Hubert H. Humphrey Building Room 337A
Washington, D.C.
PARTICIPANTS:
Lisa I. Iezzoni, M.D., M.S., Chair, Associate Professor, Harvard Medical School
Hortensia Amaro, Ph.D., Professor, Boston University
Richard K. Harding, M.D., Richland Memorial Hospital
Vincent Mor, Ph.D., Brown University
M. Elizabeth Ward, Washington State Department of Health
George H. Van Amburg, Dewitt, Michigan
Carolyn M. Rimes, Key Staff
Olivia Carter-Pokras, Ph.D., Office of Minority Health
Patricia Golden, National Center for Health Statistics
Ronald Manderscheid, Ph.D., Center for Mental Health Services
TABLE OF CONTENTS
Discussion on Federal Data Collection on Race and Ethnicity
P R O C E E D I N G S [1:10 p.m.]
Agenda Item: Welcome and Introductions and Discussion of Agenda and Hearings in San Francisco on Privacy, Confidentiality, Standards and Coding
DR. IEZZONI: This is the meeting of the Subcommittee on Population-Specific Issues. I am Lisa Iezzoni, from Beth Israel Deaconess Medical Center in Boston. I will be chairing this afternoon's session.
[Introductions were made.]
DR. IEZZONI: The first item on the agenda is just reminding people that we were in San Francisco during the 2nd to 4th of June for our Subcommittee and other working groups from the National Committee on Vital and Health which held hearings in San Francisco. They were great, very successful hearings. We had diverse and extremely thoughtful input from 44 different presenters over the course of two days. It was very intensive.
Tomorrow, hopefully, we will have a copy of the agenda so that you all can see the span and representation of the people who spoke to us over those two days.
Just in summary, we heard people from public health, from the variety of advocacy groups, from the business community, clinicians, data producers and processors, people representing mental health interests, disability. The staff is really to be congratulated on having put together a superb meeting.
Carolyn Rimes led the charge. Lynette Araki also participated in that as did -- I saw some folks from California. Who was particularly helpful to you from California?
DR. DETMER: Ann Dyer and McCafrey.
DR. IEZZONI: I think, Don, I might be overstating it a little bit, but maybe not -- it felt good to get outside of Washington, D.C.
[Laughter.]
DR. IEZZONI: I know that that is a very qualitative assessment.
[Laughter.]
DR. IEZZONI: It just felt different. What was very interesting is that even the committee members I think at the first break were kind of looking at each other saying this feels really different. You know, I think part of it was that people felt more able to be open about barriers that are confronting them and hurdles that they face.
I think that it was interesting. Because I do not think that we substantively heard much that was hugely different from what we have heard throughout the hearings this winter and spring in Washington, but the rationale, the context, the justification was a lot more kind of concrete and real-world based. So it was really very productive.
Unfortunately, George, and Hortensia, and Elizabeth could not come. But, Jim and Don, do you have any comments about the hearings?
DR. DETMER: Just to underscore what you said. I thought that it was worthwhile. I think we got a little different flavor at times.
DR. IEZZONI: There were some differences. But I think, in general, that the major issues -- it is very helpful to have kind of just reinforcement of what some of those are.
DR. SCANLON: The other perspective was to have these hearings -- rather than having separate hearings on standards, and then privacy and population-specific issues, we sort of put all three together so that we have the witnesses describing the issues from all three perspectives, and that worked out very well.
DR. IEZZONI: Yes. That is a very good comment, Jim. A number of people thought about that because it did allow for a little bit of give and take, whereas, ordinarily, we would have these kind of content-specific meetings that have not given us the diversity of issues to kind of provoke other people to their comments.
Do the two or three of you have any questions about he hearings? I did not think that it was really worth going too much more into detail. All of the testimony is on our Web page I believe, and there is a huge amount of it, like volumes.
DR. DETMER: You might want to do a search actually.
DR. IEZZONI: Yes.
DR. AMARO: I have a question.
DR. IEZZONI: Yes.
DR. AMARO: I was starting to discuss it with you earlier, but I wanted to bring it up for us to discuss and for the others who were there to also hear from you. I was wondering, one of my concerns that I expressed last time around the kind of technology that all of this will require is the impact on agencies that do not have the technology that it is going to take to do this. Since the populations who were concerned tend to be served often in agencies and in institutions that do not have all of the resources, I was wondering whether you heard any testimony that sort of helped us to get a handle on that and whether that will be an issue, how it will be an issue, and how people think that that will impact them?
