs that are included in the HIS.
Dr. Hardy then discussed briefly what linkage has meant to the HIS side, starting with the necessity to be in the field every year in January as scheduled. HIS has gone from a paper and pencil survey to a computer assisted personal interview (CAPI) medium, with potential to improve both data quality and timeliness. There are also potential advantages in terms of linked surveys. With CAPI, they are getting data regularly within six to eight weeks after data are collected in the field. Because they want to be in the field every year, they have built in a longer lead time to make sure the CAPI system is functioning, but this limits flexibility in being able to make questionnaire changes at the last minute. Because of the many changes in health insurance and welfare, they have changed the questions at mid-year, which may present problems with data processing. And because the HIS sample is limited in size, there will be competition for samples. If survey integration catches on and more surveys attempt linkage, HIS will not want to overuse their respondents. So, they have built in the capacity for extra samples that could be accessed by census.
She closed with comments about methodological issues. The next redesign of the HIS will be based on the 2000 census, and it is critical that survey integration be given full consideration; that they look more closely at how efficiently they oversample the black and Hispanic populations; that they assess new sources of survey error, e.g., problems with non- coverage rates of young black males; and that they find ways to improve collection recontact information and look at other types of information needed for followup of respondents in linked surveys. They will be assisted through a cooperative agreement with the University of Michigan to conduct evaluative activities and related tasks in linked surveys to gather data that can be fed back into the system to make it function better.
Asked to further explain case mix adjustment, Dr. Reilly identified two issues: (1) comparisons across fee-for-service and managed care to help beneficiaries choose between the two systems in terms of different socio-demographics and other characteristics not necessarily related to the care; and (2) variability due to things not deemed relevant to care but related to reporting. They have done a fair amount of analysis this past year on the managed care side in terms of education, self-reported health status, and whether a proxy was used in helping the person respond.
Regarding the CAHPS and questions across groups that pertain to gender, race and ethnicity, the question was raised whether the instrument is stratified enough to assure a large enough sample across language groups. Dr. Reilly said the survey is representative of people enrolled in managed care plans, and that it will be difficult to stratify and produce estimates by associated demographics. But samples are large enough to produce "plannable" estimates. In terms of language issues, they are tracking how many people would require Spanish questionnaires and are recording whether a proxy was used if the questionnaire was in English. In the case of telephone follow-up, they would have information as to whether the survey could not be completed because of a language barrier.
Asked if they plan to evaluate the impact of the CAHPS on health plans, Dr. Reilly said that evaluation activities are going on, e.g., a demonstration project in Kansas is looking at the effect on the Medicare population. There are also evaluations of commercial populations with Medicaid, but they are looking more at the effect on beneficiaries than on the health plans. They are routinely in touch with health plans through focus groups to determine what information is wanted from the CAHPS--this is more a continual quality improvement kind of assessment than a formalized impact evaluation on the plans. Asked whether they would assess whether impacted enrollments are going up or down because of publicity over the survey results, Dr. Reilly said they plan to track that. As to whether the Medicare and Medicaid CAHPS data would be made available as public use data sets, he replied that they are providing summary data back to the beneficiaries to help them choose among plans. The health plans are interested in getting the data back for their own quality improvement purposes. For the health plans, they will provide a fairly detailed report about their plans and other plans in their area for comparison. But making person-level data available is a very sensitive topic around HCFA, and they will probably have to set up some kind of protocol or mechanism to handle requests for certain kinds of data for research purposes.
David Brown, Center for Mental Health Services, representing Ron Manderscheid, reported that his agency is working on a behavioral health component of the CAHPS.
Unlike the other surveys, the Behavioral Risk Factor Survey (BRFS) does not have a central collection. They are in cooperative agreements with states and the issue of survey integration is less applicable at this time than for other surveys. Many of the questions on the BRFS come from existing surveys and are identical to those in the HIS. Coding categories are also identical, although the intent is not the same. Another point is that CDC, in general, has made an effort to standardize core elements of their data collection, and BRFS conforms to those, particularly as they pertain to the collection of race and ethnic information.
Dr. Powell-Griner reviewed the development of BRFS. It began in the early 1980s as a point-in-time survey, but by 1984 it had moved into a surveillance system operation in 15 states. It has grown from 6 initial topics to 15 in 1998. The number of core questions (the part of the survey not modified by any state using the BRFS) has grown from 47 to 80. Initially, there was no provision for optional modules directed largely at chronic disease and behavioral health, but now they are supporting 22 modules. States may choose to use as many modules as time permits; but if they use a module, they are not allowed to modify the wording or order of questions.
