Room 505-A
Hubert H. Humphrey Building
200
Independence Avenue, S.W.
Washington, D.C. 20201
Opening Remarks and Introduction
Current Status of Survey Integration
Subcommittee Deliberations and Work Plan Discussion
DR. IEZONNI: Thank you, everybody, for coming. This is a meeting of the Subcommittee on Populations of the National Committee on Vital and Health Statistics. We are here today to talk primarily about survey integration in the morning session, and then in the afternoon session, we will be doing some subcommittee work, hopefully thinking about our work plan, how we are going to get everything done that we have on our plate, and then adding to our plate for the future. I know that Elizabeth needs to leave at two, and so we might take a slightly shorter lunch and try to get out earlier if we can.
This morning, we have several panelists who have come to speak to us about survey integration. Let's start with going around the room so that everybody can get to know who each other is.
I am Lisa Iezzoni. I am Professor of Medicine at Harvard Medical School and at the Division of General Medicine and Primary Care at Beth Israel Deaconist Medical Center in Boston.
DR. MOR: I am Vince Mor. I am a Professor at Brown University Medical School and Chair of the Department of Community Health there.
DR. NEWACHECK: I am Paul Newacheck. I am a faculty member at the University of California at San Francisco.
DR. STARFIELD: I am Barbara Starfield, University Distinguished Professor at Johns Hopkins University.
MS. WARD: Elizabeth Ward, Assistant Administrator for Washington State Department of Health.
DR. ROUSE: I am Beatrice Rouse, Senior Epidemiologist at Substance Abuse and Mental Health Services Administration.
MR. HITCHCOCK: I am Dale Hitchcock from the HHS Data Policy Office, and I am a staff member for the committee.
DR. WUNDERLICH: Gooloo Wunderlich, Institute of Medicine of the National Academy of Sciences.
DR. HUNTER: Ed Hunter, Associate Director for Planning and Legislation at NCHS, and I am also Ross Arnet for today. Ross has had some real back problems and asked me to double head. We often tag team this presentation anyhow, and he is not really wanting to be sitting in chairs right now.
MR. HANDLER: Aaron Handler. I am with the Indian Health Service Demographics Statistics Staff. I have been my agency's OMB information collection/clearance officer about 15 years and its Privacy Act coordinator for the same time.
MR. SUMMY: I am Jim Summy with AHCPR. I am staff to the subcommittee.
DR. GREENBERG: I am Marjorie Greenberg with the National Center for Health Statistics. I have to go meet with the public health bunch in the work group, and then I will be back.
DR. DETMER: I am Don Detmer, University Professor at the University of Virginia, and I chair the National Committee.
MS. RIMES: Carolyn Rimes, HCFA. I am staff to the committee.
MS. COLTIN: I am Kathy Coltin, Director of Management System Development at Harvard Pilgrim Health Care in Boston and a member of the committee.
DR. FRIEDMAN: Dan Friedman. I am Assistant Commissioner with the Massachusetts Department of Public Health and also a member of the Committee.
DR. IEZZONI: Thank you. Marcie?
DR. CYNAMON: [Off microphone.] Marcie Cynamon, co-chair of the SLAITS Project, National Center for Health Statistics.
MS. RIMER: Carol Rimer. I am section chief of [Commenting off microphone].
MS. GOSTON: [Off microphone.] Deborah Goston, Statistician with the National Institute on Alcohol Abuse and Alcoholism.
MR. SYMPOS: [Off microphone.] Chris Sympos, NIH Office of Research on Minorities.
MS. FOSS: [Off microphone.] Betty Foss with the OPHS Office of Minority Health.
DR. IEZZONI: Sorry. Vince and I were having a slight melt down about timing. This committee operates under the Freedom of Information Act and Federal Advisory Committee Act provisions, so anybody who wants documents, let us know. We also are on the internet today, so anybody who has comments from the audience and people around the table should try to use the microphones. If you are responding to a document, you might want to just say specifically what it includes. I would like to have Stan just maybe make 30 seconds worth of comments about why we are interested in being here today.
DR. FRIEDMAN: Thanks, Lisa. Thirty seconds sounds about right. I just want to start my 30 seconds, or 60 for that matter, by thanking Carolyn for putting together what I think was really an excellent session.
The public discussion of survey integration at HHS started with the 1992 IOM Report that I think all of you have received. It is really terrific that Dr. Wunderlich is here today to start the session. As far as I know, since then, with the exception of a couple of articles by Ed Hunter and Ross Arnet, and we are lucky that we have both of them here representing us today. Since the only public discussion I have seen has been by Ross and Ed, so we are lucky that they are here.
When we move on to discussion of caps and HIS, which is somewhat clear, relatively clear waters in terms of integration, and then at the end of the morning, we are going to move into the BRFS and SLAITS and talk about integration and so forth there. So, with that, I will turn it back to Lisa.
DR. IEZZONI: Thank you. Gooloo, do you want to lead us off?
DR. WUNDERLICH: Thank you. I am very pleased to be here. As Don Detmer remarked a while back --
DR. IEZZONI: Could you pull the mike a little bit closer to you?
DR. WUNDERLICH: I said I am very pleased to be here. It brings back memories. As Don Detmer remarked earlier, how does it feel to get back here. It feels like homecoming in a way when I see my old friends here.
When I was asked by Carolyn to come and talk to you about the Academy's report towards the National Health Care Survey, I did get a little concerned, to say the least. It is now more than six years since that report was issued, and I have been gone from here for more than ten years. That is the frightening part. I thought, my god, I do not remember anything. I reread the report yesterday.
Although it is more than six years since the report was released, in rereading it, I was amazed how relevant some of the discussions, conclusions and recommendations are even today. Although some in this room, some were not in this room, feel that the report is outdated now, what with managed care and other organizations emerging and all of the new data systems and surveys established, others believe that it should serve as a starting point for examining what has been done since then based on some of the recommendations and others and what further needs to be done.
At the time that this report was released, our committee felt that the subject matter of the report and the findings and recommendations contained in the report took on added importance at that time in the context of the health care system issues and all of the efforts at developing a comprehensive health insurance plan. However, the important issues of health care systems, which appear to get more complex by the day, and the urgent need to slow down the escalating costs of the health care continue to be a subject of debate today and in some ways more so.
The health care landscape continues to change rapidly with the emergence of new health care providers and health care settings, new financing arrangements and new technologies, so it is time for you to get back into this subject matter. Recognizing the need to be up to date on meeting the needs of the country as well as realizing that some of the data systems have not kept pace with the changing landscape, NCHS developed plans to restructure their health care survey providers and service setting data. They developed what they termed an integrated national health care survey building on the existing provider surveys and merging and expanding that coverage over time and geographically linking them to the National Health Interview Survey.
NCHS asked the Academy to undertake two major tasks to advise in the development of this survey plan that they had: to identify the principle, current and future needs for health care data by public and private policymakers, health care providers, health service researchers and others; and two, determine the extent to which the proposed restructured survey can meet those identified needs for data.
This committee that we established was co-chaired by Dr. William Richardson and Dr. Edward Parren, a former director of NCHS. The report we issued, that gray book that I think you all received, responds to that request. It reviews the main features of the proposed survey from two perspectives: the extent to which the survey would enable NCHS to meet the changing data needs associated with the health care delivery and policy issues identified by the committee, and the technical features of the plan in its current state of development at that time. However, the report, like many of our reports do, goes beyond a simple review of NCHS plan and suggested a broader strategy for the surveys that the committee believed would more nearly meet the needs of the health care system.
As the committee reviewed the plans and reviewed the data systems in the department, it became very clear to the committee that no single survey is likely to meet all of the criteria, address all of the technical problems or meet all of the users' needs for data. A coordinated and integrated system of health care data collection activities making up a data system is required involving several organizational units. That is a difficult task.
In the committee's judgment, the NCHS plan as presented to the committee did not provide the full capacity to address the important questions. If it felt that it was a good step to link the survey geographically with HISPSUs(?), that, by itself, will not result in a meaningful integrated survey as suggested by NCHS.
The committee moved on and developed and 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 the design framework to see if they are feasible or make necessary changes before final decisions on the details are made. It then proceeded to make several recommendations about using the health interview survey base sample for surveying cohorts of individuals using the national nursing home survey based sample for sampling and studying a cohort of residents and using what the surveys-based sampling to survey episodes of illness. I will not go into those details here; we went into great detail in the report and you may have read it or you may read it at your leisure or may decide not to read it.
I will give one little example. The panel recognized that surveys for gathering information on medical expenditures are very expensive. Yet, we know that it is very important, especially in today's climate. We need current data, not waiting ten years. We need to do it in a way that is cost effective. The committee really deliberated on it and went back and forth. I remember that vividly. They finally came up with a recommendation of selecting a sample of individuals from the health interview survey respondents to collect the basic data on health care use and expenditures. This is not a new concept. This was developmental work that had been done a few years before then by RTI for the department when we were designing one of the predecessor's expenditure surveys. They keep changing the alphabet soup, and it is hard to remember which.
At that time, that idea was not picked up, but we were pleased to see that NCHS and the Agency for Health Care Policy Research had moved somewhat in recent years in that direction. Budget constraints and willingness to work together, which is very important, sometimes has surprisingly beneficial results. I was going to say two of the key players, but now we have two and one key player is here, and he will speak on the plans, but let's not carry this budget constraint problem too far. I think that needs to stop.
In the course of the study, now, the committee went further, as I said. It noted several issues not directly addressed in its charge. We were not specifically asked to deal with them, but the committee felt it would be very irresponsible of it if it did not. Some were broadly related to NCHS and not just the health care survey, and others went beyond to the structural issues of collaboration and coordination of data gathering and analysis within the department. It strongly believed that these broader issues must be addressed if we are serious about successfully developing and implementing an integrated data system. Its success depends on the resolution of several issues.
The committee then went ahead and made recommendations on four general areas: advisory structure for a health care data system, improving departmental coordination and collaboration, enhancing the National Center for Health Statistics' analytical capabilities, and resource requirements without which you cannot really do anything.
On the advisory structure, it went into great discussion on the fact that one needs to periodically review all major surveys. Although the National Health Survey Act was passed about 40 years ago, it was still extremely relevant. At that time, it was 34 or 35 years ago. The committee noted that no major fundamental change in the structure of the general organizational framework in response to changing health information needs had occurred since the latest maybe small changes in the 1970s.
Since we finished the study, I understand that NCHS has instituted several new surveys and are moving toward making some changes within the organization. The committee believed that the concept and operations of this kind of a data system should undergo external review.
It recommended two kinds of external reviews. one was advisory on the subject matter, and the other one was an external technical committee to advise NCHS as it is developing the survey on the data system. It gave some examples of some productive efforts done by other agencies.
When it came to coordination and collaboration, we noted the review of all the data systems that there is an abundance of data available, but they are all fragmented. There is a total lack of coordination. There even is an appearance of overlapping data and duplication. The committee felt that a major review of all of the data systems is essential. It recommended that the department establish an ad hoc, high-level, external committee composed of persons who have distinguished themselves in the field of health statistics, survey and sampling methods, and the provision of health services to undertake an independent and comprehensive review of the health statistics activities throughout the department and report its findings directly to the Secretary. This committee should have a budget and a qualified staff support independent of all agencies involved and well respected by all parties of interest. In addition, major data systems should be reviewed on a periodic basis.
This was a key recommendation of the committee. Then it moved on to the need for improving and rebuilding the analytical capabilities in NCHS. Over the years, we felt that especially in the survey divisions, with budget constraints and an effort to keep maintaining the surveys as such, as everywhere else, the analytical resources suffered. It does not help to do that not just because you cannot get the data out in time and analyze and integrate that data, but also it is important that the analytical people would be able to understand and know what kind of information is needed out there today.
Finally, the committee emphasized that without infusion of substantial resources, the course of action charted by the committee would not work. It considered the immediate implementation of its recommendations justified in view of the importance of the information not just to NCHS, not just to the department, but to the Congress and the entire Executive Branch.
An underfunded program is not going to do anything. In the final analysis, the commitment and institutional support of the Secretary of the Department, the Office of Management and Budget, and the Congress are all essential to the successful implementation of a comprehensive, integrated health care strategy. This statement was made six years ago. I guess it is true today, maybe even more so.
NCHS and the Department have taken some steps in some of the areas addressed by the committee and maybe not in some of the other important areas. I look forward to hearing you give us your plan. Thank you. If you have questions, I will try to answer.
DR. IEZZONI: We don't we just see if anybody has any questions about the specific report, and then maybe we could hear from Ed and Ross on areas where you feel that you have not been able to do some of the things that the IOM Report outlined and areas where we maybe could just kind of lend our voice to recommending reconsideration of some of those areas.
So, are there any committee questions for Gooloo about the specifics of the IOM Report?
No. Okay, why don't we near from Ed and Ross, and if you could kind of play off where Gooloo was at some point and let us know where you think the kind of major areas are that have not been addressed.
DR. HUNTER: Sure. It just occurs to me that people are listening on the internet must be really wondering who are Ed and Ross and why they sound the same way in their voice. Actually, I first want to say that I really appreciate being here. Ross was really sorry that he could not make it. He was planning to until I talked to him at home last night, and he was going and doing MRIs this morning, so he was at a level that it was not a good idea to come limping in here.
We came physically, both of us, I think two years ago and talked to the full committee about what the department had proposed, and many of you had actually been involved in discussions that the department had of the plan as it was announced as part of the vice president's reinventing program. At that time, there actually were people who were talking publicly; other than Ross and I, none of them are here anymore. So, we are left to carry the torch.
I did want to talk about and walk through first some of the things at a very basic, background level since many of you have not heard the kind of general picture. I will spend very little time about that. I will talk about things in three categories after that: things that we said we were going to do several years as part of the formal department plan that we have done and talk a little bit, not too much detail on each of those; talk about some things that we said we were going to do that we have not done yet; and talk about some things that we are doing that are within the general, sort of the spirit of the department's efforts to integrate that we did not specifically say we were going to do several years ago but nonetheless are important parts of where we are now.
Then I think there are a couple of notes that I think Gooloo is raising that I would not have otherwise covered in those things. There are a fair number of things that you can hear in the report's recommendation six years ago that we clearly have done, and there are some that we clearly have not done.
I do have something that I will leave. I did not bring enough copies. It is an editorial in the ASA magazine, Chance, that Ross and I did a little while ago that kind of lays out more -- less of the details of the plan itself at a tactical level and more of the idea of why we did it. That might be useful as background to you. Some of it sounds really optimistic. At the time, this was a previous life.
The first thing that I wanted to note is that there was something called an HHS Survey Integration Plan, and it is still an active document that the department supports. The idea at the time was not just to come up with one specific list of tasks and move forward on them. The idea was to have more of an overall approach that would change the way that the department does business in planning surveys.
There was a lot of commentary at the time. This plan was hatched in a very short period right after health reform kind of went down in flames. Fresh with that experience, I think that everyone in the department and a lot of our partners on the outside were very sensitized to what we did not have and that there were major data gaps in our ability to look at what the impact of health reform would be, to have baselines for tracking change, the research that would have supported a lot of the policy decisions that were needed and considering this within the administration and Congress. I think that was one of the major prisms through which the department's data systems were looked at. There were many of them. There were literally hundreds of data systems in the department, all of them in sort of stovepipe form, and many of them unable to answer the key questions that policy people had. We actually had a lot of people involved in this that were very involved at the senior policy level in searching for those answers and not being able to find them.
So, we went at this both from the standpoint of, gee, we have a lot of data systems that appear to be doing similar things so that we are sort of search for the smoking gun of duplication and overlap, and at the same time, this real consciousness that even with all of the things that we had, we just did not have the answers to the questions that were many of the most important ones.
There was also a real emphasis that if we did find areas that we could reduce overlap or accomplish linkages or consolidate some efforts that appeared to be similar, there might be efficiencies that we could create that would allow us to make some investments to meet those high priority needs.
We did have a plan that we cleared extensively within the department and went to OMB and got buy-in at the very senior level at OMB that many of you may have seen. It was back in April of 1995. I will always remember the dates that we published this and went to OMB to discuss it because it was about three days before my wedding. I am not allowed to forget that date, so this one is always -- the briefings were all arranged around when I was going to be out of town and not available.
There are several things that have come in addition to that plan that I think are worth noting. That plan had the specific set of tasks that I mentioned, and I will go through a number of them. There was also, I think, the presumption that the days of just independently planning surveys that would meet a narrow, agency-by-agency interest, were behind us and that we were going to think more strategically, look at things more from a departmental point of view, look for opportunities where these surveys could be modified or enhanced or linked in ways that could meet multiple needs rather than just one. That is a very big goal for a set of agencies that really do have very specific missions and whose budgets go up through very defined channels and who have very serious tradeoffs within each of their organizations. So, that was a big task.
To deal with that on an institutional basis, we put a number of things into place that we can talk longer about whether they have succeeded or not, but the department has a -- formed the Data Council. The Secretary formed the Data Council with the charter to coordinate things across the department, to take a look at survey plans as they are developed within the agencies, also to stimulate things at the agencies that they may not do themselves, and also to take a look across the department at budget priorities and sort of give input to the Assistant Secretary for Management and Budget and the Budget Review Board on things that maybe the agencies did not include in their proposals.
There is always a concern that for multipurpose efforts that where one agency only captures a very small part of the benefit but yet it addresses interests of the department as a whole. Nobody is including that in their own individual budget requests, so the question is what does the department do to kind of assure that the right investments are being made and that the wrong investments are avoided. So, the Data Council was formed for a variety of purposes beyond this, including the HIPPA types of activities that you are already familiar with what the Data Council's role is. Ross and I co-chair a survey integration work group that is a part of the Data Council.
There is then a formal process that has not yet really taken hold as a strong process, but we do have a process in place that develops an advanced plan for surveys where agencies, at the time that budgets are put forward, are also asked to identify their survey and major data systems plans. Everybody else then has an opportunity to look at those. The Data Council can then take a more detailed look at areas that are thought to be of interest or where things could be modified or expanded or changed.
It is sort of a -- one of the few times we have put into place that allows everybody to see several years even ahead of time what the other agencies are thinking about and thereby give people a chance to go and work on collaborative efforts. It used to be that, and still on an exceptions basis, things can get all the way up to the point of contractor award and clearance before you would necessarily hear about it in another part of the department. That, obviously, is later than one can deal with any changes.
Lastly, the Secretary did establish Ed Sondeck, the Director of NCHS, in a dual role of Senior Advisor to the Secretary. In that role, he has the opportunity to look across the department at these types of data planning issues and stimulate discussion of things where there might be opportunities for change.
Kind of turning to the things that we said in the 1995 plan that we would do that we have accomplished, and I think a number of these echo recommendations in Gooloo's report. I will try to footnote those that do, and you can then chime in and say that they did not really quite make it.
The first thing, and still I think the major change that was brought about in the operational sense was the linkage between the health interview survey and the medical expenditure panel survey of AHCPR. At the time this plan was first considered, the MEPS(?), what was called the National Medical Expenditure Survey, and was about to be launched as a free-standing, one-time survey. The MEPS historically has been conducted about every decade, and I think we did change the name of this every decade just to keep people confused.
I think one of the real concerns at the time was when health reform was under consideration, the most recent data from any major expenditure survey was from a 1987 survey. Paul is nodding his head; he is a user of both of these surveys that were linked. The MEPS, because of resource issues, was being planned as a point-in-time, cross-sectional study with the option to do a one-year follow-up. It was really felt at the time, and this echoes in the recommendation in the report, that we needed an ongoing basis for looking at medical expenditures. One thing that was a major change is that the MEPS was turned into an annual, ongoing survey where people are enrolled in the panel and followed for two years rather than for what was essentially a year.
The other major change was that the sample for the MEPS, rather than being drawn from an independent frame was drawn from the respondents to the HIS. Ann Hardy, who is the Acting Director of the Health Interview Survey Division, is going to talk a little bit more about that.
The major accomplishment, I think, and in addition to just the sample linkage, there was then an attempt to integrate the question or content so that we did not ask questions in different ways and that there was a crosswalk between the individuals who were interviewed in the health interview survey and the response to those questionnaires and then what was later done in the MEPS. The intent is that we follow that through into other surveys as well. I will talk about the health and nutrition examination survey as an example of that.
The major focus of this plan was the MEPS/HIS at the sample level, at the questionnaire level, and in a very big hurry, we turned around the operations of both the HIS and what was going to be the independent NEMA(?) survey and accomplished the link for the 1996 survey operations. That, I think, at the time was considered something just short of a major miracle that the two agencies, which had been at odds over these types of things for decades, actually came together, worked out a joint plan, and then made it work in an operational sense.
I think the glow of that grand accomplishment has worn off, and now it is time -- we are really having to take a serious look, now that data is flowing from these linked surveys, at just how successful has that really been. Aside from the kind of geopolitical success and the idea that we now did not have two surveys that were completely contradictory and had some linkage, not only in time but in place and in person, as well as in content, I think now we are really starting to take a serious look at the cost factors. There are some pluses and minuses. We are able to do a lot of things with the HIS sample that would have been very costly to do with an independent sample with a lot of screening.
I think we are finding that releasing data to the public is somewhat -- a lot more problematic because we now have kind of upped the ante on the disclosure risk when you have this much more information on the same individuals. I think there is still an open question as to -- forgetting the operational efficiencies or benefits, what is the analytic payoff to having done this. This was always something we intended and still is out there as a goal, but we have not yet taken advantage of the fact that we have the combined dataset with all of the depth of expenditure data on the MEPS and all of the information on non-expenditure things that we have from the HIS that are not on the MEPS, and in effect have an additional year of observations on these individuals. So, we really are needing to go back and test the waters on the analytic payoff on that.
The next thing that we -- by the way, if anybody has questions on things that you want to just stop me, I would be happy to do that. There is a limit to the detail that I am going to be able to answer questions on on the design and some of the operational things, but there may be others in the room who can.
Another area that we said we would do that we did was a -- we used to have two separate employer health insurance surveys in the department. NCHS along with AHCPR and HCFA had conducted a national employer health insurance survey in 1994, and the intent was to repeat that on a periodic basis. This was something borne out of the data needs in health reform. The expenditure survey, meanwhile, was preparing to go into the field. It would have included as part of that a follow-up to the persons who were interviewed, going back to their employers to get more information on the health insurance plans. The content of them was very similar; the operational approach was different. The goal was to merge these in an operational sense and in a context sense so that you could crosswalk, so that the same field agent could do the work in the employers that were approached because of the fact that there is a household respondent and also those that were drawn from a national frame.
We did merge those two surveys. That is now the insurance component of the MEPS. That insurance component includes both frames, in effect, where there is a frame that is derived from the household respondent and one that is done at a state level, drawn from a frame of employers that the Census Bureau has.
A third area is the National Health and Nutrition Examination Survey. We did say at the time of the plan that our intent was to use the health interview survey as sort of a launching pad or a sampling platform for a variety of other household surveys. HANES was one of them. I will talk a little bit later in the category of things that we have not done yet about the National Household Survey on Drug Abuse.
With HANES, the original plan, which was, as I said, put together in a very short period of time, called for the HIS and other subset of the HIS respondents to be approached and examined as part of the National Health and Nutrition Examination Survey. At this point, our plan for HANES, which we hopefully will launch in 1999 if we can round up the resource picture in that survey, we are going to the same sampling areas as the Health Interview Survey, and then we had a longer term agenda to look at the possibility of linking at the person level.
HANES, for those of you who are not familiar with it, is a very different animal in a sampling and operational sense. Whereas the HIS goes to 200 or 300-some, many communities and interviews relatively few people per community because the resource input is just a person who goes to a doorstep, with HANES, we go to few communities, in our current plan 15 per year as opposed to hundreds, and interviews as many. Because you are showing up with a mobile clinic, your efficiency is gained by setting it up in one place and getting a lot more people than we would with the HIS. So, there are some real issues to be worked out as to whether it is feasible to go to the person level. It may really involve changing the way that we do the HIS in some of those areas, and that is a longer fuse kind of effort.
We have linked the HANES to the HIS in the questionnaire sense where the questions that are asked in the household portion of the HANES are those drawn from the Health Interview Survey, so there is at least that questionnaire and analytic coordination. We are also going to the same areas with the plan to look at more steps on that.
We also, at the time of the plan, talked a lot about our provider surveys. This is another place where Gooloo's report is very relevant. Not only do we have the issues that Gooloo mentioned but even in a practical level, the department has many different instances where we show up at the door of a provider for somewhat similar information, and we do it very differently, with very different forms, very different contractors, very different analytic possibilities because of that.
The intent at that time was to start to push these towards more of a merger or consolidated operational platform recognizing there may still be differences in what you need to ask. An example is in the area of nursing home surveys where AHCPR has a nursing component to the expenditure survey where they go to a sample of nursing homes and get a lot of information about the nursing home and their patient population but then get a lot of detail on expenditures for those individuals. We have gone, at NCHS, at periodic intervals to nursing homes to get a lot of information about the nursing home staffing patterns, patient loads, on a broader sample but not as much detail on expenditures. The intent was that we somehow put these two together.
At this point, there are also issues in the hospital area. There is a variety of things that we all do at different agencies, at hospitals and emergency rooms. There is an emergency room survey, an NCHS emergency room survey. They all grew up for very good reasons with different approaches because the SAMSA one on drug abuse and NCHS one is on a broad base of all things that present themselves.
We have not quite solved the provider problem yet, to say the least. What we have done is that we have decided as a department we did not have enough of a grip on why we were doing all of these provider surveys and what we really wanted out of them and what sort of our analytic framework and even what our conceptual model of the health care delivery system was that we were trying to make sure that we could quantify. So, we stepped back several steps and made our lives much more complicated by trying to pose what we thought were the key analytic priorities.