DR. SCANLON: We heard from the Los Angeles County area.
DR. IEZZONI: Yes. That is what I was saying to Hortensia. Los Angeles County gave a very eloquent and very concrete, and very dire, I think, actually statement about the financial losses that they anticipate incurring this year, with a very clear statement by the woman who gave the testimony that, if they try to divert resources to improve their computerization and change their systems, their electronic information systems, that that would take away from patient care.
What I would like to do, Hortensia, is bring that up as an issue when we talk later about our work plan.
DR. AMARO: Uh-huh.
DR. SCANLON: Uh-huh.
DR. IEZZONI: Because I think that that is a very important --
DR. AMARO: Uh-huh.
DR. SCANLON: Yes.
DR. IEZZONI: -- question. Don, did you want to comment on that?
DR. DETMER: It is really frustrating because I think there is a tendency at least when you talk about what some of the privacy is, there is one area, in particular, where I think people want it, but it does not come free, and, particularly, if you look at some of these decisions like unique identifiers, and so forth. There clearly really is a real tension. It is not an imaginary thing, and yet I think there is a tendency to disregard that on some of the people's wish lists. I think that we really are going to have a tough one on that.
DR. SCANLON: We heard from others both on the standards side, Hortensia. And, on the privacy side, we heard from groups that normally would be considered disadvantaged in one way or another. Los Angeles County, as well as -- we heard from parts of the California State Health Department as well, and from some providers that are basically public providers. It was not so much that they -- the idea of what the standards -- that there should be standards and what they should be, there was not that much disagreement. But I think, as Lisa was saying, that it is a matter of when.
Apparently, the Los Angeles County Health Department, for example, is also grappling with this year 2000 millennium problem with dates. Obviously, their budget is fairly tight. In fact, HHS helped to bail them out this past year -- the Los Angeles County Health Department. So, that was their concern. It was more of a timing issue and a resource issue.
DR. IEZZONI: Yes. They definitely want to do it but they do not have the money right now to do it.
DR. SCANLON: That is right.
DR. AMARO: Right.
DR. IEZZONI: You should look up her testimony. It is only a few pages long and it is extremely compelling and concrete.
DR. AMARO: Well, we can discuss this some more when we come back to it.
DR. IEZZONI: Yes.
DR. AMARO: Okay.
DR. IEZZONI: Okay. Well, we dispensed with the first item quicker than we had anticipated, which is a good harbinger for the rest of the afternoon. But what is not so good is that I do not see Marjorie Greenberg here for the Core Data presentation. I see Lynette Araki.
DR. ARAKI: I know she is coming.
DR. IEZZONI: Well, why don't we then jump ahead to the next item and then interrupt ourselves when Lynette and Marjorie are available to give their presentation. Because I think really that what they are going to talk about should be talked about first. Because one of our agenda items is what should we do about core data elements.
Let me just talk a little bit about this third item on the agenda which is really trying to talk about our work plan and what we are going to do over the next months and year.
I would like us to be extremely concrete and, by the end of the next two days to actually have specific things that we want to do when we are going to do them, what the products of each activity will be, and a sense that we can actually pull it off given our resources, our time, and our energy. So, I would like us all to be really honest in our discussion today.
In terms of our work plan for the next year, the way that I have thought about this is that we have two broad categories of work that we need to do. The first is unfinished business from the old kind of NCVHS structure and committees, and the second is new initiatives. So, when we think about unfinished business, there were four areas that I focused on as particularly major areas of unfinished business.
The first, I think, George, relates to you, and those are your recommendations on state and community statistics that you have faithfully been trying to get us address for months now.
DR. VAN AMBURG: Right.
DR. IEZZONI: I am glad that you have not gone away. But that is one of the things that we need to address.
The second is the OMB Directive 15. We need to be able to understand what the issues are. The presentations this afternoon will help us with that, as will the presentation on Wednesday from OMB.
And we need to write a productive and thoughtful response. We need to do that this summer.
The third piece of unfinished business is the core data elements project. Ron Manderscheid was involved in developing the core data elements for mental health and substance abuse, and Carolyn was the staff person for the long-term care and disability. We have each seen the documents that have come from that effort, but it still feels like a loose end that it has not been tied back into what the committee is doing, and that is a piece of unfinished business.