An essential part of the survey that sets BRFS apart is a large section reserved for questions states are interested in. No attempt is made to standardize those questions. Often the questions that the states select are driven by programs that are helping them fund their survey.
She describing the BRFS as a "David, not Goliath" effort, with a very small staff and extremely limited budget (with a line item budget), that has to go out each year "hat in hand" to collect monies to conduct the survey. They have moved away from the survey design based on random digit dialing.
The file has increased from 12,000 interviews per year to over 135,000. For states, they range somewhere between 1500 and 5,000. Another development is the greater use of geographic stratification, with states collecting data at the county level. There would be more interest in doing local area estimates if adequate funding were available to get information at the county and city level.
A goal of BRFS is to have a set of data coming from the core items, which are comparable across all the states. A second goal is to make sure that state needs are being met.
She identified several strengths of BRFS:
She also briefly noted caveats:
Dr. Powell-Griner closed with comments on good uses of BRFS and why it is important to maintain it as a viable data system.
Ms. Cynamon described SLAITS as one of the most integrated, at both the sampling frame and questionnaire level. The survey is capable of small area estimates as well as state level. It is administered by telephone, and its primary goal is to help fill data gaps about changes in the health care and welfare systems at various levels, over time. It is also integrated into the National Immunization Survey (NIS), a list-assisted, random digit dialing survey that has existed since 1994. Its main focus is to provide state and local estimates on a quarterly basis of the immunization status of children aged 19 to 35 months. All 50 states are represented, as well as 28 metropolitan areas, including the District of Columbia.
NIS provides an enormous sampling frame: children aged 19 to 35 months represent five percent of the population, so a great number of telephone calls have to be made--2.4 million a year. The automated dialing system rules out non-working phone numbers, and ends up with 1.6 million. Those are called and screened to determine whether they are business and whether there is more than one line per household; this brings the number down to under a million households screened for children in that age group. The NIS interviews 36,000 of those households annually, but without pursuing any sort of health data--and this is where SLAITS steps in.
SLAITS, using money from the evaluation fund from the Department and CDC, developed questionnaires to be administered, using a health module largely based on the HIS that covers the basic areas of access, utilization, some unmet needs, and health insurance coverage. It is 22 minutes in length. This module has been tested in two states.
They also developed a module on child welfare and well-being. The welfare module is drawn from other surveys, and has a Medicaid status matching component linked to state-based Medicaid records that can assess people's ability to accurately report their Medicaid status.
They have developed and distributed for comment a questionnaire on children's health insurance coverage and related health insurance issues, e.g., access, barriers, health status, and health care utilization. This was designed to allow states to monitor the uninsured and changes in their status.
SLAITS makes it possible for states to customize questionnaires to accommodate their own interests. They are now discussing with CDC implementation of a module on behaviors related to HIV risk that goes into more depth than the BRFS. They are also looking at ways to use SLAITS to produce state estimates of asthma prevalence.
Some of the features of SLAITS:
Dan Friedman expressed the enthusiasm of the state of Massachusetts about SLAITS and the prospect of being a pilot site, although the flexibility of BRFS makes that the core of their public health decision making. SLAITS is in an ideal position to survey populations that BRFS is not set up to deal with, e.g., children and low-income persons. Viewing the two as complementary rather than competing systems, he suggests that more work be done to integrate them to the extent possible--not duplicating efforts but adding to one good telephone data system. He noted that BRFS has a small budget, under $7 million, of which more than half goes directly to the states in grants.
Ms. Cynamon agreed, and stated that just as there is more than one in-person health survey, there can be more than one telephone health survey. It would require a complete revision of the approach to households if BRFS were to do what is being proposed using SLAITS.
Ms. Cynamon was asked how quickly SLAITS could be operational to monitor the State Children's Health Insurance Program (SCHIP), given a fast-track procedure for OMB clearance of telephone questionnaires--that is, if money were available. She said that under those conditions, it would take three months, at most.