There has been a work group in the department that has been working for a year or more. We had a contract that went out and surveyed a wide number of people in the outside world to talk about data gaps and data needs. We have done the same thing within the department and with other federal agencies. We are trying to pull all of that together.
As part of that -- I think this really does pick up on the fact that many of the things that were discussed in the IOM SINSTAT(?) report about data needs and changes in the health care system are, with minor editing -- we talked about this on the phone the other day. With minor editing, you can just substitute a few words, and it is still very true. What we are trying to do is drill down a few steps further and actually put some specific policy questions and research questions down and then let that drive our data systems planning and this side, sort of the supply capacity delivery system side.
To help us do that, we have started what has now taken on a life of its own as a dialog with a variety of outside folks. We first invited a couple of people from the outside in to meet with our department work group, and it was a really stimulating discussion. We have people like some major professional associations that conduct surveys in the same area that we do. We thought we would hear their input on what gaps and needs were and what their survey plans were. It actually turned out to be a really interesting exchange.
Rather than sending them away and bring a different group in for the next meeting to learn more, we actually heard from them that they really enjoyed coming and talking to each other and to us. I guess the best way to describe this now is that it has taken on a life of its own. It is no longer a department-driven work group; it is sort of a free form public/private group that has met now three times and will be doing so again. We are starting to look at -- I mean, we sort of asked them to look at and comment on things that we are working on. We have been sort of exchanging information on each other's survey plans.
One issue of real importance that has been identified is this whole question of topologies and, for lack of a better term, the descriptive variables that you would want to use to describe different types of organizations. The discussion always starts at each of these meetings with the HMO, PPO and the inner study definitions do not really do it, but what does, and is there an analytic axis that you could describe that would be around risk or around affiliation or whatever. The part where there is no single thing, we are now talking about multiple kind of descriptive variables and what do we do about that, how do we stimulate that and push that along so that many of these different surveys could include something similar so that one could crosswalk these things.
I think out of this initial focus on the operational side and kind of needing to come to grips with what we were doing with some of these different surveys and what we wanted to get out of them. We had a couple of things going on in the department and outside that are the mechanisms that would be a discussion in an integrated way of getting absence of health care costs, access and insurance coverage, and health status types of things that we did not get through other existing surveys.
The thing that we did do is that we pilot tested what we now call the SLAITS, and Marcie is here to talk about that a little bit more. There are obviously some positives about SLAITS in terms of integration in that it uses an existing platform, the National Immunization Survey, and it draws questions from other existing survey activities including the health interview survey and others. There are -- I mean, obviously the things that have not really been played all the way through is that this not the only survey that might get that state level data.
There is one that we piloted tested that works, and it works for many purposes. The HIPPA risk factor surveillance system which Eve will talk about is also out there in the states, and I think that is something that could be more -- I assume there will be more discussion about that. So, that goes into both categories. We have pilot tested SLAITS and are working on it, and Marcie will talk about what we are doing now and looking at SLAITS to test it, to see how it works to attract Welfare Reform issues at the state level.
What we have not done is actually implemented SLAITS in any significant way. We have done pilots and health modules and net Welfare modules and looked at questionnaires and approaches that would use SLAITS to do the S-CHIP(?) evaluation, but there really is no SLAITS in any operational sense.
I am going to stop there on what we have done and talk for a few minutes about what we have not done that we said we were going to do. There are many things that we have not done, but some of them we never said we would so, so I guess they do not count here. I mentioned with SLAITS that we have not actually solved this riddle of how do we address state -- our needs as a department for state level data and also work with states either on the capacity side or on what might these vehicles do for them either. That has been dealt with through budget channels and other channels, and it just has not been solved yet.
We did say in the initial plan that the Medicare Current Beneficiary Survey that is done by HCFA would, at some point, be brought into this cluster of HIS MEP surveys. The MCBS does similar things to what the MEPS does but draws its sample from Medicaid/Medicare beneficiaries. Operationally, they are similar, but obviously one focuses on -- the MEPS does not oversample as it would have done the over-65 population, and the MCBS is the primary vehicle for HCFA's data collection for programmatic purposes and for their own policy purposes.
There were a lot of issues and a lot of controversy about this back when we developed the department plan. I think because there was no ongoing MEPS to say, hey, yes, this works and so obviously it is time to bring the MCBS in, and because the operational side and the efficiencies really were an open question, it was decided at that time we would come back and revisit that once the MEPS was sort of a proven entity and these other issues had kind of been resolved. We have not come back and taken that look at this point.
I talked a little bit about the provider surveys. We have not dealt with the operational side of provider surveys other than where they meet around the MEPS components.
We did talk about using -- looking at the possibility of using the Health Interview Survey as a sampling platform for the National Household Survey on Drug Abuse. There were some real unique issues with that survey that we noted at the time, that it was an anonymous survey rather than a confidential one. It was not clear at the time, and the research still has not been done to know whether the respondent really recognizes that difference. Obviously, if anonymity is a key factor, a linked design where you start off by knowing who the person is is not really a viable option.
I think it has become further complicated by the fact that the household survey has been expanded greatly to look at state level issues. There just is not the HIS sample there either. That was something that was in the original plan that is still not accomplished.
To some extent, those are, I think, at least compensated for by a couple of other things we have done that we never put down on the specific list but I think are within this intent, within the spirit of what we said that we were trying to do, which was to be a little more rational and do things -- you know, not build new stove pipes but see if we could try to find some ways around that. One thing is with the National Household Survey on Drug Abuse. I think most of you are aware of the fact that the President has mandated the department to do a survey on smoking behavior of adolescents. This was an active discussion that was a part of the Tobacco Bills on the Hill.
The initial instinct of the department and the agencies, I think, is always, gee, we have a new, important data need. In fact, there was a time line on this that was very pressing because of the provisions of the legislation. Obviously, we needed to develop a new survey and do it fast. All of the effort at first was into how do we get a contract out fast enough; we do not have time to stop and think about how this fits in with anything else. I think ultimately what -- the idea behind integrating surveys and not doing things separately kind of came more to the fore and instead of doing something separate, there is a new component added to the National Household Survey on Drug Abuse. I think many of us considered that a success not specifically in the survey integration plan but of the idea that the department can think across agencies and do things in a way that I think makes more sense, both more efficiency and also analytically.
We are also looking -- or I mentioned that the Health and Nutrition Examination Survey is returning to the field in 1999. There has been, for decades, a controversy in the nutrition community about the fact that there were -- that HANES collects nutrition data as part of its broad range of health and nutrition issues. At the core of that, there is a nutrition -- a dietary intake interview. The primary intent of HANES is to relate dietary intake to biochemical measures and health status that we observe elsewhere in the survey. USDA, meanwhile, has historically done nutrition surveys that start with the dietary and intake survey but relate that more to agricultural economics and consumption and production types of issues as well as human nutrition issues. There was even legislation passed in 1990 where the intent was to force the two departments to talk to each other across this sort of historical animosity.
I think it is encouraging now, at least. The final details have not been worked out, but it appears that the HANES and the continuing survey of food intake individuals, the CSFII of the USDA -- the current plan is that these two surveys be joined and that the HANES clinic- based interview be sort of surrounded by a telephone interview that would be linked to it geographically and also nationally, and that the two surveys would in effect be one. Not just a linkage in content but actually in more of an operational sense, these two surveys would be merged.
If that comes to pass, then of course there are a lot of issues that are even harder to deal with across departments than they are within the department in terms of confidentiality, and these are two different appropriations committees or subcommittees. If the plug is pulled on one, what does it do to the other. Nonetheless, I think that is a real positive step.
Another step is a next step on employer surveys. I think when we merged the two employer health insurance surveys that the department did, we thought we were comfortable for a while. It was not an easy thing to do to merge those two surveys. The next thing that came up when we went to OMB and elsewhere was, well, gee, the Bureau of Labor Statistics and the Census Bureau, and then there is a list of ten federal agencies that are involved somehow in looking at employer health insurance or employee benefits provided by firms or establishments and what are you going to do about that.
There is a current effort that crosses departments that is co-chaired by the Bureau of Labor Statistics, AHCPR and NCHS with the involvement of OMB and the Bureau of Economic Analysis and the Census Bureau. There is virtually every agency that touches on economics, labor or employers to look at what do we do with the HHS survey of employers and the Bureau of Labor Statistics compensation surveys.
We are also trying to drive that effort from the data needs point of view which is what are the analytic objectives, what do we really need to know, what does HCFA need for the national health accounts, what does BEA need for the national economic accounts, what do agencies need, what does Treasury need for tax planning purposes, et cetera. That is a longer term thing which is involving a lot of work by staff, particularly at AHCPR and BLS. There are a lot of other issues on content and questionnaire sharing that I think are being worked on. You have the CAPSIS as something you are going to discuss further today.
So, that is kind of the -- I probably have gone on longer than I should have here, but that is sort of the sense of what we are doing and what we have not done. There are real limits to how far you can push things when we still have a very decentralized set up in this department. Gooloo's report, I think, talked about the fact that these things exist in multiple agencies with very different missions. I think we have worked very hard to try to join things where they can be joined and have them touch on each other where they can be touched on each other, but there are some real limits given our current model.
In terms of some notes that I made on the report, just real briefly, I think we have not done exactly what the report said about these broader structural issues. The Secretary did give very serious consideration when this survey integration was put forward about reorganizing things within the department and ultimately decided that there were good reasons why different data collection mechanisms were associated with the program agencies that they were in and set up the senior advisory role and the Data Council to try to bring things together despite that. So, there was a conscious discussion of that. It was not done with an outside formal group of senior eminences, although there were some eminences brought in to discuss with her directly that question. She did make a judgment at that time that there were many benefits that outweighed that.
I think the overall question of outside input, I think obviously this committee is making an effort to step into some of the roles that the report talked about for ongoing outside input. We do not have the specific technical level input groups in all cases that the group called for, but I think between this committee and sort of the sounding board that we are using on the provider side data system and the workshop that we are heading toward with the Committee on National Statistics and the National Committee's involvement in that, I think we are hoping -- at least, I know it is a real goal of Ed Sondeck to get more outside input into all of the things that we are doing.
One last comment. The report did talk about resources and the need for a new infusion of resources. I, particularly in a taped and broadcast meeting will not say a whole lot more about that other than to say that at NCHS, since the report was published, the only increase in resources, even including inflation and cost of living type of things, has been to get a partial increase for what we needed to get HANES back into the field. Other than that -- and we are still short on that. Other than that, we have been flat lined and therefore declining.
If you look at the cumulative effect of just inflation on that, we are about 25 percent less than where we were at the time that report came out. I think that is not really to say that there were things we would have agreed with and taken up despite that, but it has been a real restraint on our ability to do more of the research that I think was called for.
On the other hand, there have been new resources put into data in the department that I think are important. At the time of that report and at the time the survey integration plan was put together, there was no MEPS in the budget, and there now is an ongoing annual expenditure survey with an insurance component that filled in two of the largest data gaps that we identified at the time. Those appear to have been somewhat institutionalized. The AHCPR budget had some real ups and downs and questions in the early 1990s, but that seems as though, if you look at the 1999 committee marks, that the MEPS is very solid. We did get the resources to get HANES at least to the pilot phase and to get it close to going into the field.
I think, in part, we are very strapped except for these very specific things that money has been put into, and that has restrained the ability to pick up on some of these ideas and go further with research. Staffing has been a problem as well. I think there were some more specific things that we could talk through, and Gooloo and I could have a healthy debate over lunch, but I think a lot of those things we have been moving on. Let me stop there.
DR. IEZZONI: Yes, I thank you for a very thorough and thoughtful presentation. Can I ask the subcommittee members? Barbara, comment?
DR. STARFIELD: You guys are terrific. That was really a wonderful report. I have two specific questions. One of them has to do with the broad provider survey, in particular, NAM-SEES(?). Is there any thought to building in a longitudinal component of that or linking it with the provider aspect of the MEPS?
Second of all, could you give us some idea of this public/private effort? What organizations are involved in that? You said surveying organizations. I would appreciate a little more specificity.
DR. HUNTER: Okay, sure. On the first one, I think the short answer is no, nothing specifically has been put on the table and discussed. I think one thing we have done is because we had not -- because within the department we had not established any set of agreed upon analytic objectives or policy needs or sort of the overall description of why we even needed to be dealing in the provider survey area, we have not -- almost as a conscious matter have not moved new plans for what specifically we would do with our existing provider surveys forward, but I think that phase is pretty much ready to finish and move on to the next one. The short answer is no, no specific plan has been discussed about that.
DR. STARFIELD: But you are going to talk about it at some point?
DR. HUNTER: I think -- yes, I think it should definitely be talked about. The second question on who all is involved in this group, we have -- it is actually a very mixed group at this point that we are not -- because of issues with advisory committees, which this group I am describing is not, we are not even in the position of determining who comes to which meetings entirely any more. We are a part of this meeting, which I think is actually -- it is so different from the way that the department usually does business that the people who are involved occasionally sit back and say, my god, we are not controlling this?
We started -- the most recent meeting was a week ago. It was hosted by the American Hospital Association, and Joe Martin was the host and chair of that. Other involved people were -- I am not going to name them all just because I do not have a very good memory. Jim Rodgers from the American Medical Association and several of his staff who work on the AMA surveys were there and presented. Don Young from AAHP or formerly AAHP, sometimes puts on his hat of being in various other roles. We have people from NAHDO, National Association of Health Data Organizations. We had a state representative; Pete Bailey from South Carolina came to this meeting. Some of the medical practice groups, the Medical Group Management Association, the Health Care Financial Management Association. Linda Bellheimer from CBO came not as a data collector but sort of putting the analyst and user hat on.
DR. IEZZONI: How are the meetings announced, and how do people know about the meetings?
DR. HUNTER: It has been a very informal -- I mean, it is not a formal meeting in any sense that has been announced. It has been sort of --
DR. IEZZONI: Word of mouth then?
DR. HUNTER: Pardon me?
DR. IEZZONI: Word of mouth then?
DR. HUNTER: It has been -- it started actually as a -- we just invited some people in to get some reactions, and it has just been -- it has evolved. It has not really been a -- we may want to talk about other involvement in that as well.
DR. IEZZONI: Paul, you had a comment, and then Vince.
DR. HUNTER: Barbara, I am failing to think of the rest of the roomful of people.
DR. STARFIELD: That is okay. You actually clarified it because I had the impression that it was the proprietary organizations that were involved, and you clarified that.
DR. NEWACHECK: Thank you both for informative presentations. I have a question for Ed, actually two questions. The first one is about the potential advantages and disadvantages and how they have played out in practice of survey integration. You indicated or you talked about some of the large potential advantages being the efficiencies or potential cost savings and the analytic possibilities from survey integration. I am wondering, though, if you can comment on, now that this has been done with HIS and MEPS, we have had some experience, whether they have played out in practice, whether there are demonstrated cost savings and whether there are demonstrated improvements in analytic capacity as a result of integration.
Second, I wonder if you could comment on whether there are any disadvantages that have emerged from survey integration in this case?
DR. HUNTER: I think the one unknown, the one thing that we have really not taken advantage of and looked at is the analytic side. I think the two staffs have been so busy getting the surveys fielded and doing sort of the first cut, in the AHCPR case, a very productive set of preliminary reports from the first rounds of the MEPS on access and transcoverage, and Health Affairs is being kept busy on some of those things. I think because of just the operational things that have driven getting these surveys together, we have not really looked very closely at one could do to analyze the two datasets together. So, I think that is a very open question.
DR. NEWACHECK: Can I make a comment on that? As a public user of these datasets, there have not been any really demonstrated advantages yet of integration in terms of the analytic capacity. In fact, there may be even some cost to it. For example, in the MEPS data that is available now and has been available for almost a year now, there are no income data on those datasets, and there are HIS income data that could be linked, but it is not available to the public, I think, because of disclosure concerns. So, in fact, it is kind of a handicap, I think, in the way that this has been designed from the public user perspective. I could see potential advantages, but if you have to be a member of the government to use those, to take advantage of those analytic capacities, it does not really serve the user community very well.
DR. HUNTER: Yes, I think one of the things that we have not solved yet, and I do not think it is -- and not to say that it cannot be done yet, because I really mean that it has not been understood as what we do with a linked file for public release. I think that is the single largest disadvantage or problematic area. I think when we accomplished the link in the first place, the interagency agreement that allowed us to transfer a sample to AHCPR and in turn their contractor was 13 single-spaced pages long. We had more lawyers involved than we did statisticians.
A lot of the complication is just the whole disclosure avoidance thing. So, we have a series of agreements with AHCPR where we each review each other's releases, and part of the -- and I think we have not really finished the discussion of what we do with the linked release, but it has proven very vexing. For that reason, and I think both agencies are talking about establishing public research data centers like the Census Bureau does that would allow people to come in, at least, and use linked files even in identifiable but not identified form. We are hoping to get something set up this fall, and AHCPR is using largely the same model.
On the other side of -- the other non-disclosure things, and I do not discount those as being real problems. I think the cost issue is something that we are really looking at now. The contract for the next MEPS is under review, and so that is something that I do not want to go too far into. I think there are a couple of things that have complicated the cost comparison. Obviously, the cost of screening in the historical NEMNIS(?) model was very high because they were looking for some very small groups of policy importance, you know, high users of care that took a little bit of work to find, as an example. From the HIS, you can more quickly find them. Traded off is the fact that the HIS is distributed widely around the country and NEMNIS used to be in a relatively small number of areas. So, the cost of distributing the sample, to some extent, offsets the cost of not having to do the screener. So, there is a lot of things being worked through now in the current evaluation. I think within a few months, we probably will have a lot more information on the cost savings.
I think there are -- I mean, there are a lot of things that we thought would be real advantages that still could be advantages. I think the opportunity to use the HIS and all that we know about the individuals in the HIS, to really fine tune the MEPS sample and go after even more carefully selected groups is there. Because the HIS, the first year we did this, was a paper and pencil survey, and we did not have the turnaround that would be required to do that and still get them into the time table for the MEPS, that really was not possible. I think the MEPS then, because of the AHCPR budget issues was not able to go with a lot of the more specific sampling that they had planned to do. So, that is in sort of an unmet positive. I think it is hard to say what the net is right now, but I think you did put your finger on the really most difficult one.
DR. IEZZONI: Vince, you had a question, and then Hortensia.
DR. MOR: I do not know the answer to this, and maybe others around the table do, but some history. Why did the committee, the IOM committee decide or suggest that a list of external senior-senior folks be involved in this as opposed to this committee?
DR. WUNDERLICH: First, before I affect anybody in my academy's bureaucratic or ruffle any feathers, this study was a joint study between the National Research Council and the Institute of Medicine. I used to be in the National Research Council when we did the study, but it was a joint study. Now, I am in IOM; that is why I introduced myself as IOM. There are some sensitivities, so I thought I had better clear that on the table.
The committee decided on that because it had -- a lot of the committee members were very familiar with and very knowledgeable about the departmental activities, and they felt like just a bureaucratic decision is not enough. You want a major review across the board of all of the data systems and see what is still relevant, what can be made more relevant, what is overlapping, what should be combined, all of these things, and then make recommendations to the Secretary. They wanted an independent, outside group to do that.
The second part of that was again to get around the internal problems. They said the group should be out, external. It should be supported by staff that is independent of all of the agencies involved. Experience tells you very fast that you do not have to be very brainy about it. There are tough problems. I have been around here for 30 years. Sometimes there are difficulties in letting go of things. We do that in every bureaucracy, not just here. We do it in the Academy, too. Don is not here, right. It just felt like somebody knowledgeable and must be respected by the agencies, but the agencies would provide the input and work with them, but it should be from an outside, prestigious group coming up with recommendations that would be more saleable to the department, to Congress, to OMB.
DR. HUNTER: A footnote on that --
DR. IEZZONI: It makes sense, Gooloo, yes.
DR. HUNTER: You had also said why not this group.
DR. WUNDERLICH: I think we have those explanations there.
DR. HUNTER: This group did not have a very specific focus on NCHS as an entity or the --
DR. WUNDERLICH: Well, we have --
DR. HUNTER: -- sort of insight of things, and I think that is something that has come about more recently.
DR. WUNDERLICH: You have seen a reference in that report if you have read it, and if not, read it. We made a big point about the National Committee on Vital and Health Statistics. One of our committee members was from here. We pointed out that it can be done by this. If you make major changes, but we were still concerned about the staffing of it.
DR. IEZZONI: Thank you, Gooloo and Ed. Hortensia, and could you introduce yourself. You were not here when we went around the table.
DR. AMARO: Thank you for your presentations. My name is Hortensia Amaro. I am professor at Boston University School of Public Health.
One of the issues that this committee, for a long time, before my tenure, has been concerned about, has looked at over the years, is the issue of the current limitations of national health data systems with respect to informing us about the health status of various ethnic and racial groups. I was wondering, in the process of survey integration, how these issues were considered and what changes that are relevant to that have been incorporated.
DR. HUNTER: Two -- well, two different tracks here. One is, I think it is clearly something that was a concern as to what impact would this plan do to our ability to estimate small population groups. I think one of the options that was considered would have been a -- what would have amounted to a great curtailment of the sample size of the Health Interview Survey.
I think the interest in at least not moving backwards with race/ethnicity data drove a lot of the discussion toward how do we work out this type of linkage plan rather than taking -- I mean, to be more clear on this, one of the options was essentially to not do the HIS anymore and to invest those dollars in a more detailed expenditure survey with a smaller sample that could get more detailed information. I think there were a lot of losses from that that were considered unacceptable, the primary one being what it would do to population groups. So, I think on preventing a downside, that was a clear consideration.
We did talk a lot about race/ethnicity data and the need to make improvements in that. We did not identify new resources that we would need to do that with the HIS or through grants or other mechanisms. Ann can talk a little bit, maybe later, about where we are with the HIS and those population groups.
Since that time, the Data Council has established a work group on race/ethnicity. What do we call this work group, Olivia?
DR. CARTER-POKRAS: [Off microphone.] Working Group for Racial and Ethnic Data.
DR. HUNTER: Racial and Ethnic Data. Olivia might want to comment on that, I mean, some of the specific steps are trying to develop a data collection plan and inclusion policy for department surveys and things. I think there really was not anything else specific in the survey integration plan where we were able to solve any of those problems at that time.
DR. IEZZONI: Olivia, do you want to comment briefly?
DR. CARTER-POKRAS: Yes. I am Olivia Carter- Pokras from the Office of Minority Health. Tony DiAngelo is the other co-chair for this working group. It is a unique working group because we have representatives from each of the agencies and several of the staff offices within the Department of Health and Human Services.
We have been up and running for now -- whether temporary or permanent, we have been up and running for about two years, and we do have a couple of accomplishments that we are happy about. One is that we did have the inclusion policy for racial and ethnic data drafted and accepted by the Secretary and circulated last October to all of the agencies. We are in the process of updating the 1995 inventory and HHS data systems which provides brief descriptions of all of the data systems within the department that are people oriented, in other words, not laboratory animals but people oriented.
We are using this update to find out how well the agencies are doing in implementing this new policy, which requires that they collect racial and ethnic data in all HHS data systems unless they have very good reasons not to. For example, the Indian Health Service is a good example of an exception. Since they are only providing health services to American Indians, members of federally-recognized tribes or their descendants, therefore, we do not need to collect information on Asian Pacific Islanders, on African Americans, for instance. So, it is acceptable for us just collect information on membership and which particular federally-recognized tribe in that particular one.
In addition to that, we have been working with each of the agencies to get input in developing a long-term strategy, and we hope to provide that in December or January to the Data Council. We have developed a list of questions that we have asked each of the agencies to respond, and we are getting some very interesting feedback. Many of the agencies have now started up their own working groups within their agencies, for instance, CDC and HRSA are examples of that. They are using that to not only monitor implementation in the inclusion policy, changes in the revision of OMB Directive 15, but also take a look at what their data needs are and meeting increasing diversity within the U.S. population for changes.
One of the things that we are running into is that there is some limitation in what you can do about block grants, so we are trying to investigate that further. We hope that eventually the subcommittee will be interested in learning what we found out.
DR. IEZZONI: Olivia, I am thinking as you are talking that we should probably have a briefing from you at one of our subcommittee breakout sessions to hear more detail because we just do not have time this morning to hear everything. We are running a little late, and I am hearing privately from subcommittee members that we need to move on, but Paul has one final question. If the response can be brief.
DR. NEWACHECK: Just briefly, you had indicated that the center has suffered a 25 percent budget reduction in real dollars during the 1990s. I wonder if you can comment on how that has affected the center's ability to fulfill its mission. Very briefly.
DR. HUNTER: Well, very briefly, badly.
DR. IEZZONI: Thank you, Ed.
DR. HUNTER: Two quick things. I think the first thing that you would observe is that we are very short on staff. We are below our FDE ceiling. We could hire, but we cannot afford to hire the way we should. We lost a lot of people in buy outs, and that is where we have been able to keep data systems afloat in the field without more money.
I think the other area, if you look at the health care surveys that were the focus of this report, we do not do them all every year anymore. We have six or seven components of which we are lucky to do four or five a year. HANES is not in the field yet and hopefully will make it.
I think in the research area, a lot of the things that we would be doing to sort of push the next level of things, we are not there. We have not been able to really deal with our own money, and there are certainly things like SLAITS where we have done some pilot work with other people's money. I think in vital statistics, we are dropping data items and components. I think across the board, there are things that are not quite noticeable at this point to the outside world, but the inside world certainly is aware of them, and the outside world will be, unfortunately, soon enough.
DR. IEZZONI: Thank you, Ed. I am really sorry to rush on for everybody because there is a lot of important information that we need to learn. Why don't we take a ten minute break and reconvene at 11:10.
[Brief recess.]