And then the fourth piece of unfinished business, I guess, we have a little bit of information on now and that is, Hortensia, your concern about the Spanish translation of the various surveys that we are interested in.
So, from my point of view, those were the four major pieces of unfinished business. I am sure that there are other things that might be added to that, but just having reviewed the materials that Carolyn prepared for us, that seems to be it.
Now, we are going to stop and talk about these before the hour is over. Since Lynette and Marjorie are not here yet, to give the new folks, Elizabeth and Hortensia, a sense of what the core data elements are so that we can really begin to have this conversation about what shall we do with them, maybe I will go on to just what my thoughts were about the new business, just so that you all can have it in the back of your minds kind of percolating over the next minutes and so that you can begin to think about how you feel about this.
For new business, I break this into two types again. The first type is discrete, focused issues that arise and need immediate attention. We need to have the ability to be able to respond quickly when such issues arise.
DR. ARAKI: She is on her way.
DR. IEZZONI: She is on her way?
DR. ARAKI: Yes.
DR. IEZZONI: Okay. Great. Hopefully, she will be here in three minutes.
Examples of this include HCFA's decision about home health agency data that we have just learned about and that we will be hearing a presentation about tomorrow during the Subcommittee's breakout session. A second issue is concerns about data release policies across the department of health and human services, and especially within HCFA that is about to make it very difficult for people outside, especially researchers to get access to data, especially, again, from HCFA.
This is an issue that has been kind of gestating, again, over the last few months, but is now kind of becoming more concrete with various draft proposals circulating around HCFA and so on. This is topic that I think we need to talk about, but we might want to do so in conjunction with Barbara Starfield's group. Okay.
A third issue is Hortensia's question about expanding data systems to Puerto Rico. It is a very discrete issue and one that is very focused, but, nonetheless, an issue that we should talk about. So, for these new business items, I really think what we need to do as a committee is be nimble and be kind of able to be tracking what is going on out there that we need the committee to think about. And the product would basically be for each of these a letter to the Secretary, not a report.
Okay. So, again, these would be very focused. A letter would be the product. We could maybe deal with some of these issues in an afternoon or a few hours. But, nonetheless, these are important concerns that are arising that we need to be able to address.
Okay. The second type ne new issues, I think, are broader issues about health data in the United States. Here I would like to go back to the metaphor that Kathy Colten and I kind of came up with as the old committee was transitioning to the new committee, and that is the camera analogy. Those of you who have been around for a while will remember this tired old camera that we trot out periodically.
But, for the rest of you, basically, Kathy's and my notion was that the first thing we need to do is ask what do we need to learn about the health of the American Public? Not until we ask that question can we begin to ask the questions about what data do we need to have to be able to look at these questions. And then we focus increasingly with our zoom lens down to what are the standards for producing those data.
So, I think what I envision as this broader category of initiatives for our committee is finding one or two health questions about the American public that are broad that we would want to really focus on because they are important and they reach a lot of people. We cannot bite off too much. But what we would hopefully do around these one or two issues is hold hearings, maybe have an outside contractor help us to do some in-depth research into the topics, and actually produce a substantive report as well as probably a letter with recommendations to the Secretary. There are, let's see, six candidates that I have come up with from having e-mailed the committee members and from having thought about the agenda. This is what we will talk about later. Let me just lay them on the table for you all to be thinking about whether they are additional, potential broad questions, or whether these are not ones that you would like to look at and we should think further. Number one, is the impact of Medicaid Managed Care on the Health of Medicaid beneficiaries. The second question is to look at the health of children in the United States. We all know that child health is a very hot political issue right now, especially under Welfare reform, with changes in Medicaid. I do not know that our committee has ever really focused on child health in recent years. I think that it is a big issue that needs to be looked at.
The third question is one that we heard about very eloquently in our California hearings and that is, as we collect data on health of Americans, we have tended to collect it in English, and sometimes Spanish. Especially in some regions of the country, there are many other ethnic and language minorities that are not being counted but, nonetheless, have very important issues, such as Asian and Pacific Islanders, Haitians, people from the Middle East, people from other countries who do not speak English or Spanish. So, the question being how do we work with states, for example, to identify ethnic populations that they need to know about and try to get information about their health?