Other health surveys are inadequate to reach all of the racial and ethnic minority populations, which are not evenly distributed throughout the country, so typical sampling strategies do not work well, Dr. Amaro expressed hope that SLAITS would be able to fill some of the holes in collecting health data for those groups. Asked whether the different samples were representative enough in the field tests to sustain hopes for this survey, Ms. Cynamon said that the health module mirrors the makeup of the general population. The large sample size for NIS makes it possible to screen at the household level for minorities. She noted that most of the interest in SLAITS has been for rarer populations, whether uninsured children or people with asthma, which takes a lot of screening and is expensive. Though a lot more expensive than BRFS, it has the capability to screen for minority status. Possible use of the GENISYS data file in combination with census track data will make it feasible to identify minorities in geographic areas, thus making the sampling of minorities more efficient.
Asked whether SLAITS will be able to provide a national estimate of SCHIP enrollees, Ms. Cynamon said that, if funded, they could and would. They have a questionnaire that monitors children's health insurance status. But even with their large sample size in each state, SCHIP enrollment is still going to be a very small proportion of the population. What SLAITS can provide that enrollment data cannot is the dynamics of the uninsured, which is the target population.
Asked whether CDC plans to include more mental health in the core BRFS questions (they now have only one question related to the mental health status), Dr. Powell-Griner explained the social context of the module that looks at a number of items related to mental health. Programs can develop modules for inclusion in the BRFS. She reviewed the process by which questions are presented at their annual conference, then cognitively tested to make sure they can get useful information. (This year, for the first time, they are asking programs that want questions in modules to provide some kind of financial support.) Another option is to work with the states to put their questions into a state-added section, e.g., a disability special study being done this year and next.
Dr. Iezzoni asked the subcommittee members to focus on information presented in the morning session and what the Committee needs to do in response to that information and the panel's questions and discussion. She asked for specific suggestions.
The following ideas were proposed:
Dr. Iezzoni recognized Mr. Handler of the Indian Heath Service, who said he had been waiting 24 years to make his comment. He observed that given the minute ratio of the Indian population to the total population, "No matter how you oversample, you will not be able to capture all of those American Indian people." He recalled that 10 years ago IHS organized a patient registration system requiring everyone who got care to register--a perfect sampling frame of 1.3 million people with names, addresses, and telephone numbers (if any). He thought his agency would make that automated list available as a sampling frame. Dr. Iezzoni suggested that he bring that offer to the attention of the HHS Data Council.
Returning to a possible resolution of the "case of SLAITS versus BRFS," Dr. Iezzoni proposed that the Subcommittee focus on specific and concrete suggestions for next steps. The following suggestions were made:
On a related topic, Ms. Greenberg asked about the value of a letter to the Secretary expressing concern over budget cuts that affect the adequacy of analytic staff. She noted the combination of budgetary and recruiting and training problems that impact at the federal and state levels. She stressed the importance of training to enhance staff analytic capacity, given the higher paying positions available in the private sector for both epidemiologists and mathematical statisticians.
Dr. Friedman suggested separation of the two issues -- work force preparedness and NCHS budget -- and that the Committee take responsibility for making a statement about the latter. He offered to take the lead in drafting a letter, and Ms. Greenberg suggested he talk with Dr. Sondik. Dr. Newacheck and Dr. Amaro will also give input. Dr. Amaro urged that the Committee address the budget issues as a sine qua non for making progress, especially with respect to racial and ethnic minorities.
Some asserted that training analysts and building capacity to learn from the health data produced is an issue the Subcommittee should continue to pursue, even if it does not propose specific solutions at this point. Apparently the difficulty in recruiting doctoral level or well- trained masters level biostatisticians has not been formally presented to the Data Council-- although it is on the radar screen of individual members.
Further discussion ensued: Dr. Starfield emphasized that merely raising the issue is not constructive enough; Ms. Ward agreed on the need to think about a broader direction to move- -possibly toward alliances with the commercial side of the health care industry, which would also benefit. Dr. Iezzoni reminded them that the Subcommittee's agenda usually is set by advocates (members) who are willing to spend considerable time on the agenda item under consideration. She asked who would assume that role -- e.g., recruiting a panel to talk about it and writing a brief, 5-page report based on their insights and suggestions. Ms. Coltin agreed to work with someone on the industry side. Dr. Mor suggested calling in three or four people from schools of public health and health administration as well as health program management to address this broad issue. The consensus was that such a panel would require only half a day. Ms. Coltin insisted on the importance of getting the right mix of people on the panel. Given the HIPAA implications, Dr. Iezzoni suggested it as a full-committee issue, and it be discussed with Dr. Detmer and the executive subcommittee. Dr. Amaro suggested they make sure of representation from programs on training racial and ethnic minorities in biomedical careers.