DR. IEZZONI: Why don't we reconvene so that we can hear from the panelists and have time for subcommittee questions. If people could come back into their seats. We have three panelists for the next hour, and if it is possible for panelists to maybe limit their comments to 10 or 15 minutes so then we will have ample time for subcommittee discussion afterwards. That would be really helpful.
Dr. Reilly, why don't we start with you.
DR. REILLY: I am Tom Reilly from the Center for Beneficiary Services at HCFA. We want to talk today about the Consumer Assessment -- Medicare Consumer Assessment of Health Plan study. We will try to keep our comments brief. We may run a little longer than ten minutes, but we will try to go as quickly as we can.
DR. IEZZONI: If you are going to use overheads, we need to mike you.
DR. REILLY: Okay.
DR. IEZZONI: The reason is, Dr. Reilly, that we are on the internet, and so we need to have everything be miked.
DR. REILLY: That is no problem at all.
DR. IEZZONI: And there is transcription.
AUDIENCE: There is a lavaliere if you want to use that.
DR. REILLY: We will be all right. Thank you.
Our CAHPS effort -- and CAHPS stands for the Consumer Assessments of Health Plans study. CAHPS was designed to assess consumers' experience with their health plans and the services that they receive. The work has been going on for a while. It began with the Agency for Health Care Policy and Research. The idea of CAHPS is to develop a standardized set of items that can be asked across populations so that it can be asked of commercial populations, for Medicaid populations, as well as for Medicare populations.
AHCPR has been working for some time in trying to develop these questionnaires and items, and we have been working with them to develop the Medicare version. I need to say, again, our colleagues from AHCPR are not here today, but from our perspective, it has been a very productive and very collegial, cooperative effort in developing CAHPS for Medicare, and we have been very pleased to work with them.
We have three parts of our CAHPS effort for Medicare. There is a survey of beneficiaries, Medicare beneficiaries currently enrolled in managed care plans. That is the survey that we have going now. It was in the field under Goldstein as our team leader on the managed care survey, and she will talk about that in a moment. We also have two developmental efforts underway. One is a survey of disenrollees from managed care plans. Again, we want to capture their experience of getting care through their plan and the reason for disenrollment, as well as a survey of beneficiaries in the fee for service sector. Again, those two are currently under development, and we will talk just briefly about those after the managed care survey.
There are several basic purposes from our perspective for the CAHPS effort. The first is to provide information to beneficiaries, to help them make informed choices. This includes the choice between do I want to be in managed care or do I want to be in fee for service, and if I have chosen to be in managed care, which health plan do I want to choose. These activities are required in the Balanced Budget Act of us. There is a requirement for us to provide information of a variety of kinds on health plan performance including satisfaction data, and that is what CAHPS is designed to capture.
We are also going to be using CAHPS data within HCFA to provide information for health plan quality improvement purposes. Again, we want to provide the data back to the plans so that they can get an idea of how their performance rates compared to others in their markets.
A third purpose for us is to provide information for monitoring of health plans by HCFA. Again, we have a regulatory function in HCFA. This gives us information to judge and identify plans that are performing well and plans that are not performing up to expectations. Then we can try to identify opportunities for improvement for those who are not performing up to expectation.
Finally, CAHPS data will be used in some of our reporting for the Government Performance and Results Act. It will be used in several places, but one of our bigger measures is to try to improve the satisfaction of beneficiaries with the health care services that they receive. The agency is being held responsible for improving that satisfaction. The way we are going to track that is through using CAHPS data as it can be wound through time.
We want to take a couple of minutes to talk through the different components of our effort for CAHPS, spending most of our time on the managed care survey because that is where we are right now in terms of getting things into the field. Liz Goldstein, again, is our team leader on managed care caps, and she will talk through some of the things that we have going on there.
DR. GOLDSTEIN: As Tom mentioned, for the overall CAHPS project, they have developed core items that are applicable across care delivery systems and also across populations. For example, they are applicable across managed care, fee for service, and they are also applicable for the Medicare, Medicaid and commercial populations.
In addition to these core items that can be used in all of these different settings and for these different populations, there are certain supplemental items that have been developed that you can use with those surveys to address the needs of some specific populations such as Medicare, Medicaid, children and persons with chronic conditions.
We call at HCFA the managed care survey the Medicare Satisfaction Survey, and that is really just a CAHPS version of the survey for Medicare managed care enrollees. In this survey, we have the core CAHPS items, and they are supplemented with additional items. Some of the items are ease of getting special equipment, ease of getting physical therapy, speech therapy, occupational therapy, ease of getting home health services. We have a series of questions to get at whether the beneficiary suffers from various chronic conditions. We have some questions relating to activities of daily living and instrumental activities of daily living.
The managed care survey right now is about 85 questions, so it is not too long, but it has a lot of information in it. Just to go over some of the major things that are in the survey, there are four overall ratings. One, an overall rating of the experience with the health plan; an overall rating of health care received; an overall rating of doctors and other health professionals in the plan; and finally, an overall rating of specialists. These ratings are scales from zero to ten where zero is the worst possible health plan and ten is the best possible health plan.
For reporting of this information from the survey, the CAHPS team has developed what they call CAHPS composites. Basically, they combine three or four questions on the survey that relate to the same area. For example, there are a series of questions related to getting needed care. There is a series of questions related to doctors who communicate well with patients. There are some questions relating to getting care quickly, to the paperwork, to the customer service in the health plan. So, what they have tried is, in terms of reporting, to try to group similar questions or questions getting at the same idea together so that it can be easily reported.
Just to give you an idea of our schedule for the managed care survey, the first round of the survey went out in February 1998, and currently we are in the field with the second round of the survey. We will be conducting this survey annually from here on. You might wonder why we are going out this year with two rounds of the survey so close to each other. Basically, we realize at HCFA, we really need to get onto a fall schedule for conducting the survey.
The reason we need to get onto this fall schedule, as Tom presented earlier, the primary purpose of this survey is to report the information to consumers. We are going to be producing every fall a handbook that is called "Medicare and You" that is going to have comparative information in it. The CAHPS information is going to be part of this comparative information.
Basically, we have a print deadline of June in order to get stuff into the fall handbook. The reason that the print deadline is so early is to produce 39 million handbooks, and they are specific to the beneficiary's area. It takes an awful long time to put all of this stuff together and also to be able to print all of this information out and get it into the mail by October 15th of every fall. So, that is why we have done two this year and we are moving onto a fall schedule for conducting the survey every year.
Just so that you know who was part of the first survey, all Medicare risk and cost contracts in effect as of January 1, 1996, were included. Six hundred beneficiaries per plan were randomly selected. The beneficiaries had to be continuously enrolled for 12 months in the plan and had to be not institutionalized. In our sample, we had over 136,000 Medicare beneficiaries.
We were very excited. For the first round of the survey, we got a 74 percent response rate. Again, this was the first time we had done this, and we thought that was pretty good.
To let you know how we conducted the survey, we sent a pre-notification letter signed by HCFA's administrator to all beneficiaries in the sample a few days before they received their first mailing. It describes to them that they will be receiving the survey in the mail, the purpose of the survey, and encourages them to complete it. We then send the first mail survey, and that is followed by a reminder or thank you postcard. Basically this postcard is just to thank those who have already responded to the survey and to remind those who have not responded yet to return their surveys to us. We did find a lot of beneficiaries did react to this postcard and send in their surveys after receiving the postcard. So, we found that was actually a very useful tool.
We then sent a second mail survey and then we do telephone follow-up to non-respondents. If for a beneficiary, if we do not have their telephone number, we follow up with a FedEx package, overnight mail, with a third survey.
The CAHPS team has basically developed two ways of presenting the information. One is really bar graphs, and the other is what they call stars. Basically, a plan can get one to three stars. Three stars says that plan is better than all plans within a given area. One star would say that plan is worse than all plans in a given area. Basically, they determine the stars for statistical tasks, controlling for differences in case mix.
We had some concern about these different ways of presenting the information and whether Medicare beneficiaries can understand these different ways of presenting the data. We have done a series of focus groups with beneficiaries and also with ICA counselors to find out do they understand these ways of presenting the data. We also conducted cognitive interviews.
Basically, what we found out is that most beneficiaries in the focus groups and cognitive interviews could understand the bar graphs. I am going to show you one in a second, what it looks like. Most beneficiaries -- almost all beneficiaries except a couple in the focus groups that had advanced degrees could really understand the stars.
At this point, we are going to be putting out some information to beneficiaries. This fall, we are just going to be putting out the bars because we have a lot of concern about how beneficiaries are interpreting the stars. Basically, the stars is a relative comparison. Most beneficiaries were saying this is an absolute comparison, and they kept trying to count up all of the stars on a page if you put different measures. They kept saying, "Where is the four-star plan? I do not see any four-star plans. Something must be wrong with my sheet." We realized those are not working right now, and we need to do a lot more work looking at reporting formats and determining what is the best way and the easiest way to present this information to beneficiaries since that is our primary purpose of the survey.
On the next slide is an example of what the bar charts look like. Basically, this is a composite, experiences with getting the care you need, and it combines three questions on the survey. Basically, the blue part of the bar are the percent who said that they always get the care they need. The red is the percent who said that they sometimes or never get the care that they need. We have an explanation of these bar charts that will be also provided to beneficiaries telling them that they should look for large blue segments, smaller red segments, and also explain to them do not look for small differences; they are really not meaningful and you need to look for big differences across the plans.
Just to let you know how this information will be disseminated to beneficiaries this year and in the future, this year this information in November is going to be going up on Medicare Compare, which is a tool on our website, Medicare.gov, that has plan comparative information. So, we will be putting up some CAHPS information and also some HEDIS information up on Medicare Compare. As I said before, this information later on is going to be part of an annual mailing to beneficiaries every fall. In that annual mailing is quite a bit of comparative information, and the CAHPS information will be included as part of that. We are also providing to health plans some fairly extensive reports about the survey findings for them to use for their quality improvement activities.
For the second round of Medicare CAHPS, we have a few changes in the survey. I guess some of these changes are something that you would be very interested in. One, the survey has changed slightly from round one to round two. We are using a survey now that is based on an effort by the National Committee for Quality Assurance and the Agency for Health Care Policy and Research. They have been working for a couple of years to combined the member satisfaction survey and the CAHPS survey, so we are going to be using this converged survey for the second round. The next slide will explain some of the changes in the survey.
We have also, for the second round of the Medicare CAHP survey, have changed the eligibility period from a six month eligibility period -- we have changed it to a six month eligibility period from the 12 month eligibility period. The reason we have made this change, as I said before, we will be conducting the survey every fall. Eventually, when Medicare has an open enrollment season and lock in, if a beneficiary enrolls in January, we would not catch them in the following fall. We would not catch them in our survey until about 19 or 20 months, and we felt this was way too long to finally catch beneficiaries in our survey. So hence, we changed it to a six month eligibility period.
Just to go into a little detail about the convergence of the CAHPS and the member satisfaction survey. The Agency for Health Care Policy and Research and NCQA have been working very diligently over the last two years to develop a common instrument. They felt it was silly to have two instruments out there and for commercial plans or other plans, they are filling out different instruments, and they cannot compare across their different populations. Basically, their goals in developing this survey is one, it should be composed of the strongest items from the CAHPS survey, the original CAHPS survey, and from the member satisfaction survey. They decided that these strongest items have to be chosen based on the actual survey results, examining all of the data that they had in on both surveys.
Also, the new survey had to reflect the input of all users, so consumers, purchasers and health plans. So, in this process, they got input from everyone into the changes to be made in the survey.
Also, one of the issues that they had to grapple with in developing this new version of the survey is they needed to take into consideration, especially for the HEDIS member satisfaction survey, trend issues, what items would be critical to have information in future years so that they could compare to information that they had previously collected. It was not as big a problem for the CAHPS survey, although it is some problem because it is a fairly new survey. For example, for our Medicare survey, we have only done one round, so it is not as much of an issue.
Also, the new survey should be as short as possible. Basically, I would say that probably like March of this year, they developed this new instrument, and we are using it for our second round. I guess everyone will start using this instrument as of January 1, 1999. It will be nice; that way we will be able to compare the information we collect in Medicare to Medicaid and the commercial population. For the HEDIS requirements, they will be using the same survey, so that will give us a lot of information to compare across populations and across plans.
For the second round of the --
DR. IEZZONI: Liz, maybe there might be some slides that you might be able to truncate because I am looking at the time.
DR. GOLDSTEIN: Yes. Actually, I just have a few more. For the second round of the survey, we have included plans that had contracts in effect as of January 1, 1997. We are also doing an evaluation of the CAHPS survey and the reporting formats in Kansas City, and I will not go into too much detail on that, but you have some information in your slides on that.
Just to spend a couple seconds on the disenrollment survey. That is another survey that is in development. The main purpose of the disenrollment survey is to ascertain the main reasons why beneficiaries disenroll from their health plan. We will also be asking some information on disenrollees' experiences with their health plan so we can combine this with information collected on the enrollees' assessments to get a fuller picture of what is going on in the plan. Just so that you know, this survey right now is in development, and a lot of the work is going to start next month through an interagency agreement with the Agency for Health Care Policy and Research.
Now I am just going to turn it over to Ted, who is going to be talking a little bit about the Medicare Fee For Service Survey.
DR. REILLY: Just very briefly, Ted is our team leader on the fee for service side. Ted, I am sorry. We asked you to do a presentation, but you have to keep it to about a minute or two if you can. [Laughing.]
DR. IEZZONI: Yes, please come to the table, Ted. Thank you. I am sorry to rush you all so much. This is really interesting. It is just that I know we have a number of other interesting presentations to hear.
DR. SEKSCENSKI: I will dispense with my slides. I think you all have copies of them, and maybe we can just run through them very quickly. The timeline for -- I am speaking on the CAHPS Fee for Service Survey. The timeline for the Fee for Service Survey in Medicare is about 18 to 24 months behind that of the managed care survey.
The pilot test is in the field. The responses are starting to trickle in, and we are excited about that. It is in five states. The five states are Arizona, Georgia, Pennsylvania, Washington and Wisconsin. In two of those states, Georgia and Pennsylvania, it is an overlap sample with the Health of Seniors Survey. Largely, that overlap sample is going to be providing information for case mix adjustment purposes for the CAHPS.
There are two different instruments in the pilot. One is an 80 question survey, and that is in all five states. That uses a six month recall period. There is not an eligibility issue in regard to the fee for service as there is in managed care, but the recall period, we are looking at whether a 6-month or a 12-month period recall is better for obtaining a lot of the information. We will probably still go with the six-month recall period in the final CAHPS survey which will be implemented in the fall of the year 2000 to be comparable with the managed care.
The size of the sample, the total size of the sample, is around 3200, and the three non-overlapping states are Arizona, Washington and Wisconsin. We are shooting for about a 70 percent response rates in both of the pilot instrument surveys.
The goals of the pilot are to assess the effectiveness of the survey instruments; to compare the alternative survey items used in version A and version B; to collect additional information case mix, on health status and use of services for case mix adjustment beyond what we will be receiving from the Health of Seniors Survey; to provide information for developing the sampling strategy for the Medicare fee for service CAHPS. There is an analytical plan that RTI, Rand and Harvard Medical School, the three power consortium that has developed the CAHPS for HCFA, with HCFA, has developed, and I will not go into that. That is laid out in the slides that you have in front of you.
In 1999 and 2000, we will be moving forward with analysis of the information that comes out of the five state pilot. We will be looking at reporting mechanisms similar to the reporting for the managed care. We want to make it as much comparable to the managed care reporting. We are not sure how it will be reported yet with managed care, whether it will be separate booklets or featured in the same booklets and so forth. That will all be developed in the focus groups and the testing that will take place after the analysis of the pilot and the data are actually coming in to us.
We will be developing various approaches for providing brochures to intermediaries beyond providing information only to the beneficiaries themselves. We want to provide information to groups that are, as a larger entity, providing information on health plans and fee for service services within a geographic area. We do not know yet the type of reporting unit that we will be using in the Fee for Service Survey. That will be the work that we will be doing with the stakeholders and experts in the Medicare fee for service sector in the states and other areas, private sector. It will assist us in regard to how we will be reporting that.
As I mentioned, the development of the case mix adjustment methodology will be moving forward in 1999 and the year 2000 so that we can implement the CAHPS Fee for Service Survey in the fall of the year 2000.
DR. IEZZONI: Thank you.
DR. SEKSCENSKI: Thank you.
DR. IEZZONI: Interesting, exciting stuff. Sorry to have to rush you. Stick around. Dr. Hardy, we would like to hear from you.
DR. HARDY: I do have some overheads that I do think I would like to go through, so I think we are going to have to --
DR. IEZZONI: Yes, can we mike Dr. Hardy, please.
[Discussions off microphone.]
DR. HARDY: I am going to provide a little bit more detail about the experience of linking the Health Interview Survey as a sampling frame for other surveys. I am going to specifically talk a little bit about our experience with the National Survey of Family Growth or NSFG, a little bit about the MEPS survey and very briefly on the HANES survey. The last survey on this slide, as Ed mentioned, is not one right now that is being considered because of the anonymous feature of the National Household Survey on Drug Abuse is kind of in direct opposition to what would have to happen in a linked survey.
I just want to point out there are various levels possible for linkage. I think the most common and the one that at least the NSFG and the MEPS have been linked at the respondent level where we actually try to get the same people who were in the HIS and these surveys. The HANES survey, because of some of the constraints that Ed already mentioned, the need to do the physical exam component, is being linked at the PSU level, which is our primary sampling unit. So, they will be going into the same areas but obviously not getting the same households or respondents.
This next slide, I think most of you could probably do maybe even better than I can but certainly as well. Just to talk about a few key features of the HIS, it is a survey of the household population. It has been continuously fielded since 1957. The data are collected by in-person interviews. It covers a broad range of health topics. It is a complex design. The sample is actually redesigned every ten years. The sample that we currently have in the field went into place in 1995. It will be in effect until 2004, at which point we will use data from the 2000 census to field a new design. I should point out that that design went into effect before and the planning for that, which actually started five or six years before it, went into effect well before this whole notion of survey integration really took hold, so this was designed not with that in mind, and that does lead to some limitations. Obviously, and we are staring to gear up now with the Census Bureau and some other statistical agencies to think about the next design based on the 2000 census, and we will be able to go into that with survey integration in mind.
Just to talk a little bit about a few features of the design that is currently in effect, we do oversample black and Hispanic persons. We are in every state, unlike the previous design that was in effect from 1985 to 1994. This has meant that we have more primary sampling units, and we are spread out a little bit more, one of the things that is kind of at odds or leads to some limitation in doing these linked surveys.
The sample is also partitioned into subdesigns. Actually, each week in the HIS is a national representative sample. This allows us to cut down on weeks if we need to, such as when the government furlough happened and we were not in the field for a few weeks, without -- it will affect your sample size and the reliability of your estimates but not your ability to make national estimates. We also then partitioned the overall designed into four panels, each of which is representative. This makes it very easy for us. In survey integration, we have not -- we have only wanted to link a particular respondent to only one additional survey, not to several, so we can just choose from certain panels for the various linked surveys. MEPS, for example, chooses from two each year, from two of our panels and two of our half of the year.
I just want to talk a little bit about the National Survey of Family Growth which actually was the first survey linked to the HIS and was first linked to the HIS sample back in their cycle four in 1988. This is a reproductive health survey. It is a very targeted population, women 15 to 44 years of age. Again, it was linked to the HIS in 1988. It used the 1987 HIS sample of women and in their cycle five, which was done in 1995, it was based on the 1993 HIS.
I will talk a little bit about some of the advantages, many of which I think Ed already touched on, so I am not going to go over these in great detail. There is some cost avoidance of having to do the listing and screening, particularly because the NSFG is even more targeted to women of a certain age group and because they also want to oversample minority groups. There is linked analysis possible, and while we discussed some of the problems and the fact that we have not quite gotten there yet with the link between MEPS and the HIS, the NSFG data file in 1995 does include about 75 items from the HIS, so we have been able to work through that and work through some of the confidentiality issues and at least use some of those. Also, I think the HIS income data was used to help impute the income data from the NSFG, at least in 1995, so again, do that amputation(?) with a little more accuracy. Also, because you know something about the people from HIS that do not go on to participate in NSFG, it provides an opportunity to make more accurate non-response adjustments. Obviously, if it was a new sample, you do not have any information on people and are limited on the people who do not participate.
Obviously, while there are advantages, there are also obviously disadvantages. The first statement, while it is true the NSFG is it really true of any of the linked surveys, you are limited by the size and the composition of the HIS sample. There are a couple of things that can be done. I think this turned out to be adequate for the NSFG, but if you need more than the sample that we can provide in any given year, you can link a couple of years. Of course, then you run into the problem that you get longer and longer away from when the HIS interview was done. It is harder to race, more difficult, more costly.
To follow upon on tracing, I think tracing turned out to be more difficult than was originally anticipated, particularly with the last round of the NSFG because it was two years from the time the HIS interview had been done. The target group, which is a fairly young group for the NSFG, is also the group that is more likely to move, about 20 to 25 people in that age range move each year. Also, the information needed for tracing such as name, date of birth, Social Security number and contact, recontact information, at least one of those items was missing from the 1993 HIS sample in about a third of the cases, making it a little more difficult.
Ultimately, they were able to locate about 95 percent of the people, of the women that they deemed eligible from the HIS that they wanted to include, but the tracing efforts turned out to be much more intensive and thereby more costly than anticipated. The response rate for these women that they were able to locate was about 79 percent, which I think is similar to response rates that NSFG has seen in the past.
I think Ed talked in detail about the linkage of MEPS and -- the initial MEPS, the 1996 MEPS was linked to the 1995 or 1994 HIS. The response rate, again this just shows you a little bit of the response rate of the people who were eligible. The response rate was about 83 percent; however, if you combine this with the initial response rate to the HIS, which was about 94 percent, the overall response rate for MEPS for this 1996 round of the household component is about 78 percent.
This is a little bit lower than the 1987 NEMIS. However, I think surveys in general are experiencing some somewhat lower response rates over time. I am not clear how that would have compared if they had again done as they did in 1987 with the sample, the unlinked sample.
Design improvements, I think we have already discussed most of these. There is the hope for improved analytic capabilities by having the two linked. We obviously have not sorted out all of the very real confidentiality considerations on how to do this, particular because MEPS has a great deal that goes into one particular aspect or several aspects into a lot of detail, and there is really the potential for disclosure. S o, we have had to look at this very carefully, but I think we are still hoping and still trying to find ways that both the federal research community and those not in the government can take advantage of this linked -- the link between the two surveys. There is still some savings achieved in terms of not having to do the screening and an independent sample.
I think some of the limitations, I have probably already mentioned, in mentioning the advantages, the restriction, concerns about confidentiality, a somewhat lower response rate, although again, it is not clear since it has been ten years since the last NEMIS was done what they would have seen since I think response rates have been declining. The need to trace people adds, I think -- has been a little bit more difficult than I think we anticipated. Also, I think, too the dispersion of the PSU's. Again, the sample design in effect now was not designed with survey linkage in mind, and the fact that we are in a wider -- a larger number of PSU's spread more across the country has added some constraints to these linked surveys.
HANES, again, I am not going to go into great detail about this, but because of the physical exam component, we are linking only at the PSU level. That will be fielded, I think, next year, and I think in 1999 they will be going to 15 PSUs that are included in the HIS that are the same as HIS PSUs.
Is the next slide on HANES? We will skip it.
I wanted to spend just a minute talking about what kind of in terms of operations this has meant -- linkage has meant to the HIS side. I think we have heard a lot talking about what this has meant to the other surveys but just talk for a moment in really some nitty-gritty detail about what survey integration has meant to or done to us. I think the first is that we have to be in the field every year in January as scheduled. I think we have a great deal of fear about what will happen if that does not happen.
The HIS, I should point out, I think I forgot to mention earlier, has gone from a paper and pencil survey to one that uses a computer assisted personal interview. We did this because KAPPI(?) has the potential to improve data quality and also data timeliness, which was important just to us for being able to get our data out more quickly, but also I think has a real advantage in terms of these linked surveys, being able to provide the information for sampling much more quickly. I think this, in fact, has borne out the initial MEPS round was linked to the HIS before we were in KAPPI, when we were in paper, and we actually had to pull those interview forms off line, as it were, and key them, key the variables needed to start the MEPS linkage separately. That was not something we would care to do all of the time.
With KAPPI, we are getting the data on a regular basis, about six to eight weeks after it has been collected in the field. So, this has really, I think, helped us in this respect. KAPPI, with KAPPI, I think in terms of survey design. It is a much longer lead time. It is not just -- and not to make that sound like a minor thing, but it is not just figuring out what the questions should be and putting them into a paper questionnaire, which we have done for a very long time. It is designing a computer program to have those show up on the screen and making sure, testing that thoroughly.
I think because we want to make sure we are in the field every year, we have built in a little bit longer lead time just to make sure that we have our KAPPI system functioning. This has also caused us, I think, to have a little bit less flexibility in being able to make changes to the questionnaire at the last minute. We cannot do that. I think certainly we cannot make changes within the two or three months before we are going to go into the field. It is just not feasible to get those into KAPPI, tested and with assurance that our instrument will not be impacted.
What this has meant is that for some things, I think with our health insurance question and some of our program participation and welfare questions where we have had to make changes, things are changing. The field is changing. We have changed the question at mid-year which is not so much a problem for operations but may be problems with processing our data when we have changed questions mid- year.
Field operations must run smoothly to produce the sampling information. Again, I think KAPPI has been a real advantage here. For those of you who have suffered through a conversion like we did, it was often to see what were the advantages and why did we do this, but I think for this reason, this certainly has been an advantage.
Again, I am going to skip the third point, and I am going to come back to that in another slide that goes into more methodological issues. There is, I think, too, a limit. Again, the HIS sample size is limited. There is going to be competition for sample. If survey integration catches on an more surveys -- as more surveys attempt to be linked, again, we do not want to overuse our respondents.