The fifth question is just information about the health of people who are under and uninsured. I know that that was something that George was particularly concerned about. These are often working people who do not have adequate health insurance.
Then the sixth item that I have added based on Hortensia's comment was how do we think about the fact that some of the K-2 recommendations that we are going to be making will require the expenditure of resources, money, basically, to implement, and that there are a variety of very important health care delivery systems for vulnerable and disenfranchised populations that are particularly going to have trouble meeting some of those electronic data standards.
Let me stop here. We will obviously go back to this as we, as a group, start to think about which of these or which other issues you would like to focus on.
We will go back now, since Marjorie and Lynette are both here, to this issue of unfinished business for the committee. Let's talk about the core data elements.
Basically, what I am hoping that Marjorie will do is kind of give us, again, the context in which the core data elements happened, talk a little bit about the full national committee's core data elements project, and then, Ron, and Carolyn, if each of you could maybe at the end say a few words about your respective core data element projects that you were involved in fostering and moving forward, and maybe give us a sense of what you think is necessary for you to feel that we have tied it into the broader picture. Okay. Marjorie, could you give is a start?
Agenda Item: Introduction to Core Health Data Elements and Segue to Introducing OMB Standards for Classifying Federal Data Collection on Race and Ethnicity
DR. GREENBERG: I was not quite sure what you wanted me to do. What you described was my assumption basically. What I thought I would do is I gave a presentation at the Medical Record Institute's Meeting on -- I was asked to make a presentation on the core data elements project. I did that in the context of the electronic patient record, which was what the conference was about, Towards and Electronic Patient Record, and in the context of the Health Insurance Portability and Accountability Act of 1996. I did not think you wanted that whole presentation. It is too long for you all.
So, I kind of made some selected overheads from it. I may have selected some that are not of interest, and I may have left out some that are. If I have a little bit of time to go through that, why don't I do that. Just feel free to interrupt me with questions. You, Lisa, and George, were there. Particularly George was there at the beginning, going back to some of the data sets that this was -- that the core data elements recommendations were based on.
But, Hortensia, and Elizabeth, and Don were not, and so some of this might be a little bit repetitive. I think that these contexts are kind of important for the continuing work of the committee. Is that okay?
Obviously, because of the purpose of this presentation, but also, I think it is important from the point of view of looking at the broader mission of the committee, and its current and future responsibilities. But I presented this to the conference as kind of a shared vision, going from core data sets to electronic medical records, and sort of everything in between. This shared vision includes several things. It includes common core data elements, uniform definitions and code sets, unique identifiers, a data dictionary, compatible terminologies and classifications, and privacy and security framework, with the last one almost probably being the one that should be on the top.
But I wanted to put this in this context because I think there has been a debate between core data sets and the electronic patient record. I really feel that they are part of a continuum and that working on one does not diminish the benefits or diminish work on another.
I had some other introductory information, at that point, including what the National Committee on Vital Health Statistics was, and I do not think that I need to do that with this group. So, if it seems a little bit like I am skipping ahead, I am for that reason.
So, having placed core data elements and core data sets in the context of the broader needs for electronic medical records, I then returned to the issues at hand, very closely at-hand, which are the administrative simplification provisions of the Health Insurance Portability and Accountability Act of 1996 and what its requirements were. Again, I grabbed this one because, without going through all of the legislative language, what it resulted in in the Department was six implementation teams, one to look at the overall infrastructure, and cross-cutting issues for carrying out the administrative simplification requirements, and then separate ones on claims and encounters. And that is really where I see big content issue, the claims and encounters, and also the next one, health insurance enrollment and eligibility. Health identifiers, which are critical, as I said in the first slide, and medical coding and classification, and then systems security. Each of these, really, other -- except probably with the exception of systems security, are issues and concerns that were addressed by the core data elements project. So the relevance of that project, I think, to the work that then was mandated for the Department and the committee is, I think, quite strong and quite clear. I also find it kind of a nice coincidence that the President signed the Health Insurance Portability and Accountability Act on the same day, August 21st, that the National Committee submitted its recommendations to the Department. So, I see that as good timing.
Okay. As you well know, and as I needed to point out to this audience, this other audience, was that there are a number of advisory and consultation mechanisms in the law and, of course, right up there at the top is the National Committee on Vital and Health Statistics.