Another issue noted was the need for strategies in the area of confidentiality and complications of data release. Dr. Mor offered to write a memo to Kathleen Frawley and the Privacy and Confidentiality Subcommittee suggesting that the issue of privacy and confidentiality be placed on the agenda for the Subcommittee to deal with. Discussions also suggested the need for full Committee involvement.
On public release of information from Medicare CAHPS and Medicaid CAHPS: Given the importance of the data sets for researchers and the need for evaluations to know how well these populations are being served, Dr. Newacheck suggested that the subject be put on the agenda in a couple of months.
HCFA staff will not determine whether Medicare information is released; it will be decided state by state. Only a few states have started the Medicaid CAHPS. To the question of how states fund the Medicaid CAHPS, Ms. Coltin said the process is in transition. HCFA paid for the first version; her sense is that eventually HCFA will expect health plans to pay, and state Medicaid agencies are likely to do the same, so the health plans will end up paying the bill. The question is who owns the data, once it is collected. Dr. Newacheck's point was that it should be in the public domain as long as it does not violate any confidentiality or proprietary purposes. Ms. Coltin raised the issue of what kinds of privacy protection, if any, should be afforded in moving beyond individuals to organizations. She asked why the confidentiality of managed care organizations is being protected, and how to draw the line between protecting individuals, physicians and managed care organizations. She cited Pennsylvania and New York as examples of differences in ways states handle the issue. Dr. Iezzoni asked Dr. Newacheck to track that issue.
Redesign of the HIS sample based on the 2000 census: A question was raised about the role of the Subcommittee vis a vis the inevitable conflict or competition between geographic areas and populations. The Committee might want to engage in a process to get adequate resources to support the type of research needed to develop and evaluate the current design and implications for the future direction. Dr. Iezzoni said this issue would be followed as they have followed the implementation of the OMB mandate. Both should be on the agenda(s) of future breakout sessions.
Dr. Iezzoni asked for specific ideas on the table, to get a sense of commitment from individual committee members and to provide lead time for staff to plan budgets and hearings.
It was agreed that the Subcommittee should move ahead on a draft report on the Islands and Territories meeting, but specific plans were not made for this process.
Dr. Friedman explained that because it is a joint effort of the Subcommittee, NCHS and CNSTAT, this is a different model for committee functioning, with its own work plan for 1999.
To a query about whether the work group is still functioning as a work group, Ms. Greenberg said that it is identified as a work group on the organizational structure, with members Friedman, Newacheck, and Starfield and with Gerry Hendershot as staff to the work group. Dr. Starfield suggested that members of that work group carve out specific pieces as areas of concentration: for herself, she expressed interest in the role of core data elements and the vision for health statistics, as well as the health objectives for the nation.
Dr. Iezzoni reviewed the fact that the Subcommittee has spent a year thinking and talking about this issue while also monitoring several other issues including the OMB Directive 15 and data on the Islands and Territories. Dr. Iezzoni added that it differs from their typical Committee processes also because of fortunate funding for the George Washington University subcontract. She noted that a similar arrangement may be needed to deal with post-acute care. The Subcommittee has not really figured out what product it wants. She asked Ms. Coltin, Dr. Mor, Dr. Starfield, and Dr. Newacheck to suggest what they see as the end product, and how to proceed to get there.
It was noted that post-acute care requires a concerted effort; it is an important issue in both private and public sectors, and is increasingly important (along with related issues of monitoring, licensure) in the states. Dr. Friedman emphasized the need to get this issue on the agenda in view of the necessary commitment of time by subcommittee members.
Ms. Coltin expressed interest in working on issues cutting across acute and post-acute settings, and monitoring quality of care across those settings. The science is not as well- developed in those settings as in managed care plans and hospitals, although neither of the latter is where it needs to be. There is not as much consensus about the way to go, what tools to use, in post-acute settings as in the other settings; so the Committee needs to push the industry toward consensus. She identified as a real problem the fact that they do not know what mix of care in what settings achieves the best outcomes for patients, since so few studies have been done looking at patients with particular conditions who have gone into rehab hospitals versus skilled nursing facilities.