We have built in the capacity with the current design so that there is extra sample ready that could be accessed by census. It cannot be accessed in a matter of weeks or a month, but it can be accessed within a few months. We do have that capability. We have not had the need or the funds to tap into that yet; it is there. Again, with a fixed sample size, there is going to be some competition potentially down the road.
Just to talk, my last slide, a little bit about methodological issues. I think the first critical is the next redesign of the HIS that will be based on the 2000 census must give full consideration to this survey integration, something that we did not do the last time because that was all done prior to this being implemented. We are already starting some preliminary discussions with census and some of the other statistical agencies about the redesign of the HIS that should go into the field in 2005.
I think we need to look a little more closely at how well we are doing our oversampling. Again, right now, we are oversampling the black and Hispanic population in this design. I think it looks like data from 1995 and 1996 that we are doing somewhat better in terms of oversampling the Hispanic population than we are with the black population for reasons that are not completely clear to us, but I think we need to continue to look at that. It is assumed that we will continue to do these kinds of oversamples. We really need to take a look at how to do this perhaps more efficiently.
We have to continue to look for ways to improve our collection of recontact information and just look at other types of information, whether we need to collect other types of information that may be useful in following up on our respondents in these linked surveys. We have a cooperative agreement with the University of Michigan to do a -- they are doing a number of evaluation type of activities together with us. One of the things they have been looking at is whether we need to ask some additional questions. We did put a few more questions in the 1998 survey to look at whether people intended to move in the next year and if so where. We will have to look in the future whether that was useful.
They are also going to be looking at several of the linked surveys such as the NSFG and the Disability Follow-Up Survey which was not really a linked a survey but involved going back to previous respondents, trying to look at what types of people, are there some common characteristics of people who do not go on to respond in these linked surveys and can we use that to feed back into perhaps our data collection system to make this a little easier. We are also now collecting information and trying to get information on two recontacts, two contact persons rather than just one that we did in the past, and we will have to see how well these work out and make adjustments as necessary.
I think the last point there, that we need to take a look at assessing new sources of survey error, for example, problems with non-coverage. To give one example, the non-coverage rate of young black males in the HIS is somewhat lower than we would like. If it is lower, if you combine that with the response rate among this group for the HIS and then ability, perhaps more difficulty tracking them for follow-up studies, and then the response rate with that, you may get down to some really what we would all consider pretty unacceptable response rates for some certain small groups. So, we need to look at coverage issues and look at ways to improve those and also more systematically look at the impact of recontacting respondents and evaluate that.
I am going to stop there. If anybody has questions, I am happy to try to answer them.
DR: IEZZONI: Okay, thank you. Those were really great presentations on both sets of presentations. What I am about to say casts no aspersions on those. I know that the subcommittee is very interested in hearing about SLAITS, and that is our next 45-minute panel. We are running a little bit late, so I think we are going to have to severely limit questions here. Does anybody have anything burning? Vince? Vince and Hortensia.
DR. MOR: This is tough. I have 17 questions.
DR. IEZZONI: And could I just ask that people give as brief responses as possible. Vince, Hortensia and Elizabeth.
DR. MOR: I guess I am going to ask about the CAHPS. You mentioned a number of times case mix adjustment. Could you just to respond to what are you adjusting and how do you know it is related to what should be adjusted?
DR. REILLY: There are a couple of issues. One is comparisons across fee for service and managed care if we are trying to help beneficiaries choose between managed care and fee for service, the nature of the populations going into those two segments may be very different in terms of sociodemographics or other characteristics and may not be related necessarily to the kind of care that they are getting there, and we want to try to take those into account. Those characteristics may be correlated with reporting but not necessarily the care, and we are trying to take that into account to the extent that we can, as well as on the managed care side just within making comparisons of the plans, the same kinds of issues can apply. We want to try to remove any variability that may be due to things that we do not really think are related to care but may be related to reporting. They have done a fair amount of analysis to try to sort that out this past year on the managed care side, and I think the items that they have chosen to -- Liz will have to correct me if I am wrong here -- education, self-reported health status, is that right, and the third is if a proxy was involved in helping the person respond, and the fourth I am blanking out at this point.
DR. STARFIELD: You said education, self- reporting, health status and what else?
DR. REILLY: If a proxy helped the person respond.
DR. STARFIELD: Are those the excellent, very good, very poor? Is that what you mean?
DR. REILLY: Self-reported health status. Then the age.
DR. IEZZONI: Thank you. Those were excellent, and I wish we had more time. Yes, they were great. This is frustrating.
DR. AMARO: Regarding the caps, maybe you mentioned this or did not get to mention it, but we know that there are some very key, important factors that impact on quality of care and access to care. Just to name two that I happen to be interested in are gender, race and ethnicity. It would be very important to be able to look at a lot of these questions that you mentioned across those groups. So, my question is, will you be able to do that? Do you have a stratified -- you know, have you stratified by that to make sure you get a large enough sample? Also, is the instrument one that will include people who have English, you know, do not have English as their main language so that you can also get at whether access and quality of services is similar across language groups?
DR. REILLY: I will just kind of briefly --
DR. STARFIELD: Can I just make something clear before you say -- I think it is important to realize that CAHPS is not a population-based survey. It is a user survey.
DR. AMARO: Right.
DR. STARFIELD: As long as you understand that we are going to underrepresent a lot of people right from the beginning.
DR. REILLY: It is a survey representative of the people enrolled in managed care plans.
DR. AMARO: That is right. I understand that.
DR. REILLY: Within the plan, it is going to be difficult to stratify and produce estimates by associated demographics. Samples are large enough to produce plannable estimates, but to go much below that plan level would probably be difficult. Once we pool across plans, though, then we certainly can make various kinds of breakdowns.
DR. AMARO: That is reasonable.
DR. REILLY: The second part of your question in terms of language issues, we are tracking how many people would be requiring Spanish questionnaires to be sent to them. We are recording whether a proxy was required to help the person with the questionnaire, and again, if they were just filling out an English questionnaire, we want to know if there is someone reading the questions to them. I am blanking, Liz. Is there anything that we can contribute on the language side?
DR. GOLDSTEIN: The only other thing is if we are doing telephone follow up, we have information. If they could not complete the survey because of a language barrier and what language it may be, that is probably the only other thing.
DR. AMARO: So basically you will be able to look at these issues, not by plan but across --
DR. REILLY: Right, across plans.
DR. AMARO: My last question was is there anything on ability and satisfaction with behavioral health services like mental health and substance abuse treatment? I notice that there is questions in here about some specialty, specific services, and I was wondering if you include questions on behavioral health care access and satisfaction.
DR. REILLY: When you say "behavioral health care," you are referring to --
DR. AMARO: Substance abuse and mental health treatment.
DR. REILLY: Not that I recall. It is more toward dealing with chronic diseases and so forth with the elderly populations. I do not believe that we have those in there.
DR. AMARO: So maybe some discussion with SAMSA. There might be interest there in collaborating on that. It would be great.
DR. REILLY: Right.
DR. AMARO: All right, thank you.
DR. IEZZONI: Good. Thank you. Elizabeth.
MS. WARD: I would like to ask that we get copies of Dr. Hardy's slides.
DR. IEZZONI: Yes, they were great.
MS. WARD: Do you have plans to go back to evaluate the impact of the CAHPS survey on health plans?
DR. REILLY: Well, you have a couple of evaluation activities going on. One is we are doing a demonstration project -- well, the overall CAHPS effort is doing a variety of demonstration projects. We are the Medicare population. We have a demonstration in Kansas where we are looking at the effect, trying to do an evaluation of the effect there. They are doing evaluations in commercial populations with Medicaid as well. Those are all looking more at the effect on beneficiaries than on the health plans.
We are sort of routinely in contact with the health plans through conducting focus groups and so forth with them to try to ask them, you know, what information do you want from the CAHPS survey and how can we best provide that to you. We will be doing that on an ongoing basis, so it is more of a CQI, continuous quality improvement kind of assessment that we will be doing rather than a formalized impact evaluation on the plans probably.
MS. WARD: Will you be asking them whether it impacted enrollments going up or down because of the publicity over the results of the survey?
DR. REILLY: We can track that in our own administrative records, and we plan to track that sort of thing.
DR. IEZZONI: Okay, Paul tells me that he has a yes/no question, and so does Dan!
PARTICIPANT: The answer is yes.
DR. IEZZONI: Oh, great, as long as the answer is yes. Paul and then Dan.
DR. NEWACHECK: Will the Medicare and Medicaid CAHPS data be released as public use datasets?
DR. REILLY: Oh, boy, you hit a button there. We are certainly providing data back. Well, the first word of business for us is to provide data back to the beneficiaries. There will be summary data by plan to help them choose among the plans. The health plans are very interested in getting the data back themselves for their own quality improvement purposes.
We plan to do a couple of things for the health plans. One is to provide a fairly detailed report back to all plans giving them information about their plans and the other plans in their area. If that does not suit their purposes, we are also going to set up some kind of facility where they can ask for specific types of other tabulations, basically at their own cost, but they can ask for additional things.
Making data available, the person level data available more broadly is a very sensitive topic around HCFA. We will probably set up some kind of protocol or mechanism where people can send a protocol in to us to request to get certain kinds of data for certain kinds of research purposes. We in HCFA will review those requests and make decisions based on our review. I know that is probably not the answer you wanted, but it is --
DR. IEZZONI: Paul, there probably are not many children in your sample yet, so [laughing]. Okay, Dan also had a yes/no question.
DR. FRIEDMAN: The question is not yes/no, but perhaps the answer could be. Dr. Goldstein, you had mentioned the convergence of the NCQA questionnaire with CAHPS, which I think is really terrific. Will that convergence go beyond the Medicare beneficiaries?
DR. GOLDSTEIN: Yes, that converts instrument will be used for Medicaid and the commercial population, for everyone. They will be using the same instrument. Although, for example, for the Medicare survey, we add a few items to it, but we will be using all of the converged items also.
DR. FRIEDMAN: Thank you.
DR. IEZZONI: Great. This is great. Oh, yes, okay, quickly.
DR. BROWN: I just wanted to --
DR. IEZZONI: Just introduce yourself for folks in the room.
DR. BROWN: My name is David Brown. I am representing Ron Mandersheid from the Center for Mental Health Services. I just wanted to add a point to the question Dr. Hortensia asked about the behavioral health component of the CAHPS. There is a consumer assessment of behavioral health that is being developed at HABIT(?). I think they are still working on that; it is not complete yet. We have been involved with that process.
DR. IEZZONI: Very good. Thank you. Olivia, really, really quickly because we have to get on.
DR. CARTER-POKRAS: I just have a quick question about the racial and ethnic data because I am a little confused. I heard education, self-report, health status and age was collected, but I did not know if you were also collecting racial and ethnic data --
DR. REILLY: Yes --
DR. CARTER-POKRAS: -- or perhaps you are linking --
DR. REILLY: It is collected. It is not included in the case mix adjustment model.
DR. IEZZONI: Yes, I am glad that race is not in the risk adjustment. Race should not be in the risk adjustment.
DR. REILLY: Yes, so that data is certainly collected; we are just not using those models.
DR. CARTER-POKRAS: Right. They listed the items in the risk adjustment, and I am glad that race is not there.
DR. IEZZONI: Thank you very much. These were great presentations. Dr. Hardy, if you could work with Barbara so that we could get copies of your slides.
DR. HARDY: Yes, I will also check. My slides actually should say the slides I gleaned from a presentation by Trena Ezzati-Rice that she made this summer at ASA. I would also say that I do not know if they do proceedings from that, the American Statistical Association, but she discussed these. I actually may include all of them because there was a lot more detail than what I --
DR. IEZZONI: That is --
DR. HARDY: -- actually went into which I think would be of use. I also wanted to point out, since I have a chance, that there are a couple of reports design and plan and operations and sample design, two reports on the NSFG which do go into some detail about the tracing and the linkage to the HIS and the success of that and some of those that people might be interested in.
DR. IEZZONI: If we could get copies of those, copies of the different colored reports that come from NCHS.
DR. HARDY: I am happy to leave these two here if that would be useful.
DR. IEZZONI: Okay, so we can get the exact citation number. Again, these were great presentations. My rushing you has nothing to do with our lack of interest.
Can we have our next panel come up. I have been asked, while you are coming up, to just remind you that we are especially interested in integration issues. Marcie Cynamon will be taking the place of Trena Rice.
All right, Dr. Griner.
DR. POWELL-GRINER: Yes, I am going to forego my slides, so my presentation may be a little more disjointed than I would want --
DR. IEZZONI: Thank you.
DR. POWELL-GRINER: -- so that we can move ahead.
DR. IEZZONI: But could you leave your slides with us?
DR. POWELL-GRINER: Yes, I will. I will. Also, I will have a paper in the ASA proceedings on the uses and limitations of the BRFFS dataset, and I can also provide a draft to anyone who is interested in that.
DR. IEZZONI: Yes. Thank you.
DR. POWELL-GRINER: All right. I think that a couple of things need to be said right at the onset, and probably the first thing about BRFFS that everyone needs to understand is that unlike all of the other surveys that you have heard about this morning, we do not have a central data collection. We are in a cooperative agreement with the states and therefore the principle focus of cooperation is with the states, and the secondary focus then would be with other federal agencies. So, to some extent, the issue of survey integration perhaps is less applicable at this point in time for us than it would be for some of the other surveys.
That said, I would like to point out that many of the questions on the Behavioral Risk Factor Survey in fact come from existing surveys such as the NHIS. In reviewing for this presentation, I actually went through the NHIS in some detail, and there are quite a number of questions that we have literally picked up from that survey where wording is identical, coding categories are identical. There are many other questions which are not identical but the intent is very much the same. So, in that sense, we do have some integration going on.
The other is that CDC, in general, has made an effort to standardize their core elements of data collection, and BRFFS is conforming to those, particularly on issues such as how do we collect race and ethnic information. Again, I think that this is related to the integration process.
The BRFFS, for those of you who may not be all that familiar with the survey, actually began in the early 1980s as a point in time survey. By 1984, we had moved into a surveillance system operation. We had 15 states. We have grown tremendously, and I think that one of the handouts that are being passed around is actually a sheet showing you exactly how much the survey has changed over time. We have gone from having initially 6 topics in 1984 to having 15 in 1998.
The number of questions on our core, which is the part of the survey which is unmodified by any state using it, has grown from 47 to 80. Initially, there was no provision for what we call optional modules, which largely are directed at chronic disease, behaviors and things connected with them. Now, we are supporting 22 modules. The states may choose to use none of those modules; they may choose to use as many as time permits. However, if they use a module, they are not allowed to modify it in wording nor order.
In terms of state questions, this is really an essential part of our survey and again is something that sets us apart from most other surveys that you have mentioned here. We actually reserve a fairly large section of our file for the states to use to collect questions that they, individually, are interested in. There is no attempt necessarily to standardize those questions. Often, the questions that the states will select will be driven by programs which are helping them fund their survey.
This is a David survey; it is not a Goliath survey. We have a very small staff. We have an extremely limited budget. We do not have a line item budget. Each year, we have to go out hat in hand and collect monies to collect our survey. So, this is a real limitation that we struggle with every single year.
We began in 1984 with a single survey design which was based on the Waxburg random digit dialing design. Over time, we have moved away from that and now are in a disproportionate design. We are ceasing to support Waxburg as of data year 1999. In terms of interviews, to give you a taste of how much our file has increased, we have gone from having around 12,000 interviews per year in the early to mid-1980s to now having over 135,000 interviews. For states, we range somewhere in the neighborhood of anywhere from 1500 to up to 5,000. In part, the states which do larger sample sizes are responding to their needs for information at the local area.
Something else that has arisen over the last few years is what we refer to as greater use of geographic stratification. Generally, that means that states are collecting data at the county level. Some states will choose to collect it at a regional level. I think that there would be much more interest in doing local area estimates if we could get adequate funding, but funding has been the real stopping point for getting information at the county or city level.
The goal of the BRFFS is that we want to have a set of data coming from the core items which is comparable across all of our states. Secondly, we want to make sure that state needs are being met by providing them with an opportunity to center their state-added questions around their own particular needs.
One of the strengths of the BRFFS is that it is very timely. This year, 1997 data was processed within seven months of final collection. We have a CD-Rom that should be produced next month. The CD-Roms are widely available. We release them through GPO; they cost $15. That is a real bargain.
It is also flexible. It is flexible in a couple of senses. First of all, we reserve up to five questions on our core each year for health issues that we refer to as emerging issues. An example of this would be when AIDS came to the forefront, we were able to immediately add five questions about that condition onto our core item. The rule is that those emerging questions stay on the questionnaire for only one year. Only after that will we continue to have them if they are cognitively tested and also now, in our new environment, if there is also some kind of financial support for having them.
It is also flexible in the sense that the states are able to incorporate new questions very rapidly. An example of this would be after the Oklahoma bombing, Oklahoma modified its questionnaire so that it could ask a series of questions about the effects of this on the population.
The BRFFS is used in many different ways. Again, the states rely very, very heavily on it to produce reports, and almost all of them produce annual reports each year. It is not unusual for them to also produce topic-specific reports, for example, on cardiovascular risks. It is used pretty heavily for media purposes. We get a lot of publicity for BRFFS when we release reports through the MMWR on particular topics. Some of the favorites are obesity. You may remember seeing the cover story on Newsweek about obesity, and that used some BRFFS data.
We have seen a dramatic increase in the use of our data in terms of refereed journal articles. Now, we are averaging about 18 per year. We have, at our count, a little over 200 reports and articles that have used the data. We also have about 30 studies that have been done on methodological aspects of BRFFS, comparing our data with other data sources. For example, one state used the BRFFS data on seatbelt usage with what they actually observed.
So, in general, the data appear to be very reliable and valid. There is, of course, some problems with overreporting of certain elements, for example, mammogram. You will get higher self-reporting there than you will find on clinical records, but within those types of constraints, generally the data appears to be useable and in fact to be used pretty widely.
There are some caveats, though, in BRFFS, and I will go through those in very brief fashion. One of the things I know that has been the concern of a lot of people is the issue of coverage. Nationally, we now have about 95 percent of households with telephone coverage; however, that varies considerably across states. Some states have as low as 87 percent; some states have as high as 98 percent. It also, of course, varies by some population within states. For example, we know that whites typically have about 5 percent non-coverage compared to about 15 percent for non- whites, compared to about 17 percent for American Indians.
So, obviously, our data is going to be less reliable for some populations than for others. Again, if what you are looking for is the ability to pinpoint areas of health concern that need some kind of intervention rather than a count of the exact number of persons with a condition, BRFFS serves you well. If you need an exact count, we do not serve you very well.
Another problem that we have or potential problem is not unique to BRFFS but is shared by any telephone survey, and that is that the data is self-reported. We have no way of verifying it either by visual means or through actual physical measurement. We, as any other telephone survey, are having problems in seeing our response rates decrease. We are seeing changes in technology surrounding telephone surveying. The increasing use of gatekeepers such as caller ID, answering machines. All of these kinds of technologies are having a real impact on us.
Also, the advent of telephones as more a personal device rather than a household device, where you have cellular phones, you have multiple lines coming into a household. Some are related to fax machines; some are related to computer lines; the increasing tendency, particularly of the wealthier families in the United States, to have separate lines for their children all complicate telephone surveying methodology and is something that we and others have to take into account.
Something that you may hear about BRFFS is, well, your items are not really comparable. That is not really that much of a problem. In general, on the core and the module items, states comply with the wording. In point of fact, we have one state which deviates from two sections of the core. One is a wording deviation having to do with mammography in which their lead-in is different. The other is potentially more severe; it affects the HIV section in which they are changing the population. Instead of surveying through age 64, they stop it at age 45. This is, in general, a minor problem.
Again, there are some variations in response rates, but we know that a lot of that has to do with demographic, cultural and other characteristics of the state population. Also, we know that it has something to do with the telephone characteristics of particular states and also to the sample designs that states are using. For example, some states will use list assisted samples; other states will not. So, all of that affects the response rates.
Now, I actually did not know very much about this committee, and so in preparation for appearing went back and went to your website and read some of the minutes of your earlier meetings. In some of those minutes, I saw a statement to the effect that we have around 55 percent response. Response depends on how you define it. If what you are looking at is the number of completed interviews overall for members called, which for our survey exceeds 500,000 telephone calls, response rates will be relatively low.
If you look at response in terms of those people who are contacted, those households who are contacted, how many completed interviews do we get, we are getting about 75 percent. About 24 percent refused and about 1 percent are terminating within the interview. Of all of the numbers that are called, around 60 percent are non-working numbers or numbers that ring no answer, or they are numbers that are not private residences.
Another variation in our survey which again a survey like the NHIS would not encounter as there are differences in data collection, most of our states use CATI, a computer-assisted telephone interviewing system. We, in fact, in the behavioral surveillance branch, do programming for over half of the states. However, we do still have two states who use pencil and paper, and we have a couple of states who use CASS(?), so there is some variation in that.
We also have a bit of a difference in terms of the extent to which small area estimates might be possible. Generally, those are going to be restricted to those few states that are doing very large sample sizes. We have had two states that, in response to their local area data needs, have moved to collecting county level data in a BRFFS-like survey, but it is not part of the main survey itself. That has been very, very expensive.
I am going to stop in just a moment, but there are two things that I would like to point out as good uses of BRFFS data and one of the reasons that I believe that it is important to maintain BRFFS as a viable data system. That is, we supply the states with most of the information that they need for Year 2000 objectives. I have included in your handout a list of those objectives, and you will also see in the folder a table of contents of our summary prevalence report which is a report we produce each year for every state that is participating in our area that lists Year 2000 objectives and how those are measured.
I would also like to point out that recently, in March of this year, there was a report on measuring health objectives and indicators, 1997 state and local capacity, which indicated that of all of the sources that states used for measuring their health objectives. Whether it is Year 2000 or whether it is their own established objectives, 21 percent of that information is coming from the BRFFS dataset.
DR. IEZZONI: Thank you. Very interesting, but we have to move on to Marcie. Stick around Eve, because we will probably have some questions for you. Marcie, again, we are especially interested in the integration side of SLAITS. Paul, did you share the letter with Marcie that we wrote to the Secretary? Yes, okay, so you know that our subcommittee has a specific interest in SLAITS.
DR. CYNAMON: [Comment off microphone.]
DR. IEZZONI: Are you wired for sound, Marcie?
[Discussion regarding microphones.]
DR. CYNAMON: All right. SLAITS is probably one of the most integrated surveys you all hear about. We are integrated at the sampling frame level and at the questionnaire level. We are capable of doing small area estimates as well as state level estimates, and it is all administered by telephone. The primary goal of the SLAITS mechanism is to help fill data gaps about changes in the health care and Welfare systems at various levels, and we can track and monitor changes over time.
We are integrated into the National Immunization Survey, which I am not sure how much you know about. I will tell you a little bit about it. It is a list-assisted, random digit dialing survey that has been in existence since 1994. The main focus of it is to provide state and local estimates on a quarterly basis of children aged 19 to 35 months on immunization status, on quite a number of different immunizations. The methodology is computer- assisted telephone interviewing. As I said, they have quarterly estimates. The NIS breaks the United States into 78 reporting areas. All 50 states are represented and, including DC, 28 metropolitan areas. This can be augmented very easily with adding from the telephone genesis which does the sampling frame for the survey.
The sampling frame is absolutely enormous. Because children aged 19 to 35 months only represent five percent of the population, a lot of telephone calls have to be made. We start with 2.4 million calls per -- telephone numbers per year. The automated dialing system rules out the non-working phone numbers and ends up with 1.6 million. Then, these houses are called and screened to determine whether they are businesses or whatever and identify if there is more than one line per household, that sort of thing. We end up with just under a million households screened for children 19 to 35 months of age. That sample is what is available to us, just under a million.
The NIS only does interviews in 36,000 of those households annually, so they end up hanging up on close to a million people without pursuing any sort of health data. So, this is where we stepped in. SLAITS, using money through the evaluation fund process from the department and from CDC has developed some questionnaires to be administered. We have done a couple of tests that I will tell you about, but let me first explain something about the questionnaires that we are using.
We have developed a health module that is largely based on the HIS. In fact, it is more like a mini-HIS. It covers the basic areas of access, utilization, some unmet need, health insurance coverage, and it was household based similar to the HIS, 22 minutes in length, unlike the HIS.
We have also developed a questionnaire on child welfare and well being that we just called the welfare module. I will tell you a little bit more about its content, and it is drawn from other surveys. It has a Medicaid status matching component where we are linking back to state-based Medicaid records to assess people's ability to accurately report their Medicaid status. These two modules can be picked apart into subcategories and reassembled based on whatever is needed.
We have also developed and distributed for comment widely a questionnaire on children's health insurance coverage and related health insurance issues such as access to care, barriers, health status, health care utilization, et cetera. This is designed to allow states to monitor the uninsured, changes in the uninsured, especially with regard to CHIP(?).
There exists also the possibility to customize questionnaires should states want to. We have spoken with a number of different states that are interested in small issues pertinent to their states, and that potential exists. We have not realized it yet, but the potential is there.
We are talking to one of the offices at CDC about implementing a module on behaviors related to HIV risk that goes in a lot more depth than the BRFFS module does. We have also just recently started talking to an office at CDC about using SLAITS to come up with state estimates of asthma prevalence.
Some of the key features of SLAITS are that it is centrally administered. That means that we have one contractor, and we control all of the methodology that is used, which allows us to easily compare states, compare estimates over time, and combine all states to come up with national estimates. We adjust for non-telephone households. One of the reasons for basing our questionnaire on existing surveys, especially surveys that are in person and cover all states, is that we can use comparable data to make these adjustments. We have rapid implementation and timely release. We have worked out an arrangement with OMB where clearance can be done in a matter of weeks rather than long months.