The National Committee has been required -- just a little bit of review here -- to make recommendations on the administrative simplification provisions, also with respect to privacy of individually identified data, uniform data standards for the patient medical record information. So, even the Congress recognized this continuum -- and then standards for the electronic exchange of that medical record information.
Now, what I also pointed out, which probably does not need to be pointed out to this group, but I will just so that you will know that I am out there trying to let people know as much about the committee and its history as I can -- was that this was not some kind of exceptional activity for the Congress to say that the National Committee needs to be involved with these decisions because the Committee really has a long history of involvement, actually going back to 1949 when it was established in classification systems, going back to the early days in the late 1960s, 1970s, on uniform core data sets, the populations at special risk. I do not think that you had your new title yet, but I was thinking of you all obviously, and state and community health statistics. Each of these is an area that the Committee has a long history of involvement with, and I think, although a lot of focus has been on some of the immediate issues at-hand, certainly I know that folks in this room agree that these are areas that the committee continues to have responsibility for and that issues like the core data elements and those recommendations are relevant too.
And then I talked to them somewhat about particularly classification issues and some of the others. I am not going to go through that here. But some of you are quite familiar with that history.
In the area of uniform core data sets, this has really been progressive. There was a uniform hospital discharge data set and its early recommendations, actually, in late 1969, 1970; the uniform ambulatory care data set which followed after that. And then each of these were updated by the Committee in 1989 for the ambulatory data set; in 1992 for the UHDDS. So, these have been on the Committee's horizon, agenda, plate, since 1970, really going on 30 years.
Also, the Committee has made recommendations in the past on long-term care data sets, not only on patient-based long-term care data sets, the facility-based long-term care data sets, and on the minimum data set for nursing home residents, which the Committee is very actively involved with. Lynette Araki was actually staffing the committee at that time, so she could answer questions about that better.
So, all of this led up to this project, the Core Health Data Elements Project which, as I said, reached completion in August of 1996 and actually grew out of the health reform enthusiasm. Jim is shaking his head because, obviously, the Department was very helpful in providing funds to the Committee, as it continues to obviously in the current work -- but to look at what would be -- what elements would there be for eligibility and enrollment in core data sets, because this was being discussed during the time of the health reform proposals in 1992 and 1993.
In fact, it was at that time that some of this tension between should we be focusing on core data sets or should we be focusing on electronic medical records came up, and I was kind of trying to put that to bed in this particular presentation which actually I can report to you that there was a tremendous interest in the topic and the committee's work in this in the meeting that I attended.
In addition, throughout all of that work and in some additional work as well, the committee has talked about unique identifiers for individuals, facilities, and providers. So, the HIPAA legislation really picked up on many things that the Committee had been advocating for many years. I think that is actually very exciting.
Because this is your committee, I thought I would just share with you what I said about this Subcommittee and how it relates to all of this, and that is -- I guess I did know your new title by then -- that you were coordinating the two-day hearing in San Francisco, and I assume you have already had some discussion about that or will later -- that you were seeking input from populations at special risk which, in fact, we did get at that meeting and it was excellent, and that you were also considering core data elements for disability, long-term care and mental health, and Lisa alluded to that, and that, also, you are going to comment on the proposed federal standard on racial and ethnic data. All of this relates very much, I believe, to the core data elements, requirements, and related activities.
So, just to have a mid-way summary, getting back to what I was talking about in the beginning with the shared vision. We have got this privacy framework. And this is where I have really tried to put these in a hierarchy of needs as I see them. This is my own vision, and it is not necessarily the committee's, though I think that the topics are all of interest to the committee. I do not know if you have put them in the same hierarchy.
First, we have the privacy framework, which, hopefully, we are going to be moving more towards in the next year.
The classification systems, I think, are critical, not only for procedures and diagnoses, but also for issues like racial and ethnic identification and other types of classifications that you need in order to collect some of the core data elements. I believe that we need agreement on core data elements, their definitions, et cetera.
That leads to a data dictionary, which might include elements that some people -- that everybody is going to want to collect and others that not everybody is going to want to collect. But, if you collect them, let's try to do it this way. So, that was always, I think, the philosophy of the Committee, not to come up with a set of data elements that everyone must collect or mandate.