Dr. Iezzoni asked her to outline specific steps that would serve as a work plan for the subcommittee, with the output specified, and Ms. Coltin noted the need to better understand the following issues, among others:
Dr. Mor then outlined the following tasks:
Dr. Iezzoni recalled that HCFA talked about stepping back from the silos of Congressionally mandated data collection to develop a conceptual framework for thinking about data for various users.
Regarding the time frame, Dr. Mor said this project would take longer than a year to unfold. Dr. Iezzoni suggested aiming to produce a document about data policy by the end of 1999 -- not necessarily a specific core data set, but a conceptual framework for post-acute care that would range across the continuum from 500-gram premature babies who require extended care to the very old. It also would consider all the different settings, the practical issues raised by Dr. Mor, and the conceptual and operational issues involved in getting the data.
To Dr. Starfield's point that they may not be able to specify how to measure each of the core data elements yet, Dr. Iezzoni said they could first figure out what various stakeholders need and the gaps, and they might get to the point where they can specify data elements, as they did for Medicaid managed care. Ms. Coltin noted the similarity with what has been done about surveys: identify a set of data needs; understand where they are currently and where the gaps are; look at the options for filling the gaps; and then decide what to recommend among the options.
Noting that this was shaping up as a plan, Dr. Iezzoni added that they might need to ask states about Medicaid and disabled children, and that minority issues would be key. Ms. Greenberg observed that it is not just a HCFA issue and involves the private sector.
Dr. Iezzoni asked if the subcommittee saw a need for an outside contractor to write a report synthesizing the information, possibly doing a survey or compendium of all the relevant federal surveys. Members felt it was an idea to consider at a later juncture. Ms. Greenberg mentioned the possibility that ASPE could get money for limited research projects and that individual agencies could give additional help for this project.
One of the uncharted areas, Dr. Iezzoni said, is the role of the private sector in paying for post-acute care, and whether the issue is one solely for state or federal responsibility. It would be valuable to find out more about the roles, policies, costs, and activities of the private sector in paying for this. Under the BBA, within a couple of years managed care plans will have to submit full encounter data.
Dr. Iezzoni suggested that she, Ms. Coltin, Dr. Mor, and Ms. Rimes work together to present a proposal at the October 30 Subcommittee meeting, outlining how they will begin to address the issue of post-acute care. They also will identify possible dates for hearings for next winter and spring. Meanwhile, the group should be thinking about the need for site visits.
Ms. Greenberg asked whether the quality work group also will focus on this issue. Ms. Coltin reminded her that there are now two projects on quality: one on Medicaid managed care for vulnerable populations, and one on quality of care in post-acute settings. She suggested that a distillation of what they learned in the first effort could serve as a road map to direct their work on post-acute care.
Dr. Amaro reiterated her strong opinions on survey data needs and gaps in data for racial and ethnic minority populations, SCHIP, states, and the disabled. She stated the interests that brought her onto the Committee: to reinforce its role in addressing those gaps in the health data system in the development of integrated surveys; to promote the inclusion of territories; and to integrate mental health and substance abuse into the Committee's considerations. The issue must again be addressed, in light of the President's initiative on eliminating racial disparities, even though they don't know how they will measure eliminating racial disparities when they do not have data on the things they want to eliminate. Dr. Iezzoni reminded her that the June breakout session, in which the Census and OMB people talked about the implications of the new classification on longitudinal tracking of experiences of racial and ethnic minorities, was a good start, and that suggestions for specific and substantive next steps would be welcome.
Dr. Starfield suggested as a good venue the Healthy People 2010 Objectives for the Nation, which addresses inequities and has opportunities for public comment through December 15. Dr. Iezzoni suggesting setting aside a few hours for the full Committee to provide commentary. Ms. Greenberg felt the work groups and subcommittees could take the time to discuss recommendations and then remind the full committee of the issue.
Dr. Carter-Pokras pointed out the deadlines for written comments and the dates of five regional meetings, which start in October. A big data session in Seattle (Dec. 2-3) will have a session on eliminating disparity, which Dr. Sondik will attend. She suggested timing the Subcommittee report to take advantage of what they learn after the regional meeting but before the deadline for public comment.