We are also studying dual frame estimation techniques where some of the interviews can be in person and some over the phone. We are looking at ways to combine with the HIS to enhance their estimates as well as ours. These methodological aspects are being more closely examined through the use of a contract that is underway now.
The health module that we used drew from not only the Health Interview Survey but from SIP, from the National Household Education Survey, the National Survey of Family and Households, and the new standard demographic questions. So, not only did we integrate questionnaire-wise with HIS but we had several others from which to draw, and we benefitted greatly from that.
The areas, as I mentioned briefly, but they are health insurance coverage, health status, self-perceived inactivity limitations, access to medical care, utilization of services, program participation, and demographics. Here are the response rates.
DR. IEZZONI: We will want copies of your slides, so if you could leave those with Barbara.
DR. CYNAMON: I will. Our response rate, I just want to say, when you see this, for the health module, the response rate using the CASRA(?), which is one of the more restrictive rates, it was only about 68 percent, but we were running out of money, and we decided to cut off interviewing several weeks early. We are confident that we could have worked some of those more difficult cases. My next slide would show you that our findings were comparable to CPS. Then I have another couple of slides, characteristics of the uninsured from that first module.
The welfare or child well-being and welfare module that we are doing targets households with children under 18. Half of the sample will be households at or below 200 percent of the poverty level and half above. This sample will be larger than the one that we used for the straight RDD health module.
The content includes family composition and stability, program participation, cash and non-cash benefits, employment and earnings, early education experience, child care, health status and insurance, access to medical care, food, security, neighborhood characteristics, psychological well-being, and socio- demographic characteristics.
An aside to this, the NIS uses the AT&T language line which is a service that AT&T provides at a certain cost that engages AT&T interpreters. If we can find a household and identify which language they prefer the interview to be in, we link up a three-way type of conversation with the interpreter. This worked for the health insurance module. We had some problems with getting this passed through our IRB for the child welfare and well-being module because this module has some perceived sensitive questions about turmoil within the household, and the AT&T contract requires that if an interpreter has reason to suspect any illegal activity in a household, they are required to report it, which directly contradicts NCHS privacy and confidentiality regulations. We are currently working on that; I do not know that we will have it worked out in time to do the child welfare and well-being, which would be a bigger problem if we were not doing it in Texas. Texas is one of our sites for the welfare module, and we do have a Spanish translation, so we will not miss that population, but we will miss some of the other rarer languages.
I have talked about this mechanism as though we are full blown and going, but I am kind of mixing up potential with actual. We have done two states for our health module test. We are definitely doing Texas starting in about one month with the child welfare and well-being module, and we hope to add Minnesota; we are negotiating with them right now. The reason this negotiation is going on is because one of the components of the child welfare and well-being module is this Medicaid -- matching to Medicaid records that the state maintains. We want to assess the quality of Medicaid reporting, and we have to engage quite a number of different offices and people at the state to get this to happen. It takes longer than you can imagine to do. Plus, the contracts have to go through the individual Office of General Council, the State, and that takes a while, too.
Current and future activities include continued analysis of the health pilot module, which is a public use data file, two files, one for Washington State and one for Iowa; implementation of this welfare module in these two sites. We are also working with evaluation funds to conduct a SLAITS test on an Indian reservation. Those plans are in the works right now. We are probably going to implement this in Alaska using the health module.
We are constantly trying to strengthen our partnerships and collaborative arrangements with states. Once we use up our evaluation money, we are not guaranteed any additional funds. We would like very much to have baseline funding for SLAITS to be an ongoing mechanism that states can add to or that we can add to or the department can add to. Right now, we do not have the promise of that, but we are seeking other means of support. We are continuing to consider ways to customize the questionnaire that so that we can enhance the capability and the cost sharing promotional aspects.
That was all I prepared, so were --
DR. IEZZONI: Great. Both of your presentations were wonderful.
DR. CYNAMON: -- there other integration aspects?
DR. IEZZONI: We are going to have some questions. Dan?
DR. FRIEDMAN: Thanks to both of you. Let me just start out very quickly with a moment of full disclosure which is that we have been very enthusiastic about the development as SLAITS. As Marcie knows, we have been interested in participating as a pilot site.
DR. IEZZONI: "We" being the State of Massachusetts.
DR. FRIEDMAN: "We" being the State of Massachusetts, yes. On the other hand, BRFS, for us, is really a core of our public health decision making. We actually, in 1999, we are going to have a sample around 7,000 translated into five or six languages. So, it is absolutely essential to what we do. One of our employees is Marchella Ponsi and a close relative of the person who is known -- who developed the Ponsi scheme. We often feel like we are selling and reselling the Brooklyn Bridge. That is how we finance it. We are well aware of the budgetary problems around the BRFS.
Clearly, each of them have major strengths. Certainly, for us, the BRFS, the flexibility of the BRFS, the process of collaboration with the states for us has been a real model of collaboration with CDC. SLAITS has some obvious major methodological strengths. Long term -- having said that, and I sort of apologize if you feel like you may be getting set up here, but long term, it is really difficult for me to imagine how or why we should have two independent state surveys.
DR. POWELL-GRINER: I think Marcie and I both will respond to that. I do not believe that SLAITS and BRFFS have to collect the same types of data. I think that there is room for both of them, and there is a real need for both of them. SLAITS can do things because they are going to collect data centrally that BRFFS would never be able to do. Because SLAITS is going to collect data centrally, they are not going to be able to do things that BRFFS does very well, which is work with the states.
We do not require that states standardize, for example, their state-added questions. I was asking Marcie about this earlier, and I do not believe that they have yet come to a resolution as to what their plans will be for handling state-added questions, but it would seem to me unlikely that they are going to be able to allow as much variation as BRFFS does at the present time.
Also, I think SLAITS is in an ideal position to survey populations that BRFFS is not set up to deal with. For example, children. For example, low income persons. These special populations where there are a number of issues and a number of policymakers very, very interested in data that BRFFS simply cannot provide. So, my hope would be that we not look at these two systems as competing systems but as complementary systems, and that we do try to work to integrate them as best we can so that we are not duplicating efforts but are really adding to one good telephone data system.
Now, whether or not there is money to fund both, I do not know. BRFFS is a small potato survey. We are under $7 million total, and that includes the costs of CDC personnel. Out of that, more than half of it goes directly to the states in state grants.
DR. IEZZONI: Marcie, do you want to weigh in on that?
DR. CYNAMON: Ditto.
DR. IEZZONI: Thank you!
DR. CYNAMON: Was that short enough? No, I just feel very strongly that just like there is more than one in-person health survey, there can be more than one telephone health survey. They meet different needs and can use different populations and different sorts of questions. I think that is very important to note. For BRFFS to do what we are proposing to do would require a complete revision of the way they approach the household. It would require rostering families and selecting from within that, and they do not do any of that right now which allows them to be so fast and concise.
DR. POWELL-GRINER: Nor do we have plans to do that at this point in time. We do not have the resources to do it.
DR. IEZZONI: Dan's brow is furrowed.
DR. FRIEDMAN: No, no.
DR. IEZZONI: We are going to turn to Paul. Paul, do you have a quick question?
DR. NEWACHECK: Yes, I wanted to ask Marcie about how quickly the SLAITS could be operational for purposes of monitoring the State Children's Health Insurance Program. As you know, we are interested in that in the committee. You indicated earlier in your remarks that there is now a fast-track procedure for SLAITS telephone questionnaires through OMB clearance. Could you talk a little bit about that and how fast you think it would be to get out into the field if money were available?
DR. CYNAMON: When we went to OMB the first -- for initial SLAITS clearance, we wrote the package and included a pool of questions which drew from about 12 different surveys, lengthy surveys, and said please, almost all of these have been cleared through other OMB reviews, please clear this whole package and let us draw questions from this and just keep you informed. We would send a letter to OMB and the department saying this is one we want to start and this is what the sampling frame looks like. We asked for ample burden to be able to do this in the first couple of years. Our time, our clearance is drawing to a close, and we will be asking for another three-year generic clearance that would allow us to do a similar sort of thing.
DR. NEWACHECK: So, if money became available tomorrow, how fast could you get down [off microphone].
DR. CYNAMON: Three months at most.
DR. IEZZONI: That is pretty quick. Hortensia, you had a question?
DR. AMARO: Thank you for those presentations. I guess there have been a lot of hopes hanging on SLAITS, and one of them that we have discussed previously with Dr. Sondeck is that because of the inadequacy of the other health surveys that we have and their ability to really reach all of the racial and ethnic minority populations that are not evenly distributed throughout the country, and so the typical sampling strategies do not work very well. One of the thoughts way back in some of our previous conversations had been perhaps SLAITS would really be a good mechanism for getting information on racial and ethnic groups because they would just naturally bear out more, larger samples of racial and ethnic groups at different states and that through that we might be able to get some of the holes filled that we have had in terms of getting health data for those groups. Can you tell us a little bit about how that has sort of panned out so far in your field tests in terms of representation from different samples and whether at this early stage you think that is still sort of a hope that we should have for this survey?
DR. CYNAMON: The health module we did, the demographics did mirror the general population, so we were pleased with that. Because the sample size for NIS is so large, we are capable of screening at the household level for minorities. As it turns out, most of the interest in SLAITS has been for rarer populations, be it uninsured children or people with asthma, whatever. This takes a lot of screening, and it is expensive. SLAITS is going to look a lot more expensive than BRFFS, but they are not screening households for rare events, and we are. Minority status, obviously, would be. The capability exists to do that.
We are also looking at using the genesis data file in combination with census track data to come up with the ability to identify minorities in geographic areas and use area code information to help us more efficiently sample minorities.
DR. IEZZONI: Marcie, are there any written documents about SLAITS for some of us on the committee that have not seen?
DR. CYNAMON: Sure, general types of write-ups?
DR. IEZZONI: Caroline, could you talk with Marcie afterwards and see if you can get some kind of summary descriptions of it for subcommittee members. Olivia, you had a quick question.
DR. CARTER-POKRAS: Yes, a real quick question. I was wondering whether SLAITS is going to be able to provide us with a national estimate of CHIPS enrollees? Otherwise, are we going to have to go back to each of the states to get that information? Are we going to get sort of like an aggregate estimate?
DR. CYNAMON: If we were funded to do that, yes, we could. I mean, that is our plan. We would like to be able to do that. We have, as I said, this questionnaire that monitors children's health insurance status. Even with our large sample size, CHIP enrollment is still going to be a very small proportion of the population. We are proposing to use fairly large samples in each state to screen for this, and we think we can do that. Now, to describe in depth the characteristics of CHIP enrollees, I think you might prefer to use enrollment data. What SLAITS can give you that enrollment data cannot is the dynamics of the uninsured which is the target population.
DR. IEZZONI: Okay, Barbara, quickly.
DR. STARFIELD: Just a clarification. The estimates of CHIP eligibles is about three million children nationally.
DR. IEZZONI: Thank you. That really makes sampling -- that will be a needle in a haystack. Yes, quickly.
DR. BROWN: I had a question for Eve. For mental health, you have only one question that has to do with the status of people and their mental health. I am wondering, for the future, do you have any plans to include mental health in the core questions?
DR. POWELL-GRINER: It is true that we only have the one question on the core. We do have a module that is social context which looks at a number of items that would be related to mental health. There is always the option for programs to develop modules for inclusion in the BRFFS. That process is, first of all, the program has to develop the questions. They have to be presented to our state at our annual conference, and then they have to be cognitively tested to make sure that they are picking up useful information.
As I said, this year, for the first time, we are requesting that programs that want questions in modules or on the core provide some kind of financial support for that, but there is nothing to rule it out. You also have the option of working with the states to put it into a state added section. We are doing that with a disability special study this year and also next year.
DR. IEZZONI: Thank you. Thank you guys. I am sorry that I am so informal. Thank you very much for those excellent presentations and for coming to speak with us today. As you can tell from the interest about the table, we will probably be back in touch with you, and we will certainly be following the progress of SLAITS.
We are a little bit behind schedule, folks, so we are supposed to reconvene at 1:15. Could I ask people to go upstairs and bring your lunch back here. What we will do after lunch is kind of talk among ourselves about whether there are any recommendations that would flow from what we have heard today that we might want to make to the Secretary and then try to get a handle on the work plan for the subcommittee. Thank you very much.
[Whereupon at 12:55 p.m., a recess was taken until 1:15 p.m. that same day.]
DR. IEZZONI: Okay, we are going to try to reconvene at this point. Thank you. People are eating their lunch, but the sooner we start talking, the sooner we can all get out of here, so I think that is probably a good thing.
I have asked Cathy to move over there, and when Hortensia comes in, I might ask her to move maybe next to you, Barbara, because I think this part of the meeting is going to be more of a conversation among us. It will be easier for people to kind of interact if we are organized that way.
Basically what we had on the schedule for after lunch was something that we called a Panel Discussion on Survey Data Needs and Gaps in the Data for Minority Populations, SCHIP, States and the Disabled. I do not know. I think we talked about the SCHIP fairly extensively in Paul and Barbara's preparation of the letter that the full committee agreed to send to the Secretary yesterday. I think what I would prefer to do, if everybody is agreeable, is at this point just think about what we have heard this morning, which was actually very, very informative, and just think -- help us think through whether there is anything that we want to do based on what we heard this morning such as writing a letter to the Secretary suggesting that she ask for more money from Congress, or if there is any other specific suggestions from this morning's discussion. Dan, do you have any thoughts?
DR. FRIEDMAN: I remain concerned about the BRFS/SLAITS issues, both from the point of view of sustainability as well as from the point of view of some, frankly, common sense. I do not know if anybody else feels that way, but I would be interested in having somehow or another some follow-up especially on that issue.
DR. IEZZONI: Elizabeth, you are nodding, and you are also chewing. Sorry to call you when you are chewing.
MS. WARD: I think logically, you just look at the two and say this will be hard to sustain over time, but I also know, being in a state that uses BRFS a lot, that we have reached a limit of the number of questions. We are still feeling a constant pressure to develop more modules. If we go much farther, we will have not a single person being willing to stay on the phone with us. We have two things, and that is why SLAITS looks very good because there is a way to divert potentially some of those additional module requests that are coming too. Again, unless you can somehow describe them as being a piece of something, two pieces of something bigger, we also are going to run into a problem.
DR. IEZZONI: Dan, you had something.
DR. FRIEDMAN: I know this is not the place to get into that substantive discussion, but let me very quickly say that we can think about samples, we can think about using the same sampling frame, et cetera, et cetera, so I think there are ways of doing that. We have the same problem with the BRFS. We are -- we have so many balls in the air that they are really hitting us in the head now.
MS. WARD: Right, and I think a national committee like this directing that kind of comment from a scientific analysis of what is feasible from people who are not in the programs.
DR. IEZZONI: Yes, not advocates, in other words.
MS. WARD: Because I understand how defensive I get about my own sort of things of my own.
DR. IEZZONI: That is an interesting idea.
MS. WARD: But to get other states involved in the comments, I think there is potentially a solution.
DR. IEZZONI: Paul?
DR. NEWACHECK: I agree with what we have been talking about. I do think, though, that the two surveys have different purposes or different utilities. Particularly, I think the SLAITS survey is well designed for the purpose of looking at rare events very quickly with rapid turnaround on national issues as well as providing state level information. BRFS has much more utility for state purposes and can be customized easier for state purposes as well. So, I do think that they do have distinct advantages and differences in terms of their utilities. I agree, though, with what Dan has said, that there may be some potential savings that could be incorporated in using say a similar sample frame or the same sampling frame, a uniform questionnaire administration system or something like that. So, there may be some savings, but I would hate to lose what we have in terms of the advantages of both of those surveys as they are now being used.
DR. IEZZONI: Yes, Vince.
DR. MOR: I think on this same subject, both of these surveys end up going hat in hand to a variety of different federal, state and other sources on an ongoing basis. Perhaps the BRFFS now has the advantage of it being so form-fit and ongoing serving the states whereas, as Paul mentioned, the SLAITS has a real opportunity to be responsive. In that responsiveness, the unit cost is going to be higher, and it is going to therefore sort of have to be at readiness, so to speak. The question is, there is a difference between having sort of maintaining some infrastructure to be able to be ready at the response, and then someone has to pay for surveys on a sort of as-needed basis. I think the role that this committee can serve is perhaps looking to or talking about the interrelationship between federal data and non-federal data needs, that this kind of survey's potential might meet.
I am particularly encouraged by the example of HCFA. They talk about having the caps, lots of AHCPR funding, finally integrated a little bit with the NCQA structure and modules, which have largely come totally from the non-federal, almost in some sense from the commercial sector. This might not be a bad example of a context in which some integration could be brokered between a federally-led but state-pertinent data collection effort that might meet both public as well as not so public needs.
DR. IEZZONI: Okay, practical. Aaron, can I just push the committee, because we have to get a work plan in place. Is it --
MR. HANDLER: I do have a comment.
DR. IEZZONI: Okay, quickly.
MR. HANDLER: I have been with Indian Health Service 24 years, and I have been waiting 24 years to make this comment.
DR. IEZZONI: I could not deprive you then.
MR. HANDLER: The Indian population in the United States comprises about eight-tenths of one percent of the total U.S. population. The Indians living in the area that my agency has responsibility for is about six-tenths of one percent. No matter how you oversample, your oversampling will not be enough to capture all of those American Indian people. Someone mentioned earlier that 17 percent of American Indians do not have a telephone at all. On the Navajo reservation, that is about 50 percent, and that is the largest tribe we have, about 150,000 people. Half of them do not have a telephone.
Now, what do we do. Ten years ago, my agency organized a patient registration system where everyone, to get care, has to register. That is a perfect sampling frame, 1.3 million people with names, addresses, telephone numbers when they have them. I think my agency would make that automated list -- it is an automated list -- available to some organization who would like to use that as a sampling frame. We have not used it ourselves; all of our money goes into patient care, and statistics is a by-product like stepchild. We do have that possibility. If someone wants to use it, we could work it out.
DR. IEZZONI: That is something that you might want to bring to the attention of the Data Council, the HHS Data Council, because there might be interest there. It sounds like a very good possibility.
MR. HANDLER: Tony DiAngelo in my office is a member of that.
DR. IEZZONI: Okay, great. Now, back to the committee and the work plan. I am trying to be very concrete here. Wait, wait, let me finish. Am I hearing around the table that we want to have a hearing, we want to have a meeting, we want to have a report, we want to have specific follow-up about -- I am just going to phrase it for convenience, SLAITS versus BRFFS, not that it is necessarily that bald. Am I hearing that is what people would like to do?
DR. STARFIELD: Can I just make it -- I think the issue is pretty simple, that we just make a recommendation that the Data Council consider the potential in somehow merging at least the sampling administration of the two surveys.
PARTICIPANT: That would be a letter?
DR. IEZZONI: How do people feel?
DR. STARFIELD: Do we need to spend a lot of time on it? I do --
DR. IEZZONI: I am not sure what I am hearing from the group. It sounded to me that Marcie and Eve would say it is a little more complicated than that.
DR. STARFIELD: Not that this has not been on people's radar screens.
DR. IEZZONI: Right.
DR. STARFIELD: Long standing.
DR. MOR: That would be very difficult to do. It is not just a minor adjustment; it would be a really substantial adjustment, and you might lose from both. Whereas, there is some pertinence to having the BRFS in place as an ongoing thing that is useful to the states, but the question is for national policy purposes, that is what this SLAITS issue is about. It is a radically different structure insofar as it is centralized as opposed to state mediated.
DR. STARFIELD: I am just talking about the sample. That is all I am talking about.
DR. IEZZONI: But what I am hearing from Marcie is that they are trying to identify specific populations and so they screen millions --
DR. STARFIELD: You can do that with the same sample.
DR. IEZZONI: They screen a million households --
DR. STARFIELD: I understand that.
DR. IEZZONI: -- before they get, you know. I think that I would probably feel that we needed to have a lot more discussion about that before I would make a recommendation about that.
PARTICIPANT: Right.
DR. POWELL-GRINER: May I interject.
DR. IEZZONI: Yes.
DR. POWELL-GRINER: I believe that the committee needs to look at a very pertinent player in this issue, and that is the states.
DR. CYNAMON: The states, exactly.
DR. POWELL-GRINER: I believe that you need to contact the state organizations and talk to them and get their impact. This is not just a CDC issue, and it is not just this committee's issue. The states are the ones that have the real -- that will ultimately have the say in what is done. I believe you need their input.
DR. IEZZONI: I agree. What I am trying to figure out right now is what additional work does the subcommittee want to do on this?
DR. STARFIELD: Can I weigh in on that. I would think that we have so much on our plate, I do not think that we should take this on in the work plan right now.
DR. IEZZONI: Okay. How do other people feel?
PARTICIPANT: SLAITS?
DR. IEZZONI: SLAITS versus BRFFS type of --
PARTICIPANT: Oh, that issue. No.
DR. IEZZONI: Not right now. We just put this into kind of our meeting minutes and hopefully people will read the meeting minutes and -- yes?
MS. WARD: I would not like to sort of close it forever. What I would like to do is put it on the page. When we have other things finished and we are not as overwhelmed as we are, I think there are one or two or more people on the committee or on the subcommittee who could begin to look at some additional questions that need to be answered --
DR. IEZZONI: Okay.
MS. WARD: -- who would need to be involved. I do not see more hearings, but I see some analysis that needs to be done before we can come to any conclusion.
DR. IEZZONI: We could also, at a breakout session, have an hour and a half spent. One of the things that I am hearing from Marcie is that they are in Texas, and they are hoping to get into Minnesota, but they are really still very early in the whole process. I thin it is not ripened yet enough to be able to go that much further. So, what we will say is at this point we will probably not make any specific recommendations or pronouncements.
Given that, yes, Olivia?
DR. CARTER-POKRAS: I just wanted to point out that the Data Council had been briefed, the comparison of the two in addition to looking at other surveys that potentially could provide state-specific information like the Youth Risk Behavior Survey and that sort of thing. There was a nice chart, a spreadsheet, that was helpful to us. I do not know whether that could perhaps be updated or something just to brief some people who might not be as familiar with the pros and cons. I think age, for instance, was one of the issues that were talking about. There is a very different target population.
DR. IEZZONI: Excellent suggestion. Let's see if we can get a copy of that.
DR. MOR: I do have something that disengenders an idea that I think this subcommittee should actually consider. The whole notion of linkage, geocoding, survey depth was raised repeatedly as a potential problem with the NHIS redesign in terms of that quote issue of confidentiality because it increases the burden on the confidentiality of data and complicating data release. This is an issue that comes up over and over and over again, and we have to solve this. We have to stimulate folks to begin to solve this problem because to the extent that it holds up, complicates or modifies the availability of these kinds of data for analysis, we all lose.
There are hopefully some strategies that go beyond what the Bureau of Labor Statistics does as basically say we have this data center that you can come down to Baltimore, to Silver Spring, and work in our little shop and have a good time, but you cannot take the data home with you. Or, for instance, the HEAD data which is funded by NIA. Because of the fact that it is all of this linked information, they are actually instituting procedures where someone goes and audits the security systems in your university computer because of the number of break-ins that are going on now, people vacuuming data out of computer systems. So, these are really big issues.
There must be -- we may not have a technological answer or an invested answer today, but I think if we do not push for this, it is going to be a continual problem greatly complicating any efforts for integration and data availability in the big HIPPA vision, so to speak.
DR. IEZZONI: That might be something that we should talk about jointly with the chair of the Data Privacy and Confidentiality Subcommittee. Is Kathleen Frawley still the chair of that subcommittee?
PARTICIPANT: I think so.
DR. IEZZONI: She just was not at this meeting. So, we should maybe try to follow up with her on that because I know Paul was frustrated by that and getting the combined NHIS and the MEPS data. Kathy?
MS. COLTIN: In that regard, I thought the other thing that I heard was while a lot of the possibilities of the survey integration have been realized, a lot of the realities in terms of the availability of analytical staff to actually make use of the linked data files is a real problem. So, the data are coming in with possibilities to look at longitudinal cohorts, possibilities to link the data, and there are not the staff resources or the analytical planning to say this is what we are going to work on, and we will have this many staff, so we are going to let out an extra research contract for people to look at these other things or whatever. I do not know that there is any such strategy as to how they are going to address those limitations. It plays into the issue that you are raising in terms of one way to do it is to put the data tapes out there and allow researchers to use them.
DR. MOR: Researchers cannot use it.
DR. IEZZONI: Talking concretely, what should we do with this? I agree with your sense. What should we do with that as a subcommittee?
DR. MOR: I would be happy to write a memo, I guess to you all, which could then be sent to Kathleen Frawley and the other committee, suggesting that this should go on somebody's agenda at some point for the overall committee to deal with.
DR. IEZZONI: The privacy piece, absolutely. I think the issue about budget, frankly, that is what it boils down to and FTEs, is this something that we -- I mean, Marjorie, would it be of any value to us to write a letter to the Secretary saying that we are concerned that because of budget cuts that there are not adequate analytic staff on board to be able to maximize the utility of the information that is being generated? Is that of any value?
DR. GREENBERG: Well, this sounds like a reprise of what we were talking about at the state level, which I guess is much worse, but on Wednesday or Tuesday, whatever day that was. Some of this really requires methodological work. It is an effort to see really how these linkages will lead to -- what ways to benefit from it. We have nine vacancies in the Office of Research and Methodology. We have been decimated, no, decimated means only one in ten, so that is not good at all.
DR. IEZZONI: Marjorie, can I be clear about this? Are you recruiting for those positions and you cannot fill them, or they have been frozen?
DR. GREENBERG: A mixture of recruiting. Certainly, we are recruiting for some. We are not able, often, to recruit at a level where you can get people who really have those kind of qualifications. It is complicated, and I do not know that --
DR. NEWACHECK: It affects morale, as well. I think this is a big issue.