To some degree, the UHDDS and the UACDS were more in that mode. These are elements that you must collect. In fact, they have had quite a lot of influence. Certainly, the UHDDS has had tremendous influence in standardizing hospital discharge data in this country. But the 34 states have -- I think 39 states have mandates, 34 are actually collecting data. For the most part, they are collecting the elements in the UHDDS, at least as it was promulgated in 1985, and not all collecting some of the newer elements recommended by the committee. But some of them are doing that as well. Some of those state activities inform the national committee in looking at those data sets. So it has really been an iterative process.
From those elements, you get some kind of a core agreement on enrollment and encounter data, and then, of course, you need the vocabulary and terminology to get at the greater level of specificity for the electronic patient record.
Okay. The major findings from the core data elements project. One, the health core information field was solidly-supported standardization. This was very clear in all of the meetings that were held and, I think, continues to be clear today. Although, a Bob Gellman might point to us, that is where the disagreement starts and where the agreement ends because everybody has their own twist on this. But I think that it is actually broader than that.
But the field was seeking a recognized leader and a group who could forge consensus. This was before the HIPAA. This finding came out of the process before the HIPAA legislation and after the demise of health reform. So, there was a lot of receptivity toward the committee providing some of that leadership and the department. And the response to the committee's request was very significant in people attending hearings, letters, et cetera.
We found that a consensus existed for a significant number of data elements, but there was less agreement on data definitions. I know that this is what the infrastructure team is finding now as they are trying to map all of the different transaction standards to a data dictionary in which everybody is even calling something the same.
One of my most sobering experiences was when somebody we had worked with quite a bit in the Committee told me that he had adopted these core data elements and he was using them in his community, and he really was a tremendous supporter, and he was. But then when he actually showed me the data set, several of the elements had the same name, but they were defined totally differently than what the Committee had proposed. So I think that the devil is definitely in the details.
And then there were several elements that the Committee considered critical and wanted to get on the map but admitted that they really required additional evaluation, such as functional status. That was a key one. But they really wanted to at least get people thinking about this.
So, the committee made its recommendations. I am getting to those here. What the committee recommended was not only a set of elements, but a process. We asked the Data Council to circulate the report within the Department and place it on the home page, and the Data Council did do that.
They also asked the council to refer the report to the National Uniform Claim Committee, and they certainly have had the benefit of the report in their deliberations.
The Committee also asked the Data Council to provide stable resources to the project. Well, as Jim can tell us, the Data Council has no resources of its own other than some staff and good will who are actually the staff who also have other full-time jobs.
But I think that the continuing involvement of the committee with the Data Council, certainly Dr. Detmer's attendance and participation in the monthly meetings have provided some of that.
One of the areas that I think there is not really any resources going into is that, on a number of the elements that the committee felt that additional work was needed, but that they were critical. That is something that I will close with. Because I think that that is still kind of on the back-burner, but it is part of the report and part of the report's recommendations, the research agenda.
They also asked, not knowing that HIPAA was going to pass, the Committee asked the Department or the Data Council to pursue within the Department those elements that were ready for standardization. All of you have seen the report. If there is anyone here who has not seen the report, just give Barbara your card. We actually have to do a second printing of it, which I think is good news. We will send you one subsequently.
But, we categorized the elements according to whether they were ready for prime time, whether they were close but not really ready, and then whether they needed a lot of additional work. So, they fit in those three categories.
So, the Committee was asking the Department to try to pursue the implementation of those recommendations where the Committee thought that it does not need anymore work, it just needs to be implemented.
Now, I think that is being, you know, it is part of the agenda on the HIPAA transactions and that I think the Executive Subcommittee agreed at its meeting in San Francisco last month or earlier this month I guess it was, that those elements that were recommended in the core data elements report need to be looked at against how these transactions are being standardized and just see what the relevance is, continuing relevance.
Then we put the data into two different categories. One was personal enrollment data. That included the personal unique identifier, the date of birth, the gender, race and ethnicity, marital status, living or residential arrangement, years of schooling completed, the current or most recent occupation in industry, the patient's relationship to the subscriber, self-reported health status and functional status. Those are actually quite a few elements that are not currently in the standard transactions.
Part of what is happening now, I think, with these is that they are not necessarily elements that you would want to collect on every encounter. In fact, the committee made this point, that they did not need to be collected on every encounter. They should be in some kind of an enrollment or eligibility database.