There was discussion of whether there is time to involve the full Committee, since the Medicaid Managed Care Report also must be addressed at the November meeting. Dr. Iezzoni suggested they have a draft letter at the November meeting for Dr. Detmer's signature, seeking full committee approval. It would be nice to have the report of the work groups just before discussion of the letter. Dr. Carter-Pokras agreed to work with Ms. Greenberg, Jack Anderson, and others on a report and draft letter. It was suggested that two or three members work informally with Dr. Carter-Pokras to identify issues to bring to the Committee's attention.
Dr. Carter-Pokras mentioned that their agency has already drafted questions and sent out a 12-page newsletter on Healthy People to over 16,000 individuals and organizations. They have six work groups, plus the data work group for the Eliminating Disparities Initiative, and there is a definite overlap with the Healthy People 2010 goal. All have been asked to look at the data on the national, state and local levels. She pointed out areas of gaps and omissions in data. Dr. Amaro felt that the issue should be integrated into the general work of the committee.
Since the document (Healthy People) is so massive, and the committee members have so little time, Dr. Iezzoni asked for suggestions for tackling it. Dr. Starfield suggested that she work on the socioeconomic disparities, and that she and Dr. Amaro could flag important parts that the committee could respond to in its statement. They will have a draft ready for the November meeting to get approval to finalize it in time for the December 15 deadline. Ms. Rimes was asked to coordinate their comments and pass them along to the new staff person, Sheila Merriweather. Dr. Carter-Pokras will help make sure that the finalized statement goes through the appropriate formal channels.
Dr. Amaro asked whether the Healthy People 2010 mental health and substance abuse goals are satisfactory to SAMHSA. Dr. Rouse said there are problems in some areas, but in others the data needs are addressed. Dr. Rouse consented to work on that area for the Subcommittee.
Ms. Greenberg asked if there is a mental health/substance abuse component to the post- acute issue. Dr. Newacheck and Dr. Mor discussed the complicating factors in MH/SA treatment: risk adjustment and social factors related to discharge for premature infants and other complications associated with children with disabilities. (Probably 85 percent of the children with disabilities in institutional settings are there for mental health rather than physical reasons.)
Dr. Carter-Pokras said that all the work groups responsible for drafting objectives for Healthy People 2010 were asked to break out data as much as possible by racial/ethnic minority, at least in the categories of OMB Directive 15; in addition, they were asked for measures of socioeconomic status for the disabled, women, and some age groups. The socioeconomic status indicators presented even more problems than the racial and ethnic data.
Dr. Iezzoni raised a question about the Subcommittee's role in working on new core data elements that were not ready for inclusion in the last set of recommendations. This unfinished business is technically the domain of the Data Standards and Security Subcommittee. Dr. Iezzoni pointed out that in the two years that have elapsed there have been changes on the Committee and the issue and process needs to be talked through at greater length. She suggested that the Subcommittee think of an operational plan for integrating it with the standards process.
Members discussed the need to clarify the Subcommittee's role as an advisory body to NCHS in population-based issues. Some saw the advisory role as appropriate given the work on the visioning process; others regarded tieing the role to NCHS as too limiting, since other agencies also are producing population-based information. Ms. Greenberg noted that Dr. Sondik asked the National Committee to serve as an advisor, particularly in respect to visioning data for the 21st Century, and this commitment by the Committee must be considered. It was noted that one idea had been to set up a completely different advisory committee, which does not seem necessary. Mr. Hitchcock pointed out that the Committee should avoid being perceived as an NCHS committee, as happened in the past. Dr. Friedman voiced concern about any statement that would give the impression that the Committee's role vis-a-vis population-based data would be limited to NCHS, even though NCHS is an important part. The consensus was to defer action until they have a very clear work plan, and that they monitor this periodically to see if specific activity is required that needs to be placed in the charge. Dr. Newacheck and Dr. Friedman will discuss further how this issue can best be handled.
Dr. Newacheck suggested revising the charge in regard to data needs for disabled persons. His suggestion was considered important, especially in view of new approaches by the Social Security Administration to evaluating disability. It was agreed that the Subcommittee should be monitoring this. The topic was deferred for discussion at a future meeting.
Ms. Rimes will see that everybody is asked to get their comments back to her regarding the revised charge.
Ms. Ward was asked to attend the Seattle meeting on Healthy People 2010. Dr. Iezzoni then adjourned the meeting.
I hereby certify that, to the best of my knowledge, the foregoing summary of minutes is accurate and complete.
/s/ Lisa I. Iezzoni January 21, 1999
Lisa I. Iezzoni, Chair Date