DR. GREENBERG: HIS. I mean, actually in both HIS and OREM, we have actually lost people because of very premature death. It has been complicated.
DR. IEZZONI: That has been a problem for you.
DR. GREENBERG: This whole issue of people being adequately trained. It is getting somewhat away from the budget thing. One thing that NCHS is also losing its capacity to do is to -- I do not know if Doug Williams is still here. I guess he is not. It is to even provide training the way we have through the ASTE(?) courses. So, I think it is complicated, but it impacts at the federal level, and it impacts at the state level, too.
DR. IEZZONI: We certainly heard it in Arizona and Massachusetts and from the hearings. Maybe this is sounding like a potential letter about the state of preparedness of the work force to be able to do the analysis that we need to be able to do.
DR. GREENBERG: Frankly, both epidemiologists and mathematical statisticians in the private sector will -- like pharmaceutical companies --
DR. IEZZONI: But even Kathy cannot get --
DR. GREENBERG: I mean like pharmaceutical companies can just pay so much more
DR. MOR: They just do not make them now.
DR. IEZZONI: We just are not making enough of them. So, is this something that should be -- I mean, we have heard this repeatedly at all levels.
DR. GREENBERG: Getting a man to the moon, we did not --
DR. IEZZONI: Right. Is this something -- how would people feel about addressing this? Paul?
DR. FRIEDMAN: I think there are really two issues that we have been talking about here. One is the work force preparedness issue and other is the budget issue for NCHS. I think the budget issue is one that has concerned me for a long, long time because it has been this gradual but progressive deterioration of real dollars or real support for the agency. It has tremendous impact on not only the ability maybe for them to conduct their work but also for morale, for attracting new people, et cetera, and it goes on from there. I think that it is an area that this committee has a responsibility to make a statement.
DR. IEZZONI: Yes.
DR. FRIEDMAN: I would be willing to take the lead on a letter.
DR. IEZZONI: To draft a letter.
DR. GREENBERG: You might want to also maybe draft something but then meet with Ed Sondeck and talk with him about it. He is not only the Director of NCHS, he is the Senior Advisor to the Secretary on Health Statistics.
DR. IEZZONI: Paul, why don't we draft a letter, and barbara and --
DR. STARFIELD: I guess this is sort of fine and we should weigh in. We weigh in periodically about the budget. That is all we can do is write a letter, but I do not think we should delude ourselves that we are going to have any impact, because we are not.
DR. IEZZONI: No, but I think --
DR. STARFIELD: It is a broader question. We have to get advocacy groups to advocate successfully. It is not something that is going to be done within HHS. It is not going to be done by the Secretary because I do not think the Secretary has a lot of discretion either.
DR. IEZZONI: I think we are all hearing you. I just think that Paul is saying that we have heard this, and it is our responsibility to raise it one more time. Dan?
DR. FRIEDMAN: I think that a letter would be fine. Having said that, my impression is that the state of the budgets really have varied substantially even within CDC and between CDC and NCHS. I think that if -- in other words, CDC is doing better than NCHS, and different centers in CDC are doing better than NCHS. So, I think there is an NCHS-specific issue. While a letter would be helpful, I think what would frankly probably be more helpful, if we want to get involved in that, is somehow or another working with NCHS around vehicles for -- I am trying to find the right word.
DR. STARFIELD: Developing a constituency.
DR. FRIEDMAN: Yes, developing a constituency. NIOSH has been doing it very successfully for the last couple of years. There are some other parts of CDC that have been doing it very successfully.
DR. GREENBERG: Frankly, that is part of the agenda with the Health Statistics of the 21st Century Plan that Dan has been so helpful with.
DR. IEZZONI: Why don't we keep this on our plate, then. I think it is probably a good idea to write the letter but continue to be thinking about it. Hortensia, thinking about the advocacy?
DR. AMARO: Yes, I would favor that broadly in terms of all of the issues that have been brought up and then specifically, every time I ask a question about what is happening with addressing the limitations of our current data system with respect to racial and ethnic minorities, and it is not just the ethnicity identifier, getting the labels right, but it is the sampling, different methodologies. The answer is that we do not have the money. So, even along those lines, I think we are not going to be able to make any progress, and we are going to be stuck in the same place unless the budget issues are addressed.
DR. IEZZONI: You will not have information on people who are Spanish speaking but even there maybe not.
DR. AMARO: There are lots of others, Asian populations, for example --
DR. IEZZONI: Exactly.
DR. AMARO: -- and other issues beyond that. That is like the first two items on a list of issues.
DR. IEZZONI: So maybe, Hortensia, you could draft a paragraph that would go into the letter that Paul would be putting an initial draft together based on what you just observed.
DR. AMARO: Okay. Send me yours, and I will add it.
DR. IEZZONI: Okay, great. Then what we can do is if we write such a letter is see whether we could have it on the agenda for the November meeting for the full committee to look at.
Now, this issue of training analysts and building capacity to learn from the health data that we do produce. Is that something that our subcommittee wants to take on as a major issue or is that -- I mean, how do we want to raise the visibility of that issue? Is simply a letter again a way of doing it? I think maybe -- has the Data Council heard this issue? Marjorie, do you have any idea?
DR. GREENBERG: It has not -- I am looking to see if Olivia is here because although I do attend them regularly, and it really has not been on there. Certainly the budget issues have.
DR. IEZZONI: But the capacity issue has not been.
DR. STARFIELD: Who are we talking to when we are talking about capacity? Who are we talking to? Are we talking to the Department of Labor kind of issues? What are we talking about? Traineeships within the department? What are we talking about?
DR. IEZZONI: I think we just need to raise the issue and say that we need to come up with a solution. I do not think that we are in a position of proposing specific solutions at this point, but I think, Barbara, we heard this in Arizona. We heard it --
DR. STARFIELD: Oh, I know.
DR. IEZZONI: We have heard it, and it sounds like the Data Council has not heard it, and I agree with Kathy that the Public Health Service training programs back in the 1970s were --
DR. STARFIELD: Do we want doctoral trained people? Do we want masters trained people? What specifically can we --
MS. COLTIN: I think we need a mixture of both. I think we need a mix of both. We have tremendous difficulty recruiting biostatisticians.
PARTICIPANT: Oh, yes.
MS. COLTIN: That is crazy, so that is doctoral. I am thinking mostly doctoral level, although there are some excellently masters trained biostatisticians.
DR. STARFIELD: Well, we can have more ACHPR traineeships. They will not put out these people though, I mean, you know. That is really what we want.
DR. IEZZONI: I think we need to raise the issue. It does not seem like it is on the radar screen --
DR. GREENBERG: I think it is on the radar screen of individual members of the Data Council, but it has not been a Data Council issue.
DR. IEZZONI: It has not been a Data Council issue, and I think that it is something that we are hearing repeatedly, and it is impeding capacity everywhere.
DR. STARFIELD: I am just thinking about that we ought to be constructive instead of just always raising questions.
DR. AMARO: What we can do is raise the issue as something that we need to bring people around the table to think about and articulate the extent of the issue and possible solutions. I think we do not have to, right now, come up with what are the solutions. I think it is a serious issue because it is professional preparation of people in the field that represent the work force, well, for this agency but more broadly as well. I think it is an appropriate issue.
DR. MOR: Barbara, it is not just the AHCPR Ph.D. or post-doc training programs. It is basically masters prepared people who are able to actually handle, visualize and extract information from the reams an reams of data. That applies to the commercial, the industrial and the public health sector. It is the schools of public health that are not producing them. The MBA schools are producing them, but the schools of public health are not.
DR. STARFIELD: I agree, and that is exactly why I think we need to be more specific. Just raising the issue is not enough. It is not constructive enough.
DR. IEZZONI: Elizabeth, you had a comment.
MS. WARD: I was going to say that what is different about this conversation is that it is not just government funding for government that we are talking about, and it gives us a very different angle. It is one of the solutions that I think Barbara was thinking about. I think that I agree with Barbara, that we need to think a little bit more about sort of the direction, and probably that is to the benefit. If we could find some alliances with the health care industry commercial side to begin to concur, this could take a much more likely to be solved issue, actually, than one more government complaining that government is --
DR. IEZZONI: Is this something that we want our subcommittee to spend some time on? Do we want to spend an afternoon? Do we want to spend a day on this?
DR. STARFIELD: I think the answer depends upon whether there is someone on the committee who wants to take this on and lead it.
DR. IEZZONI: Right. I agree. You know, actually, Barbara, that is a perfect lead in for what I was talking to Paul about at lunch. Everybody on this subcommittee is busier than even if we cloned ourselves we could handle most of the time. The way, frankly, that some of our agenda gets set is because there are advocates on the subcommittee who want to spend extra time, like Hortensia has spent considerable time on the islands and territories. George VanAmberg and Elizabeth spent considerable time on the Medicaid managed care. So, we really need to defer to people who are willing to spend considerable time as we are thinking about agenda items to pursue further.
Is there anybody on the subcommittee who would want to take this on as a role? Now, keep in mind that this might not be a big deal. It might be that we have an afternoon where we bring in six to eight panelists to talk about it and based on that we write brief kind of five-page report from just observing what people say and suggestions that people give us. That is not really all that much work. It is work for staff, obviously, to pull together the meeting, and whoever is the advocate of this would work with staff to come up with the right people to be at the table. Is there anybody on the subcommittee who wants to take up this issue? Any educators on this subcommittee?
[Laughter.]
DR. MOR: Kathy, do you have some interest from the industry side?
MS. COLTIN: Yes, I would be willing to work with someone on it.
DR. MOR: What I would probably do would be to call in three or four people from schools of public health and health administration arena plus somebody from a health program, health management, and then just sort of try to address this issue. I think it is a broad, broad based issue.
DR. IEZZONI: Yes, it is.
DR. MOR: Personally, I need to deal with this other post-acute thing.
DR. IEZZONI: I know. So, maybe like half a day or something like that instead of having something for a full day?
MS. COLTIN: I think a panel for half a day would be --
DR. IEZZONI: Right, having a panel for half a day.
MS. COLTIN: I think we do need to give some careful consideration to getting the right mix of people on the panel. I am not sure it is just schools of public health --
DR. MOR: Elizabeth is absolutely correct in terms of the broad -- it has a broad ramification and broad implications and therefore the message would be more readily received.
DR. IEZZONI: Frankly, Marjorie, when I listen to this, I see this as a full committee issue.
PARTICIPANT: Oh, it is.
DR. IEZZONI: I see this as a full committee issue. So, I would see maybe the half a day being with the full committee because this has an issue for training people and how to deal with data every which way and the privacy. It is really -- it is about HIPPA implications. So, why don't we bring this up to Don at the executive subcommittee. Can you make a note to bring this up?
DR. GREENBERG: Sure.
DR. IEZZONI: I think that a lot of people around the full committee would have a feeling about this. Hortensia?
DR. AMARO: If and when we do get to the point that this will happen, I would really like to suggest that we make sure that we have representation from -- I forget what the exact names are, but all of those programs, Training Racial and Ethnic Minorities in Biomedical Careers. Can anybody help me? I know NIH has one, right?
PARTICIPANT: I know what you mean.
DR. IEZZONI: Yes.
MR. HANDLER: My agency, the Indian Health Service, has a problem in recruiting health professionals, doctors, nurses, dentists, pharmacists. We have two programs that we have in place that has been going for at least ten years, a scholarship program and a loan repayment program. We either pay in full for a scholarship for a student to attend medical or any other health professional school or if they are already in the school and they decide that they want to align themselves with us. Then we pay their loan if they have made a loan. They have to obligate themselves to work in our facilities for at least two years, and then the loan is forgiven or the scholarship is not paid back. Maybe that same type of thing could be done to fill vacancies in other organizations, not just my own.
DR. IEZZONI: Okay, yes, a lot of us are familiar with that kind of approach.
DR. GREENBERG: I was wondering about HRSA.
DR. IEZZONI: Yes. I'm sorry?
DR. GREENBERG: We were just wondering whether -- the extent to which HRSA --
DR. IEZZONI: Someone from HRSA.
DR. GREENBERG: Is anyone from HRSA -- is Mike Millman your staff person?
DR. IEZZONI: Here he is. I saw him come in earlier.
DR. GREENBERG: Is this something that HRSA concerns itself with?
MR. MILLMAN: Not specifically this. We have had public health training fellowships, but it has not been specifically.
DR. IEZZONI: Okay, Marjorie, I think we need to bring this to Don's attention and say that we have heard this repeatedly, it is an NCHS problem and everybody problem.
Dr. GREENBERG: I think as you were saying that it cuts across the whole committee. I think obviously our data resources are also -- I mean, our data systems are being stretched and compromised by resource issues, but also I think there is always a little more money, it seems, to start a new data system than to really deal with how you are going to -- what is going to be done with the data, who is going to be there.
DR. IEZZONI: I think there is so much emphasis on education right now throughout the entire kind of government that bringing our voice to the table around this type of education would be useful.
Okay, I see people moving on mentally. Are there any other things from this morning that make you think we should do something immediately like another letter to the Secretary? Yes?
DR. NEWACHECK: I think one of the concerns that I had from the morning's discussion about the CAHP survey was whether or not information from Medicare CAHPS and Medicaid CAHPS will become publicly available. I think it will be an important dataset for researchers both inside and outside of government to look at to evaluate how well we are doing in serving these populations.
I was a little concerned with what seemed to be a little bit of trepidation about releasing the data publicly. I understand there may be some barriers in the department now towards that, and I would like to put it onto the agenda as something that we come back to look at in perhaps a couple of months to think about whether or not there is something we can do there to encourage public release of that information.
DR. IEZZONI: I think that is fair. It is just that the sample is so weird. It is 600 people within specific health plans for CAHPS. They are not even doing Medicaid CAHPS yet. That is planned.
[Simultaneous discussion.]
That is planned, but only in a few states have they really started doing that. Well, but Medicare, the HCFA people are not going to be the people who determine whether that is released. You know, it is going to be state by state by state, just like the Medicaid tape to tape. Medicaid data has always been --
DR. NEWACHECK: Well, it would depend on who is funding it, I think, in part.
DR. IEZZONI: Well, who has funded the state CAHPS? Elizabeth, did you do CAHPS in your state in Medicaid?
MS. WARD: No, we did our state purchasing agencies, Medicaid and what we call our Health Care Authority. We tested it on our --
DR. IEZZONI: So, your people -- you paid for it?
MS. WARD: -- public employees.
DR. IEZZONI: Yes. You paid for it. Kathy, you had a comment about this?
MS. COLTIN: I think that you are looking at a process that is in transition. HCFA paid for the first version of CAHPS. Whether HCFA will continue to pay for it or have health plan pay for it is a moving issue. My sense is that eventually they are going to want the health plans to pay for it. They have done that with Healthy Seniors, and I can see it coming in CAHPS. I think that state Medicaid agencies are likely to end up doing the same thing, and so it is going to be the health plans that are going to end up paying the bill to collect the data. So, what happens with the data once it is collected. There is this question of who owns the data. Is it whoever paid for it?
DR. NEWACHECK: Yes. My point --
MS. COLTIN: Is it who commissioned it?
DR. NEWACHECK: I'm sorry. Yes, my point would be that as much as possible, it should be in the public domain as long as it does not violate any confidentiality purposes or proprietariness purposes.
[Simultaneous discussion.]
MS. COLTIN: -- issue around the intersection with the privacy and confidentiality subcommittee is when you start looking at data at higher levels of advocation, at the provider level, at the medical group level, at the hospital level, at the health plan level, what kinds of privacy protection should be afforded when you go beyond individuals to organizations, if any. I mean, I was very surprised to find that in Massachusetts, in our hospital discharge abstract database, even though they have detailed payer codes, they do not want to release reports --
DR. IEZZONI: They will release it for Blue Cross. They release it for Blue Cross, but the commercials are all lumped together.
MS. COLTIN: They have it. They have it with the detailed payer codes for all of the individual HMOs, but if you request a custom report -- I know this because I have already inquired. I will get a dataset, and it will tell me here are the rates for Harvard Pilgrim Health Care, and then here is the rate for HMO A, HMO B, HMO C. It will not tell me which one they are.
DR. NEWACHECK: Yes, right.
MS. COLTIN: Why are we protecting the confidentiality of a managed care organization? Where do we draw the line between where -- I mean, I understand individuals, even though I have concerns about protecting physicians. I think what they have done in Pennsylvania and New York has been different from what they have done in other states. It seems to me that is another interesting area and issue to think about.
DR. IEZZONI: Okay, why don't we keep track of that issue, Paul, because it is a moving target. I mean, this is really their first year of doing it. Let's just ask them again. It sounds like they are thinking about it, but they are really just trying to get their first report out. The administrator is in front of Congress, who is really mad at them for not getting out their first report on quality. So, let's track that issue. Hortensia?
DR. AMARO: I was a little concerned, you know, and I put it into the context of the budget cuts and all of that. I was a little concerned to kind of hear that in this process of survey integration, people just are having -- are very limited in how they are able to, as they do that address, the limitations that we have had in our data systems pertaining to racial and ethnic groups. I can understand why based on the budget cuts and all of that, but I am very concerned that they are recreating or making these changes at this large level in our health data systems and that we are going to continue to basically replicate a lot of the problems. Maybe there was not enough time for discussion of that, but I would really like Dr. Sondeck to come back and kind of give us a broad sense of how he sees addressing these issues considering what the department is facing.
DR. IEZZONI: We will definitely have your paragraph in Paul's letter about the budget cuts, and we can think about whether, during one of our breakout sessions, Ed would be willing to come in and talk to us about that further. Last July, Ed was happy to come in and talk to us about language, for those of you who were on the committee at that time, and so hopefully he will be willing to do that.
Okay, Marjorie?
DR. GREENBERG: One thing that they kind of alluded to that is future oriented that you might want to play a role in or follow or have some discussions about and see what role you want to play is that we are beginning to look towards the redesign of the HIS sample based on the --
DR. IEZZONI: The census.
DR. GREENBERG: -- the 2000 census. You have to start that quite far ahead of time, and of course, it is part of an overall governmental effort to redesign all of the surveys that are based on the census. The conflict or the kind of conflict or competition that always exists is for doing more in geographic areas versus doing more for populations, and you kind of try to do both, but you do not maybe do either very well. You have the integrated survey design, which really actually was a factor in the previous -- in the design that is being used now. That is how NSFG ended up being able to be linked to NHIS because it was part of a past redesign. These are the critical issues that come up in redesign. Of course, it is a question of resources and sample size, but it is also a lot of methodological issues. So, you might want to engage on that, both that there be adequate resources to even do the type of research that is needed to develop, to evaluate the current design and the implications of other approaches and just follow it through its process.
DR. IEZZONI: Yes, well, we should follow that just like we have been following the implementation of the OMB and how it is going to affect longitudinal data analysis. It is a very related issue involving census as well. So, we should have that as breakout sessions.
Okay, we need to come up with some notions about major initiatives that our subcommittee wants to undertake starting in January or so. Joan, we talked on Tuesday about how you are preparing a draft report for us for the islands and territories meeting. Hopefully, we will be able to respond to that very quickly and get some recommendations coming from that once we have had a chance to draft it. We are working to draft the Medicaid Managed Care Report, and we are having a meeting on October 30 to talk about the subcommittee's views of recommendations. So, hopefully, we will be able to move on that.
We need to decide what our initiatives should be starting in January of 1999 because we need to give staff adequate time to plan and think about budgets and hearings. So, the table is now open for people to make specific suggestions about areas that we as a subcommittee would like to look at. Kathy? We will go around the table.
DR. STARFIELD: Let me just ask you, does this have anything to do with the work plan? Are we going to discuss the work plan?
DR. IEZZONI: It has to do with the work plan, Barbara, yes, but I would just like to get ideas on the table because I need to get a sense of commitment from committee members around issues because frankly, we need people to be committed to specific things. A work plan is a piece of paper, a piece of paper reflecting what we thought in June. I just need to get a sense right now what people are thinking. Kathy.
MS. COLTIN: I was going to make a pitch for post- acute care. It has been on our agenda for a while, but we really have kind of dealt with it in patches here and there as opposed to a concerted effort. I think it is a very important issue in the private sector, you know, for managed care plans. I think it is a very important issue in the public sector in terms of HCFA's issues, and I think that it is becoming and will continue to be an important issue in the states. Now states are going to be getting the MDS data. What are they going to do with it? How are they going to use it? Do they even know? What have they thought about it? So, I think there are state issues around post- acute care and monitoring as well. Also, obviously, the licensure and other types of state functions that relate to post-acute care settings.
DR. IEZZONI: Vince?
DR. MOR: That is what I would like to spend time doing. All of the issues that Kathy mentioned plus I think the overriding issue of the public/private cooperation/collaboration or availability of data because that is really the key. It is not sufficient to just say that we have data from nursing homes. It has to be somehow or another linked to data from hospital material, and the more people who are in HMOs, the less data there are about what happens in hospitals. You cannot crosslink those at the present.
There is the issue of what data and how do they crosswalk across, let's say, the MDS on the one hand versus the OASIS(?), which is the home health system, on the other hand, versus what has become, I suppose in some sense, a de facto standard in the hospital rehab world, which is the FIM. What is going to happen under prospective payment to those rehab hospitals because HCFA sort of in some sense has really declared war on them saying, you know, We are just not going to do it anymore; we are going to subject you to prospective payment, and we are going to probably impose some kind of resource utilization group that is largely based on a SNIFF(?) with some modifications. That may be fine, and that may be good, but that is part of the context, and it is very regionally differential as well.
The other issue is right now, we will have data from these little funnels --
DR. IEZZONI: The smokestacks or whatever.
DR. MOR: The little smokestacks of providers. Yet, we have no information nor even any data to suggest that it is possible to use the same kind of data that is collected in one setting as a premeasure of a postmeasure taken in another setting. Do they mean the same thing? Can they be used in the same way? If you are going to be talking about quality performance and three-month performance rate improvements, post-hospital discharge for this plan versus another plan, this is a massive leap of faith as to whether one can do this.
I think those are all -- and then there is the final issue, which is mentioned in Don's piece that we talked about yesterday. The big vision of all of this integrated stuff is what should the federal role be in training, teaching, setting an example for how to train the world out there to collect the data comparably, to do it right and to then use it in some way. While in retrospect HCFA has done a reasonable amount on the nursing home side, had it really made a big, big investment in doing that right, it would have -- from the outset, it would have been a whole different kettle of fish. The agency was not geared up, and I think there was real ambivalence about what the role of the government relative to the market should be on this kind of thing. I think that is a general issue all together. It is a whole bunch of stuff, but it is very pertinent to this issue of data for this context.
DR. IEZZONI: The data element of functional status and how you define that is a core data element issue. Remember, Barbara, on the core data elements, we never figured out how you would do that. That would be part and parcel of this. My question to Kathy and Vince--
DR. MOR: That is easy.
DR. IEZZONI: No, wait, wait. Listen to what my -- wait. Let me finish. My question to Kathy and Vince was are we talking about adults or is there a pediatric piece to this because I know that we have two pediatrician or pediatric interested people on the subcommittee, and I would want to make sure, Vince.
DR. MOR: There has to be pediatrics.
PARTICIPANT: Oh, yes?
DR. IEZZONI: I just want that to be on the table so you guys do not think that all we are talking about is adults.
DR. STARFIELD: Right. Thank you, Lisa.
PARTICIPANT: Lisa, I appreciate that.
DR. STARFIELD: Because the issues across different kinds of settings are just as pertinent for children.
DR. IEZZONI: Now, Dan, I see you kind of shifting restlessly in your chair, and I am not sure if it is because you have a bad back or what.
[Simultaneous discussion.]
DR. FRIEDMAN: It is -- I will be glad to wait my turn, and it is because I get that.
DR. IEZZONI: Yes, get antsy. Well, now is your turn.
DR. FRIEDMAN: Okay, if now is my turn, let me sort of bookmark, as it were, that the health statistics' vision process, which is a joint effort with us and NCHS and hopefully SINSTAT is it is a major activity. It is a somewhat different model for committee functioning because essentially what we have done is we have had a work plan that was developed last spring, and then since then, Caroline, Gerry and I and Barbara and somewhat less so Barbara and Paul, you know, it has primarily been my working with the staff to start fulfilling that work plan. Then in 1998, it was primarily in 1999 that it is going to be hopefully affected.
DR. IEZZONI: The thing that was different about that also from our typical committee processes is that you had budget commitments from -- that were bigger than what our usual budget commitments can be. We have been very lucky to get funding, for example, for the GW subcontract, and we might need something like that for post-acute care, but that is a slight difference from what you are describing like with the SINSTAT and stuff.
DR. FRIEDMAN: I think that we clearly need to have that on the agenda both because it is going to be requiring time as well as because it is a committee commitment.
DR. IEZZONI: I think that we need to figure out how our subcommittee integrates with that specifically. Is that a work group?
PARTICIPANT: It is.
DR. IEZZONI: It is a work group.
DR. FRIEDMAN: I am not clear what happened. It is not -- I do not know if it is a work group or not.
DR. GREENBERG: We are carrying it as a work group in our organizational structure of I think Dan and Paul and Barbara. Gerry Hendershot is --
DR. IEZZONI: Is staffing it.
DR. GREENBERG: -- staff to the work group, and Caroline has been wonderful in working with him.
DR. STARFIELD: But I think that even the rest of us in that work group can carve out specific pieces of that. I am especially interested in a couple of things, the core data, the role of the core data elements and the vision for health statistics, as well as the health objectives for the nation. I think they are all very interrelated. So, I see myself in that work plan, but these particular interests in that.
DR. IEZZONI: You know, I know people are going to start peeling off as Elizabeth did, and so let's try to be more specific.
DR. MOR: Yes. I was --
DR. IEZZONI: Yes.
DR. MOR: What do we do next.
DR. IEZZONI: Well, okay, what we do next is let me just give you an example of what happened with the Medicaid managed care. We decided that we wanted to spend a year talking about it but that during that time we would continue to monitor things like the OMB Directive 15. We would do the islands and territories, which was basically a single meeting, and so we would do other things, but we would have a couple of separate hearings and site visits around Medicaid managed care. We were grateful; we got money for a subcontract to be let to George Washington for that. But, it took us a long time before we figured out what the product was that we wanted and exactly how we wanted to tackle this Medicaid managed care. Jim Scanlon sort of stepped in and saved us after a meeting in July of 1997 where we just were wishy-washy all over the place, and it felt really bad.
DR. GREENBERG: Jim who?
DR. IEZZONI: Jim Scanlon.
DR. GREENBERG: Oh, Jim Scanlon.
DR. IEZZONI: By suggesting what the unique role of our committee could be. So, what we need to do right now is hear from Kathy and Vince, who have proposed this, and Barbara and Paul, who recognize that there is an important piece here which is specific to kids, and Vince, you know more about the elderly and adults. What would you see being the end product of all of this, and how would you see us getting there?
DR. MOR: Let me just offer. I thought about this the last time that we talked in the summer and broke it down into five sort of different proposals or goals for sets of hearings. I do not see this activity in the same way as the Medicaid managed care perhaps because of the timing of the real world --
DR. IEZZONI: It is happening now.
DR. MOR: -- and perhaps -- yes, probably because of the timing in the real world. There are three or four major initiatives that are underway in HCFA in response to PPS and a Balanced Budget Act which will take several years to actually unfold. There are a lot of open pieces of that. In addition, as the proportion of the Medicare beneficiaries in managed care goes up, the real world out there begins to gain increasing experience and tries to replicate those plans with experience and grapples for position, how do we learn how to do the care clinically over the continuum. They are not trying -- at this point, they are not worrying about their data issues, although they will have to.
Those who have volume, like Harvard Pilgrim, for instance, are now beginning to worry about how they are going to deal with this data issue. Most of the plans still do not. This is still being and becoming. I do not think we can just say, okay, let's hold 2 sets of hearings in the next 12 months and then have 6 months to report on the same mechanism. At least that is how I would view it.
I think amongst the issues that need to be looked at, one of our public/private data gaps that are facilitating the private tracking of patients across multiple providers, and that is just what are the gaps operationally and at the big picture level. Secondly, what are the implications of converting data systems in individual sort of networks as they try to just talk, get these systems to talk to one another, which is not a trivial matter at all. Is anyone doing this right in some sense.
Next is, what are the implications of again learning data systems which absolutely look like they are going to be set up for prospective payment as opposed to at least thought about initially for clinical applications, which is again one of the points. If you are going to have a data system that is just for money, then it is probably going to have one kind of use. It is going to have one kind of characteristic as opposed to data that is for clinical purposes.
Then the issue of measurement of the function. There is a placeholder in our minimum dataset, and I think there are a variety of candidates that could readily fit in there that would meet -- or at least as good as insurance as a data element.
DR. IEZZONI: Kathy, did you?
MS. COLTIN: My slant is very similar but may be a little bit different. I am interested in both the cross- cutting issues across acute and post-acute settings and the ability to monitor quality of care across those settings. I am also interested in the silo in terms of looking at quality of care in post-acute settings because I do not think that the science is well developed in those settings as it is in managed care plans and in hospitals. I am not saying that I think either one of those is where they need to be by a long shot, but I think these are even further behind. We are not seeing as much consensus being achieved around the way to go, what tools to use, in those settings as we have seen in some of the other settings. I think we need to try to push the industry toward consensus in some way.
DR. IEZZONI: So, Kathy, what would you see as a work plan for this specific project, for our post-acute care initiative? What would you see as the steps, and what would you see as the output of this activity?
MS. COLTIN: Well, you are putting me on the spot --
DR. IEZZONI: Well, but you are so good at this.
MS. COLTIN: -- to do that without having really laid it out.
DR. IEZZONI: We are going to need to try to -- I mean not today necessarily, but we need to push on this.
MS. COLTIN: One of the things that I think is a real problem for us is we do not know what mix of care in what settings achieves the best outcomes for people. There have been a few studies, very few, but a few that have looked at patients with particular conditions who have gone into rehab hospitals versus skilled nursing facilities. We are making choices when we discharge patients as to what is the right setting. What happens when you shorten the length of stay in one setting and trade it off against more time in another setting.
We are doing it with really not enough information or maybe even some assumptions about the fact that the content of care in a given setting is the same regardless of payer. I know that our patients in ECF get a lot more physical therapy than some of the Medicare patients in some of those same ECFs, and it is because our payment system does not limit that. We are saying that we are taking you out of the hospital earlier. We want you to get more PT in these settings, and we will pay for it. But you do not have that same ability, so an HMO patient in an ECF with a particular condition may be getting a very different mix of services than a fee for service patient, Medicare patient, in that setting. I do not know that we have data to know that.
If we make policy decisions about what settings people should go to without understanding that in fact if you change the mix of services, there may not be differences in those settings, to me, that is an issue. I would like to understand better what data do we need to be able to look inside the box at what services people are getting, what mix of services, and how that impacts on the outcomes of care in these different settings, what are the best tools available to measure the outcomes, to measure the inputs to the care process and so forth, and then to be able to look at those tradeoffs across settings.
DR. IEZZONI: When we were at HCFA back -- I think it was March 2nd that we had a day there which was a day that Caroline sweated to organize that turned out to be great. HCFA talked about the fact that the understood that they were having these silos of data collection which were basically mandated by Congress, a lot of it, but that they were going to try as an institution to step back and look at the conceptual framework for how you think about data for people. I mean, we are talking about people being treated in all of these different settings, and you can only collect so much about people. It would seem to me that we should kind of maybe start there. I know, Vince, that you feel the pressure of the hoofbeats, and you are saying that we have to do this in real time, but --
DR. MOR: Oh, no, I did not mean that we have to do it right away. What I was saying is that it is going to take longer than a year to unfold.
MS. COLTIN: That is true. That is true.
DR. MOR: I am not saying that we need to have an answer tomorrow. In fact, what I am saying is that I do not think it is possible.
DR. IEZZONI: Because I would see what might be useful -- and this could be too wishy-washy, and I want people to tell me if it is too wishy-washy. At the end of the year, if we have a document about data policy around post-acute care and what kind of data elements, maybe not necessarily a core dataset specifically but a kind of conceptual framework around this ranging from tiny babies, you know the 500 gram babies who go on to require all sorts of extended care due to being born prematurely all the way to the very old, thinking about all different kinds of settings and touching on the issues that Vince raised about the practical issues, who is doing this well, but also the conceptual issues about how should you be doing it, what are the data. I mean, is that too vague?
DR. MOR: It doesn't get at Kathy's issue of what data are needed to really understand what goes on inside the --
DR. IEZZONI: That is what I meant it to do.
DR. MOR: What?
DR. IEZZONI: That is what I meant it to do. I meant it to say what data, what should be the conceptual framework for the data that you should be thinking about for this purpose. I meant it -- ditto what Kathy said.
DR. STARFIELD: You said needed data elements, but we do not -- I mean, as with the core data elements, you may not be able to specify how to measure each one yet. It is the framework.
DR. IEZZONI: Maybe, I mean, we might get to the point where we feel that we can actually specify data elements, but we have to kind of first figure out what various stakeholders need, you know, what they are interested in learning, where the gaps are. It is kind of like what we did for Medicaid managed care.
MS. COLTIN: I do not think that it is terribly different from what they did here regarding the surveys. I think we have to identify a set of data needs. We have to understand where we are currently, and where are the gaps, and what are the options that are out there for filling those gaps, and what would we recommend among those options.
DR. IEZZONI: Okay, hopefully the tape will transcribe Kathy's comments verbatim because that sounds like a plan for -- in general, for where we should be heading with this. What do people think? Bringing in the kids issue, I think the two of you are -- because our tendency is going to be to go ask HCFA, and we might need to ask states, you know, around Medicaid, disabled children. I think minority issues are going to be very key here, Hortensia. I think that the extent to which people are impoverished and cannot afford certain types of care is going to be a very important issue. Marjorie?
DR. GREENBERG: This is not just a HCFA issue, obviously --
DR. IEZZONI: That is right. It is the private sector as Kathy --
[Simultaneous discussion.]
DR. GREENBERG: Well, and also others.
MS. RIMES: We also brought in ACRA(?) and NCHS because it is always viewed as HCFA except that it is not.
DR. GREENBERG: Kathy was talking about going from the acute care and all the way through. The reason -- this may not be on track, and we do not have to discuss it, but the reason Gerry is not here today is because he is at this meting that NCHS convened for two days on summary health measures, things like disability --
DR. IEZZONI: Is this in Baltimore?
DR. GREENBERG: I thought it was at NCHS, but in any event, disability, adjusted life years, et cetera. It ties in with the objectives for the nation. It ties in with functional status because looking at more than just how long people live but really what the quality of their life is, potentially, and making some real broad assumptions there. There are nursing homes. NCHS has, potentially, this whole range of surveys in a lot of different settings. As Ed said, we have not been able to field them all on a regular basis or to reap their benefits as we might, and to even begin to look at integrating them, which is in a way like what we discussed here, and integrating them maybe with the HIS, which was discussed here. So, you know, there is a lot there.
DR. IEZZONI: Let's talk really nuts and bolts then and Marjorie, since you are here. Does anybody see the need for an outside contractor to write a report, you know synthesize the information, maybe do a survey, or do a compendium of all of the federal surveys that might touch on this? You know, because it sounds like from what you say that there is a lot going on. Is there a role for seeing whether we might spend a little money to help assist this project? Are there specific things that people would see an outside contractor being able to do that would be valuable to us?
DR. NEWACHECK: There would be something, but at this point --
DR. IEZZONI: It is too premature to define?
[Simultaneous discussion.]
DR. MOR: I agree. I am still trying to get a handle on the broad brush here.
DR. IEZZONI: I just wanted Marjorie to put it in her thinking that we might --
DR. MOR: No. I agree.
DR. IEZZONI: We might come and ask for some money.
DR. NEWACHECK: Yes, definitely.
DR. IEZZONI: Yes.
DR. GREENBERG: We know you people will come for money.
[Laughter and simultaneous discussion.]
DR. GREENBERG: -- if ASPE continues to -- it is a commitment to get research money for limited research projects, and in some cases, like with Division 1 and individual agencies will step up to the plate, too, beyond what they already are. I mean, I think you obviously have to think realistically. I guess what would really -- to think and to identify, which may take some more discussions with people in the agency, et cetera, as to what would be helpful to the people in the agency who are thinking about this and worrying about this. I assume what you are talking about Jim Scanlon helping you focus was the Sarah Rosenbaum activity?
DR. IEZZONI: No, it was more a kind of a sense of what we could expect to achieve and what our unique role --
DR. GREENBERG: That conflict --
DR. IEZZONI: -- and that was very valuable.
DR. GREENBERG: That conflict that you have with Sarah is something that we have been kicking around in CDC, getting something like that, and at NCHS for some time. Things kind of seemed to converge, and so this was something that certainly part of the agency really felt would be useful, too, and would be useful to different parts of the department. So, if you can identify something like that.
DR. IEZZONI: Although, Marjorie, one of the areas that I think is really uncharted is the role of the private sector in paying for post-acute care. Everybody does think that it is a government either state or federal issue, but I think to the extent that we can maybe begin to find out more about the roles and policies and costs and activities of the private sector in paying for this, that could also be valuable.
DR. GREENBERG: The government may start using -- you know, if you can take care of all of the who owns, the big issues, the data ownership issue and the confidentiality issues, et cetera. I mean, I know the potential that Ed Sondeck -- you know, that we kind of convinced him of and he really believes in for the recent HIPPA implications is, you know, you may have data out there that is being produced as part of the health care system, which is as much private as public, but it can replace or supplement or interact with these public data sources. They are not either/or.
DR. IEZZONI: Another opportunity or way to think about this is that under the BBA, eventually, within a couple of years, managed care plans are going to have to submit full encounter data.
DR. GREENBERG: Right.
DR. IEZZONI: You know, starting out with the hospital data now, the acute hospital data, but moving fairly quickly, I think. I think managed care will actually push for moving quicker to full encounter data because we have --
[Simultaneous discussion.]
MS. COLTIN: We are going to be very concerned because policies like saying maybe we will not count the two-day hospital stays, that changes the whole practice. I mean, there are a lot of patients who you put in the hospital for two days and then they go into the ECF, but that does not get figured in. They are worried about gaming, that you are going to admit somebody for one or two days just to get the payment up when in fact --
DR. IEZZONI: Oh, I know, yes.
MS. COLTIN: -- that person really is part of a longer stay, but part of it is in a nursing home, and they are not collecting that piece, so they do not know. So, I think we are going to want them to have it all and see the picture. When that happens, there is going to be a wealth of data out there on patterns of practice in managed care and fee for service that are going to tell us an awful lot about how these post-acute settings are being used and being used quite differently.
DR. IEZZONI: Although one could argue that diagnosis code information, ICD9CM codes, do not tell you what you need to know about --
DR. GREENBERG: ICIDH might.
DR. IEZZONI: ICIDH might! Exactly!
MS. COLTIN: But again, this is an issue where data linkage can be helpful because if you have the MDS data and the claims data together --
DR. MOR: HCFA is putting into place what is called -- it is going to be called a post-acute MDS. It will actually span the multiplicity of settings, probably within two years. It will be smaller than the current nursing home MDS, but it will be done actually within 48 hours of entry into that setting and on a fairly frequent basis.
DR. IEZZONI: I see, Paul, you have to leave, too?
DR. NEWACHECK: Soon.
DR. IEZZONI: Pretty soon. Why don't we do this. I see exhaustion around the table. Why don't we say that we have identified our major area, although we will continue to do the other kind of smaller things and wrap up our other work. Why don't Kathy, you and me and Vince work with Caroline -- Caroline is spurring this -- to try to get -- Sorry. You are lead staff. [Laughing.]
[Simultaneous discussion.]
DR. GREENBERG: -- help Caroline, that would be good.
DR. IEZZONI: Well, but let me --
DR. AMARO: This is my area, too, research.
DR. IEZZONI: Let me just finish what I was going to suggest, that October 30th we are having a meeting to talk about the Medicaid managed care report, but it would be nice to have a proposal to the committee about how we are going to begin to address this and begin to get some dates on our calendars for next winter and spring, about days that we will want to come in to hold some hearings that we inevitably will. We will need to think about whether any site visits would be necessary; they may or may not be. It might not happen.
PARTICIPANT: Going to HCFA?
DR. IEZZONI: Right, going to HCFA. Well, that was a site visit. We had to get special dispensation to go to Baltimore.
[Laughter and simultaneous discussion.]
DR. GREENBERG: We do not object.
DR. IEZZONI: Yes, the cafeteria was much better, Marjorie.
DR. GREENBERG: I cannot wait to go back.
DR. IEZZONI: I know. So, why don't we, the four of us, kind of come back to the subcommittee with a proposal brief that will not -- you know, we will not have to spend too much time talking on October 30th about but that we can begin to really try to -- oh, Hortensia, you were not here when we decided on October 30th.
DR. AMARO: Yes, I have a standing retreat with about 30 of the staff.
DR. IEZZONI: We will miss you greatly, but everybody else can be there. We had a devil of a time kind of coming up with this, so --
DR. GREENBERG: Are you tied up the whole day?
DR. AMARO: Yes, I am at a retreat with all of my staff for the entire day, so I cannot --
DR. GREENBERG: Because otherwise I thought we could tie her in by phone.
DR. GREENBERG: Can I ask a question?
DR. IEZZONI: Yes.
DR. GREENBERG: Will this be the focus probably of the quality work group then also?
[Simultaneous discussion.]
DR. IEZZONI: Yes, I think it has to be, and I think it is a fit.
DR. GREENBERG: That is what I was hearing, and I think --
MS. COLTIN: Actually, I think there are two focus areas on quality right now. I think we should take the report that we are doing on Medicaid managed care and identify the quality of care for vulnerable populations issues that come out of that and what we have learned and distill those and then think about where we would need to go beyond that if we do in the quality area and this post-acute quality. To me, if I can get something going in the quality area in both of those, I feel like I will have done a lot in that period of time because those are two important areas.
DR. GREENBERG: I know that you had talked about trying to follow up from the two reports, the commission report--
MS. COLTIN: Those are two areas identified in the report.
DR. GREENBERG: I was going to say, taking pieces out of those reports also on the road map kind of flows from that, too. Okay, good. I just wanted to be clear on that.
DR. IEZZONI: So, Caroline, can you try to interface with the three of us over the next month to try to begin to generate ideas about specifics because we need to be concrete.
DR. MOR: Should I try to set up a conference call?
MS. RIMES: I think that would be lovely, or I can try to set one up.
DR. IEZZONI: Would this be a public conference call, Marjorie?
DR. GREENBERG: What?
DR. IEZZONI: Would such a thing have to be a public conference call?
DR. GREENBERG: It sounds to me like you are just trying to set up some kind of a work plan. Set it up, though -- it is difficult if one of you sets up the conference call and then tries to get reimbursed for it. It is much better for us to set it up.
DR. MOR: No, I understand. I just care about the date and time.
MS. RIMES: Barbara is sitting there looking at me in hatefulness, but no, we will get it. Why don't we set it up.
DR. GREENBERG: Oh, Barbara Hexler. I thought you meant Barbara Starfield.
MS. RIMES: No, no.
DR. GREENBERG: I never called upon her to set up conference calls before.
MS. RIMES: No, no, I did not mean that, Barbara.
[Simultaneous discussion.]
DR. GREENBERG: Barbara is so efficient, she already has it set up.
MS. RIMES: I know.
DR. IEZZONI: Hortensia wanted to make a comment.
DR. AMARO: Just before we end, so I do not know if you are through with this.
DR. IEZZONI: We need to try to just reach closure on the charge that Dan and Barbara drafted for us, so we could just vote on it and say whether we think it is okay or not. Hortensia, why don't you make your comment now while people are kind of -- your charge is in your little folder at your desk.
DR. AMARO: I came in a couple minutes late from lunch, so I do not know what was decided about that item on our agenda right after lunch, which was supposed to be a panel discussion on survey data needs and gaps in data for minority populations, SCHIP, states and the disabled. So, I do not know whether we are going to put that at another time. I do want to kind of come back to this issue of data on racial and ethnic minorities, and I know that I feel like a broken record.
One of my interests in coming on this committee -- I had three interests. Originally, one was to really try to reinforce the committee's role in addressing the gaps in the health data systems we have, especially since this was a time of change and integration, development of integration of data systems. It was an opportunity to do that. The second was the issue of inclusion of territories, and we are working on that. The third was trying to integrate mental health and substance abuse into what the committee considers.
I think that we have sort of here and there addressed the issue of race and ethnicity, but I feel like there is more that we could do there. It sort of feels like we mention it but it kind of -- I am not quite sure what we are achieving on it. Especially with the President's initiative now on eliminating racial disparities, I think it is time for us to maybe take a look at the issue again, especially because I do not know how they are going to measure eliminating racial disparities when we do not have data on most of the things that they want to eliminate or very limited. So, I think it is a time for us to kind of reconsider what can the committee do, if anything else. I would like to just have some time for the group to discuss it again.
DR. IEZZONI: That seems very fair. I mean, one of the things that we have done, I actually thought the breakout session in June was really excellent where we had Census come in and OMB come in and talk to us about the implications of the new classification on longitudinal tracking of experiences of racial and ethnic minorities. So, we have been trying to periodically hear from those people and kind of weigh in on what our feelings are about that particular issue, but suggestions for more specific and substantive things would be welcome.
DR. STARFIELD: Can I also suggest, too, it seems to me that the health objectives for the nation would be a good venue for doing that. Now, I understand that those are going to address inequities. I have not seen the report, and we have copies coming to us in the mail.
PARTICIPANT: [Comment off microphone.]
DR. STARFIELD: Oh, you have yours.
DR. AMARO: -- are open for public comment now.
DR. STARFIELD: Right, so we are all getting mail. That is a good way to do it. If there are objectives at reducing inequities, that is the way to deal with the data issue.
DR. AMARO: Yes.
DR. IEZZONI: I think given that it is a healthy people issue, it might actually be interesting to have this be a full committee. If we could have a few hours for the full committee, that would really tighten --
DR. STARFIELD; [Off microphone.]
DR. IEZZONI: -- the visibility around the whole issue.
DR. STARFIELD: -- eliminate disparities --
DR. IEZZONI: Yes, exactly, because I think that, you know, to be quite honest, Marjorie, the full committee meetings have not been as substantive in some sense, you know, around specific issues recently.
DR. GREENBERG: We have been trying to give more time to --
DR. IEZZONI: Right, the MPRMs and stuff like that.
DR. GREENBERG: The work groups and subcommittees can meet and then give adequate time to discuss recommendations.
DR. IEZZONI: Yes, it is just that it might be nice to maybe have a couple of hours on this. I think we would need to remind the full committee that this is an issue.
DR. AMARO: I just want to say, just to tell you, when you look through this report, and you look at the categories by race, and it says, "Not available. Not available. Not available," the data, even for something like diabetes for Hispanics.
DR. IEZZONI: Yes, that is --
DR. AMARO: You know, I mean, I feel like this is a committee-wide issue, and I would like to have everybody's thinking on it.
[Simultaneous discussion.]
DR. STARFIELD: When is the public comment period?
DR. IEZZONI: Right, exactly, it is.
DR. GREENBERG: I think one of the things that we are going to hear at that workshop in November is the frustration that states have in getting race and ethnicity data that they get from administrative records or encounter data.
DR. AMARO: Yes.
DR. STARFIELD: When is the public comment period over for this?
DR. AMARO: What does it say?
DR. IEZZONI: Because we should probably weigh in with public comments.
DR. STARFIELD: Absolutely. Right.
DR. GREENBERG: That would fit nicely.
DR. AMARO: I am looking for the date. December 15th.
DR. IEZZONI: Olivia has been waving her hand there.
DR. CARTER-POKRAS: I just wanted to point out, in addition to being able to send in written comments, we are also going to have five regional meetings starting in October. I think October 3 and 4 is Philadelphia, and then we go also to New Orleans, Chicago, Seattle and Sacramento. Seattle is going to be a special data session. Actually, they are going to have data sessions for everywhere except for Sacramento, but they are going to have a big data session in Seattle. Dr. Sondeck is going to be attending that. They have both an eliminating disparity session where we are going to get some input, and they are also going to have data sessions for everything except, as I said, Sacramento. So, it may be as far as timing to report out on that, we may want to do it either during or immediately after the regional meeting so that you can get input as to what we heard, the input that we heard at the data sessions.
DR. IEZZONI: That is a great suggestion.
[Simultaneous discussion.]
DR. CARTER-POKRAS: -- but before the public comment period has closed.
DR. IEZZONI: That sounds like a great idea, having somebody from us --
DR. GREENBERG: From the committee attend the Seattle meeting if that is the one where data is going to be emphasized the most.
DR. CARTER-POKRAS: Right. Data, as I said, data is in every single -- it is going to be in Philadelphia, New Orleans, Seattle and Chicago but not the Sacramento, which is the last one. I do have the dates, but they should be in the materials also.
DR. GREENBERG: They are here, actually.
DR. CARTER-POKRAS: Right.
DR. GREENBERG: Seattle is December 2 and 3.
DR. MOR: Hortensia, for those of us -- I plan to go through that, and Hortensia, I am sure that you are.
DR. AMARO: No, I am going to be in Israel.
DR. MOR: No, go through the document.
DR. AMARO: Oh, yes.
DR. MOR: Go through the document.
DR. CARTER-POKRAS: If you planning to attend, please sign up immediately because it is limited seating.
DR. MOR: As we go through that, it would be helpful, since all of us know an awful lot more about what data are available. I just find it kind of strange that there are probably some bits of information. I see this frequently where it says data are not available and I know data must be out there. They just have not been run. That may get back to the prior issue that we have discussed about inadequate availability of analysts. The data have got to be available.
DR. GREENBERG: On a national level are you saying?
[Simultaneous discussion.]
DR. FRIEDMAN: -- are two different things. I do not mean to be -- John thought about it(?), but there is an incredible amount, particularly on prevalence, that just is not available.
DR. IEZZONI: Okay. I think, though, that Marjorie, we need to look at the healthy people comment period as a committee and see if we can weigh in on it. When did we say that the comment period ends?
DR. GREENBERG: December 15th.
DR. STARFIELD: It cannot be in -- oh, December because the last hearing is in Sacramento in December, isn't it?
DR. GREENBERG: In December, that is right.
DR. AMARO: IF we can hear a report on the hearings and the data issues that people raise at the hearings.
DR. GREENBERG: Does the comment period end right after the last hearing? Is that what I was hearing?
DR. AMARO: It says here December 15th. Comments will be accepted from September 15 through 5 p.m. eastern standard time on December 15.
DR. STARFIELD: So, if we are going to -- I would think that you are on an advisory committee to the department, you already have their one presentation. Actually, Healthy People 2000 was on the agenda of this committee for quite a while, that they will not stop hearing from you on December 15th, but there is a meeting on November 12 and 13, and then there is not another full committee meeting until February.
DR. IEZZONI: Dan?
DR. FRIEDMAN: It might -- especially given that and also given how pressed committee and subcommittee time seems to be, it might be helpful for two or three folks to spend some pre-meeting informal or even public conference call time, certainly together with Olivia, identifying particular issues that we might want to bring to committee attention. I do not think that it necessarily needs everybody on the subcommittee at the same time. I think it also could make the committee process a little bit more efficient. So, I feel very comfortable filling Hortensia and Olivia's time.
DR. CARTER-POKRAS: Actually, our office has already drafted some questions that we certainly wanted to get some input along with the Disease Prevention/Health Promotion. We have a 12-page closing the gap newsletter that has been sent out to over 16,000 individuals and organizations. The whole topic is on healthy people, and we can make sure that if the national committee has not received copies that you get copies because it explains what is Healthy People, when are the regional meetings --
DR. FRIEDMAN: Closing the gaps?
DR. CARTER-POKRAS: Closing the gap.
[Simultaneous discussion.]
DR. CARTER-POKRAS: Hortensia, she must be on our mailing list.
DR. AMARO: No, Miriam brought it to me.
DR. CARTER-POKRAS: So, we will see if we can make sure that you get copies of them.
DR. AMARO: My only concern about -- and I want to work in the most effective way possible, is just that we get sort of new people looking at this, and it gets more integrated into kind of the general work of the committee.
DR. IEZZONI: David Takeuchi, unfortunately, could not be here today because his moving van was arriving in Indiana, but he clearly has this as a high priority issue and spoke to it on Tuesday, Hortensia, but unfortunately, he could not be here.
I think the basic point is that we should, as a committee, respond to the Healthy People in the context of closing the gaps as Olivia said. Once again, highlighting the issue of data around racial and ethnic minorities and the question is how do to that.
DR. CARTER-POKRAS: I should also point out that the work group - there are six work groups plus the data work group for the Eliminating Disparities Initiative. There is definitely overlap with the Healthy People 2010 goal of eliminating disparities. There are six health focus areas that we are looking at with the Eliminating Health Disparities Initiative, and they have a work group for each one of those. They have all been asked to look at the data, not only at the national level but also at state and local levels. So, they have found out, for instance, end stage renal disease.
We have data on whites and blacks and native American Indians and Alaskan natives, but we do not have any information on Hispanics, and that is a major complication of diabetes. That is an example of an obvious gap. So, those are things that also could possibly be brought to the subcommittee at a later time.
DR. IEZZONI: Yes, the point is that probably none of us have the time to go through this massive document and identify exactly what the kind of things that -- Yes, Barbara, do you have a suggestion for how we should do this?
DR. STARFIELD: We talked about this a couple of times, and I think it is not only the racial and ethnic disparities, but it is also the SES disparities.
DR. AMARO: Well, yes.
DR. STARFIELD: So, we have to include that, and I certainly will be glad to work with you on that.
DR. AMARO: But are there any goals related to that here? I do not even know if there are goals related.
DR. STARFIELD: We were assured there were going to be.
[Simultaneous discussion.]
DR. AMARO: That is what they said at the --
DR. IEZZONI: I am sorry, goals related to data?
DR. CARTER-POKRAS: Actually, the recommendation was made to each of the work groups responsible for drafting objectives for Healthy People 2010 that they break out data as much as possible by race/ethnicity for at a minimum the OMB Directive 15 or its caption categories. In addition to that, some measure of socioeconomic status, disabled, women, by age, that sort of thing. Now, some of the work groups did a better job than other groups, and that is the truth of it. What we find, as Dr. Amaro is bringing up, we are basically shedding light on the fact that data are missing. This is not a new story; this is an old story. It is just that we are seeing it because this is data not available, so it draws your attention to it in a way that it had not been drawn before. Whatever problems that we have for racial and ethnic data, it seems that it is even worse for the presentation by socioeconomic status indicators.
DR. IEZZONI: The question is, who has the time to go through the document and come up with a written response to --
DR. STARFIELD: I am going to go through the document anyway.
DR. IEZZONI: Barbara!
DR. STARFIELD: And Hortensia is, so --
DR. IEZZONI: And Hortensia. So, the two of you can flag things.
DR. CARTER-POKRAS: Maybe you could call David.
DR. IEZZONI: Dave Takeuchi.
DR. AMARO: Yes, and so the goal is for us to come up with a statement about --
DR. IEZZONI: We are responding.
DR. AMARO: Quotations of the fact that there is this -- sort of the problems that we see and then suggestions?
DR. IEZZONI: Yes, it is kind of like the open response letter. We are responding.
DR. AMARO: What you could do is if you could bring something, draft or whatever, to the November meeting, then probably get the approval to finalize it, if more needs to be done, in time by the December 15th deadline.
DR. AMARO: That sounds good.
DR. IEZZONI: Okay, so Barbara, I know that you go through these things because I have seen how you do it. Like the HEDIS document, I was really impressed, during your vacation.
DR. STARFIELD: So was I.
DR. IEZZONI: Hortensia, also the OMB Directive 15. So, why don't the two of you -- and Caroline, can you coordinate getting the comments from them?
MS. RIMES: Sure, and I think we will pass it to Olivia --
DR. IEZZONI: To Olivia, or Olivia, your new staff person, Sylvia? No, what was her name?
DR. CARTER-POKRAS: Sheila Merriwether.
DR. IEZZONI: Sheila. I knew it was an "S" name.
MS. RIMES: If you would not mind doing that, I would appreciate it.
DR. CARTER-POKRAS: No, we would be happy to coordinate informally as well as make sure it goes through the formal channels.
DR. IEZZONI: Once we get the letter from Don Detmer, we will want it to go through the formal channels, but at this point, it is more coordinating.
DR. CARTER-POKRAS: There may be opportunities to bring up some of the issues during some of the regional meetings if some of you are in attendance.
DR. GREENBERG: I really encourage you to include David, too.
DR. IEZZONI: Okay, and David. David Takeuchi.
DR. GREENBERG: He is overcommitted in the rest of his life, but he is not overcommitted to this committee yet.
[Laughter.]
DR. IEZZONI: Oh, he will be. We will work on that.
DR. AMARO: Can I just ask a question about whether mental health and substance abuse goals, whatever, was included in here to the satisfaction of SAMSA or is that an area we should look at also?
DR. ROUSE: We have some problems just like the other areas, but there are areas in there where we are addressing the data needs.
DR. AMARO: Okay, so it might be helpful to know if there is a list of issues that we might be able to look at in case we can integrate any of those in our --
MS. RIMES: Beatrice, why don't we include you with Olivia and the other people in working that?
DR. ROUSE: Okay, I would be glad to.
DR. AMARO: That would be great.
DR. GREENBERG: May I ask just a yes/no question?
DR. IEZZONI: Yes
DR. GREENBERG: Is there a mental health/substance abuse component to this whole post-acute care issue?
DR. AMARO: That is very important.
DR. NEWACHECK: That was not an easy yes/no question.
DR. MOR: It is not framed that way, but it is always a complicating factor.
DR. STARFIELD: I would think so.
DR. MOR: So, if one is doing any kind of risk adjustment, it is a big issue. I assume there are social factor issues that are related to rate of discharge for preemies and other complications associated with children with disabilities.
DR. NEWACHECK: Probably 85 percent of the children with disabilities who are in institutional settings are there for mental health related reasons and not physical problems.
DR. MOR: Mental health rehab, too, all of those facilities.
DR. AMARO: And substance abuse treatment, is that?
DR. MOR: Substance abuse treatment would not fit into what Kathy and I have been thinking about in post-acute world because that is largely --
DR. AMARO: Where does it fit?
DR. MOR: Not specifically in this. It is not something that comes following hospitalization, which is mostly what -- or independent of whether it is a hospitalization, it is something of the emergence of some kind of an acute event or an exacerbation of a chronic condition that requires more aggressive medical rehabilitation.
DR. IEZZONI: Olivia?
DR. CARTER-POKRAS: Without getting myself hopefully into trouble with my colleagues, I am looking at the calendar. The meeting for the National Committee on Vital and Health Statistics is November 12 and 13. By that time, we will have Philadelphia, Chicago and New Orleans' meetings over with. It will just be the big data meeting in Seattle that will not be there, so we could potentially get sort of like an interim report to the subcommittee if that is of interest about the data recommendations that have been made at the regional meetings for Healthy People 2010. In addition to that -- Richard Kline is the one who is organizing that at NCHS, the data sessions.
In addition to that, we may want to have the data work group for the eliminating health disparities initiative come and report out because the work group reports are supposed to be finalized by the end of this month. So, that gives them some time to prepare, sort of pull out the sections about the data recommendations from each of the six work groups and present that to the committee. I am just tossing that out as possibilities for the November meeting.
DR. IEZZONI: That sounds really great. I am just worried about -- we had thought that --
DR. GREENBERG: I wonder if that would be a good thing for the full committee although --
DR. IEZZONI: The full committee.
DR. GREENBERG: -- I am worried because we are totally limited to two days unless we decide to have a Saturday meeting, which I --
DR. IEZZONI: No. No, no, no.
DR. GREENBERG: -- am totally against. Even evening meetings are not a good idea.
DR. IEZZONI: No. Except for Don.
DR. GREENBERG: Except for Don and John. You probably want what, a few hours?
DR. IEZZONI: We are going to need to work on the Medicaid Managed Care Report at that meeting.
DR. GREENBERG: You are going to need a few hours. Every subcommittee and work group --
DR. IEZZONI: And probably the charge.
DR. GREENBERG: -- wants a few hours.
DR. IEZZONI: It does not look like we are really going to get to that today. Probably the charge as well. Olivia, I think that the full committee -- having a report about this to the full committee would be --
DR. GREENBERG: It is very timely. If you want to have recommendations that you actually consider and get in by the December 15th deadline --
DR. IEZZONI: Yes, you need to have it by November.
DR. GREENBERG: Do you think that deadline is important even for the committee?
DR. CARTER-POKRAS: There usually is a little give, especially since it is approaching the Christmas holidays. There is a little give, but they need to make sure they build up enough time to have the work groups then respond to the comments, and just reading the comments is overwhelming.
DR. GREENBERG: So, February 2 and 3, that is --
DR. CARTER-POKRAS: no, that would be too much give.
DR. GREENBERG: We are talking about a timely thing here.
DR. IEZZONI: We could have a draft letter that we would be putting up at that November meeting for Don's signature that we would want full committee approval for, and it would be nice to maybe have the report of the work groups right before the discussion of the draft letter.
DR. GREENBERG: That is what I am thinking.
DR. CARTER-POKRAS: Certainly if we notify them now, that gives them plenty of time to prepare because that is going to be absolutely critical.
DR. GREENBERG: You will work with them.
DR. CARTER-POKRAS: Right, I will work with Marjorie and Richard and also Jack Anderson. He is the co-chair for the Data Work Group.
DR. IEZZONI: Yes, Kathy?
MS. COLTIN: After the mid-November meeting of the full committee, we do not have any meetings for this subcommittee --
DR. IEZZONI: That is right.
MS. COLTIN: -- or the full committee until the beginning of February. Do we want to poll for meetings in December or January?
DR. IEZZONI: We do. We do. That is what I was implying by saying that we need to get things on our books and we need to get a sense from Vince and you and me and approval from the whole subcommittee about what we want to do. We need to get dates.
MS. RIMES: We will do that.
DR. IEZZONI: You have a list of good stuff to do.
MS. RIMES: Right.
DR. IEZZONI: Paul?
DR. NEWACHECK: I have to leave in a minute because I have a cab waiting downstairs to take me out to Dulles. You had asked us earlier if we were in a position to approve the charge. I agree with it, and I feel it is a good charge. I wanted to suggest two potential inclusions, though, or two additions that we might want to consider. One is an item that this committee is in a position to serve as an advisory body to NCHS on population-based issues. This came up earlier in the discussion when we were talking about the why NCVHS was not considered to be the advisory body in the discussion of IOM Report.
DR. GREENBERG: I really wanted to engage you in some discussion on that.
DR. NEWACHECK: I think having that as a specific function of our committee on population-based issues would be important. We have it kind of indirectly through the work on the visioning process, but to have a specific statement that we view ourselves as an advisory body to the NCHS on population-based issues, I think, would be a useful statement or a useful part of our charge.
DR. STARFIELD: This committee was originally, as you know, advisory to NCHS, and that was felt to be limiting. It really interfered with our ability to deal with the other agencies. So, I would really be careful about doing that.
DR. FRIEDMAN: I think it is important to work out that relationship and to have that kind of relationship. At the same time, I do think that it takes some discussion to the extent that there are other agencies also producing population-based --
DR. GREENBERG: Maybe I am missing something here, but NCHS has asked -- Ed Sondeck has asked the national committee to serve in that capacity.
DR. FRIEDMAN: Right.
DR. GREENBERG: It was my understanding that this whole effort that we are doing for the 21st Century with SINSTAT and NCHS is to kind of kick off that, jump start that process, and that is something that, at least under the current chair, the committee has committed to.
DR. STARFIELD: That is why we put it into the broader framework. We do regard ourselves as advisory, but I think to put it in there specifically would interfere with our ability to --
DR. NEWACHECK: I am not sure about that. I think that --
DR. STARFIELD: We have history to show it.
DR. NEWACHECK: It seems to me that we can provide that assistance, and I would hate to have another situation come up where NCHS goes to somebody or IOM points to some other committee or some other group to comment on something that is as important as this, as has happened in the past.
DR. STARFIELD: The IOM was not only commenting about NCHS. It was more a data issue.
DR. GREENBERG: Actually, it was the national committee on a subcommittee chaired by David Mechanic that suggested that IOM do this study. That is exactly how it came about. So, there is synergy here. I thought there was a commitment to provide advice to NCHS in this area over the long term.
DR. NEWACHECK: Isn't that on behalf of the full committee as opposed to just the subcommittee?
DR. GREENBERG: It is on behalf of the full committee, although to the extent that you are looking at population-based data, it is probably going to fall more on this committee.
DR. NEWACHECK: Right. I think what I was suggesting was that we make that statement for our subcommittee just with regard to population-based data, but it does not sound like there is a consensus around the room that that is something that we would want to go on the record as saying.
DR. GREENBERG: I wonder, though, if you need -- in this section here on the 21st Century vision for health statistics, you need sort of another sentence that says this will jump start or kick off an ongoing role for the committee.
MS. COLTIN: Actually, I think it falls under the heading of population-based data.
DR. GREENBERG: That is where I am.
MS. COLTIN: The goals for the nation and the enhancement of health surveys.
DR. GREENBERG: Right.
MS. COLTIN: Rather than under the developing --
Dr. GREENBERG: Wherever, yes. It would not be under that.
MS. COLTIN: All three of those.
DR. GREENBERG: That was the objective as I understood it. I mean, there are suggestions that there should be a separate advisory committee, that SINSTAT should be advised. There have been lots of different suggestions made and generally, I think, our position at NCHS has been that we have the National Committee on Vital and Health Statistics set up a completely different advisory committee on that broad area does not make any sense.
MR. HITCHCOCK: If I could just interject something. I would be reluctant to emphasize totally the nature of this committee to NCHS because that gets us back into the position again where we were a few years back where we were perceived as sort of an NCHS committee. We do not want to be that again, I do not think we want to be an advisory group.
DR. GREENBERG: I hear what you are saying, and I agree, but how do we mesh that with this kind of commitment that has been made with Dr. Sondeck? I mean, I guess there is just this disconnect.
MR. HITCHCOCK: That is just one item. that is just one item, and I think that one item is fine.
DR. GREENBERG: Not that that is the only thing that this committee is, by any means --
MR. HITCHCOCK: Right.
DR. GREENBERG: -- but that it is one component of the committee's work.
MR. HITCHCOCK: Yes.
DR. IEZZONI: Dan, did you have a comment on this?
DR. FRIEDMAN: Yes. As I hope Marjorie knows, it is an extremely important role for the committee, and frankly, my only concern is that we may not be able to resolve it today; maybe we can. I just do not want to see any statement in there giving the impression that the role vis-a-vis population-based data would be limited to NCHS. That is my only concern because obviously, while NCHS is an extremely important part of that, there are obviously other parts as well.
DR. GREENBERG: Right. Yes, obviously.
DR. FRIEDMAN: Like the CDC, et cetera, et cetera.
DR. GREENBERG: We scarcely have the budget to be the only game in town.
DR. FRIEDMAN: Right.
DR. GREENBERG: Maybe it is self-evident here, Paul.
DR. NEWACHECK: I am not sure that it is. I feel that it is something that we should be doing, and it is really an important mission, I think.
DR. FRIEDMAN: Maybe we should spend a little bit of time, I do not mean now, but we could --
DR. NEWACHECK: Yes, we could --
DR. FRIEDMAN: -- try to come up with a sentence or two.
DR. NEWACHECK: Yes, we could do that.
DR. GREENBERG: It is something that the committee does more of.
DR. IEZZONI: It seems to me if you want to do that, we need to have a very clear work plan for doing that. We need to have, okay, every six month we will monitor this in some way or there has to be some periodicity, and there needs to be some specific activity. Ed Hunter was right. I remember when he came and talked to the full committee, I think it was two years ago, about survey integration. It is nice to hear from him, but if we put something in our charge, I think we really need to have a plan in place to follow up on it. So, that is maybe what you and Dan could talk about, as well as how we could put, given the resources and time limitations that we have for a very important issue and an interesting one, how we could effectively do what you are suggesting that we do.
DR. NEWACHECK: The other thing that I would like to put on the agenda for future consideration, not for today, is data needs for disabled persons.
DR. IEZZONI: Yes, I agree. We have tried to track that a little bit. You know, for example, Social Security is completely redesigning the way that they are evaluating disability, which is kind of scary what is happening, frankly. That should be something that we should monitor.
DR. GREENBERG: Gooloo is the head of --
DR. IEZZONI: Gooloo is the head of the IOM evaluation of the SSA, but the contractor on the SSA evaluation right now is really trying to figure out where they are going with this, and it is very complicated.
MS. RIMES: They have not selected a contractor yet.
DR. IEZZONI: There is this institute for Michael Kane.
DR. GREENBERG: That is another issue.
DR. IEZZONI: Okay. That wa not on the record maybe. Paul has to leave. I do not think that we can deal with the charge right now because so many people have kind of left and are leaving, and people like David Takeuchi are not here and so on.
MS. RIMES: Do we want to get people's comments on the charge and then maybe be ready? I mean, I have some comments. I have Aaron's comments, and I did get some from --
DR. IEZZONI: While Barbara is here and Marjorie, I do have to raise a question about what we mean by the core data elements because there is this quote, "The subcommittee will continue to work on the core data elements for the purpose of making recommendations about several elements that were previously considered not ready for implementation." I do not know whether I feel comfortable having this.
DR. STARFIELD: Well, we have unfinished business in the committee.
DR. IEZZONI: Excuse me?
DR. STARFIELD: We have unfinished business in the subcommittee, and the other subcommittee is not going to do it, so we need to do it.
DR. IEZZONI: Well.
DR. STARFIELD: It has been a two year lapse. If we thought it was worthwhile to do it in the first place, we need to finish the job.
DR. IEZZONI: I know, but things have changed since --
DR. STARFIELD: Like what?
DR. IEZZONI: -- two years ago. Membership on the committee, what people might think is important, the way that it is viewed in the world, the whole goal for having it, the context. So, I think if we are going to have it as part of our charge and work plan, we need to talk it through a little bit more.
DR. STARFIELD: That is fair enough, we can review it, but we cannot just drop it. Because we have it. It is there as our policy.
DR. GREENBERG: It really built on about a 25- year history that frankly is still playing out in these billing committees and other places where --
DR. IEZZONI: Exactly, exactly, and the data centers.
DR. GREENBERG: -- hospital and ambulatory data are being defined and will --
DR. IEZZONI: See, that is frankly what --
DR. GREENBERG: -- play out in other ways, too, so.
DR. IEZZONI: That is why I think it really belongs in the data standards area because we are talking about definitions, code sets, how you define data elements, and it seems to me that --
DR. GREENBERG: They are not going to address it.
DR. IEZZONI: I know that they are not.
DR. GREENBERG: There is no doubt.
DR. IEZZONI: So they punted it to us.
DR. STARFIELD: They did not punt it to us. I mean, we punted it to us. They are not dealing with the population issues. The core datasets have at least a population element.
DR. MOR: I do not understand. I mean, I remember this material quite well. I thought the committee, when HIPPA started, was going to adopt all of those code sets that had been firmed up within the core data elements, and that obviously not all of them did get set up, but that there are additional things in the history of HIPPA and the next round of data elements that are going to be recommended as part of code sets that would be pushed through. Does that mean to say that the Data Standards Committee just does not want to deal with that anymore or they only want to deal with diagnosis and CPT codes or what?
DR. GREENBERG: For example --
DR. MOR: Because I do not think that it is a population issue only.
DR. STARFIELD: Maybe we just need to ask Don to resolve who chairs the issue.
DR. GREENBERG: Lisa, you and I were discussing that one of the real problems with hospital data is the lack of this qualifier for whether secondary or other diagnoses were present at onset or not present at onset or uncertain. That was recommended in the core data elements; it is not in the encounter datasets. It is something that New York and California are trying to push and will come up at the HIPAA workshop. Does the committee want to have -- and actually, New York was able to get some kind of recognition of this at the ANSI X-12 meeting recently because the National Committee had recommended it and actually got California to agree to the definition in the Core Data Elements Report. I mean, I do not know. They still sort of live.
DR. IEZZONI: Marjorie, I thought that the 1992, I think, recommendation for adding the modifier was a UHDDS recommendation not a core data element.
DR. GREENBERG: It was carried forward in the core -- that is what the core data elements did. It basically carried forward recommendations from those earlier periods.
DR. IEZZONI: Yes, but it was both ambulatory care and hospitals. I think that it is more kind of a sense of history, I guess, is what you are saying. We have done this in the past, and so we should continue doing it. It is just that I am not sure what this means --
DR. STARFIELD: This is the policy of the committee right now. Okay, if we want to change it, we have to change it, but we should finish it also. It is the policy of the committee; it is there.
DR. IEZZONI: Kathy, did you have a comment about this?
MS. COLTIN: I agree with Barbara and Marjorie on it. I think there is some unfinished business on it. I do not see it as a major initiative of this committee.
DR. IEZZONI: Well, that is what I wondered, what it meant operationally.
DR. GREENBERG: It is sort of a monitoring and weighing in.
MS. COLTIN: I think we need to check in. I think what I would like to see, very succinctly, is just to take a list of the core data elements and the recommendations that we have, match it up against where we are in the standards in terms of whether these data elements are in the datasets, what --
DR. GREENBERG: That was done in an article in the Journal of the American Public Health Association.
MS. COLTIN: Well, it is and it is not. I think part of the issue is, is it a required element in the dataset or is it an optional element. If it is an optional element, what is the position that we would like to take in terms of encouraging its use. So, that sort of thing, I think, is --
PARTICIPANT: You mention your November meeting is also going to come up with data needs that may not be met by what is in the core data elements now, and some of those may be in the post-acute area right now.
DR. IEZZONI: I guess I really need to understand why John Lumpkin's subcommittee is not working with us on this because I really think that --
DR. GREENBERG: Because they do not seem to want to look at anything at that level of detail.
MS. COLTIN: I kept getting them to try to look at content. At one meeting, they tried to do it, and it was like -- it was just --not done.
DR. STARFIELD: One of the reasons that I switched committees was because of this issue, because they could not deal with things like that.
DR. IEZZONI: we just need to come up with an operational plan for how we deal with this in the context of everything else that we are doing, and I really do think that it needs to be integrated somehow with the standards process. Maybe what it would be is that we would make a recommendation to John Lumpkin and say you guys need to consider this as part of the --
DR. MOR: We can specifically incorporate this perspective in the subacute, and it so happens that all of the --
[Simultaneous discussion.]
-- data elements that I know of in the three existing systems actually probably have virtually all of the code and have, obviously then, conflicting definitions and operational code sets for the functional status. All of the others have -- they are there.
DR. STARFIELD: I already volunteered to work on it in the context of the vision for the 21st Century, and I am sure --
DR. IEZZONI: Great, so we will have a list of what is there, what is not there.
DR. GREENBERG: I really see it as Kathy mentioned it, though, kind of a monitoring, support, changing. I mean, functional status is going to have to be addressed in the post-acute care.
DR. IEZZONI: That is why I mentioned it earlier as an unresolved piece of business.
DR. GREENBERG: It does not require a big initiative.
DR. IEZZONI: Okay. All right, we still need to vote on the charge. I do not think that there are enough people around the table right now to do that, but we are going to need to make sure that people get comments back to Carolyn. So, Carolyn, could you once again ask everybody who is not here to get comments back to you about this?
MS. RIMES: Even people who are here we will ask again.
DR. IEZZONI: Okay, thank you. So, we are going to adjourn.
DR. AMARO: We need to identify somebody from the committee who will go to one of those meetings. I do not know if we want to ask David.
DR. GREENBERG: You do not have to. I was just saying that --
DR. AMARO: No, I think it is a good idea.
DR. GREENBERG: We could certainly support one member to go to one.
DR. MOR: Elizabeth lives in Seattle.
DR. IEZZONI: Yes, Elizabeth is in Seattle.
DR. GREENBERG: You are right!
DR. MOR: Elizabeth is right in Seattle, and it would be a data thing, and it would be something of interest to her.
DR. CARTER-POKRAS: Yes, good idea.
DR. GREENBERG: That would be great. Ask her.
DR. MOR: When is the meeting in Philadelphia?
DR. CARTER-POKRAS: Philadelphia is October, I think the 3rd and 4th.
DR. GREENBERG: Oh, too soon.
DR. IEZZONI: Thank you, everybody.
DR. GREENBERG: We could just send Elizabeth an email.
DR. CARTER-POKRAS: I am sorry, it is October 5 and 6. Then the next one is New Orleans October 21 and 22.
DR. GREENBERG: A nice time to be in New Orleans.
DR. CARTER-POKRAS: Yes, that is a tough duty one. Chicago is November 5 and 6.
[Whereupon, at 3:15 p.m., the meeting was concluded.]