As a result, they were not really considered by the National Uniform Planning Committee when they looked at the standard for encounter data for a claim, for the outpatient patient claim or the encounter data, equivalent encounter data.
And I do not think that the National Uniform Billing Committee has looked at them either because they are focusing on the claim itself. These are elements that the Committee felt needed to be in the patient's records so that they could be linked whenever you did -- their position was that you are not going to collect race and ethnicity at every single encounter. Probably this is not the best way to collect it, at least that has been the position of the committee in the past and was for this report. But, you need that information in your patient enrollment or eligibility data so that you could link it when you do analyses and look at health conditions or however you want to cut the data. It is the same with occupation and industry. But, here, you are probably going to want to have to update it on an annual basis, or only if it changes. So, just as each time you go into the doctor and say, is any of the information that you gave us changed? If it has, give it to us.
So, some of these were only to be collected once, except under extraordinary circumstances, such as maybe your gender. Address would have to be updated every time it changed, occupation also, and self-reported health status. Clearly, what you report one day might not be the same three years from now.
I think that the idea that Cathy Coltin, a member of the Committee at that time as well, kind of added to this. It was not just maybe what you would collect at enrollment or encounter, but maybe at that first visit. Because often very little information is collected at enrollment or eligibility. It is just enough to show that the person is enrolled in a health plan or is eligible for coverage. But that, at the first visit, or the first visit each year, or however it might work, and depending upon what kind of a plan the person is in that you might want to collect or every five years. All of that was kind of open for discussion, health status.
And then functional status you might also want to collect at every visit if you have a patient in rehabilitation. So, each time, you want to see how much progress are we making. But that is not something that you would collect on every visit for other types of patients. So, there is a lot of room for work here, developmental work. But I think that the concepts the Committee was quite committed to.
Just to mention it. Obviously, the personal unique identifier was considered critical and was considered by all of the participants in the process as the most important core data element, I might add, but probably the most difficult. But certainly, everybody -- and they continue to say that we need to solve this problem, and that it clearly is a critical element for electronic records.
I do a little summary which is in your agenda books for the meeting tomorrow and Wednesday, the kind of recent history of the Committee on the subject of personal unique identifier. The Committee did recommend in the UHDDS and the UACDS the Social Security number as probably the most feasible option, with all of the caveats about that it must be protected, it should be provided in encrypted form, if in a public setting, et cetera recognizing that it was a de facto standard.
But in the core data elements recommendations, other possible alternatives were alluded to, and the need not to just collect a number, but some other confirmatory information about the patients and basic demographics was also included. We heard about that again when we were out in California earlier in the month.
Okay. I am not going to go through all of these elements by any means, but I did want to mention the years of schooling. Because this one, I think, is probably quite important to this particular Subcommittee because it was seen as kind of a proxy for socioeconomic status, not a perfect one, but one that was probably a lot less intrusive than asking people their income, and also could be rather readily collected. There are a lot of examples in health services research at least where it has been easily collected in survey mechanisms. Also, it would probably be something that a health plan or physician would want to know about a person as to what their level of education was so that they could target educational materials appropriately and some other kind of clinical needs to have some idea, if not the exact number of years, but of the educational status.
The recommendation -- I think that I put the wrong one up first, but this was what I was talking about, the proxy for socioeconomic status found not only as readily -- pretty readily able to collect it, but that it is highly-predictive of health status and health care use. And we had several members on the committee who are not with the committee anymore, but who had a number of references in that regard.
And it had been recommended by the National Committee for both the most recent update of the Uniform Ambulatory Care Data Set and the Uniform Hospital Discharge Data Set. Both of those, the 1989 and 1992 versions, did recommend collecting years of education completed.
Returning to this one. It can be collected for a child probably from the mother if the child is under 18, although, probably that is a little high. A considerably younger child could certainly report accurately.
The Committee recommended selecting the highest grade of schooling completed. I just wanted to share that we subsequently got some comments from the National Center for Health Statistics on the recommendations which pointed out that the way the Bureau of the Census is now collecting it, it is in relationship to credentials received. This is also what the NCHS is doing in the Health Interview Survey because it is hard to match years with credentials, and whether or not you have a high school diploma, or you are a high school graduate will impact in lots of aspects of your ability probably to work, et cetera, and your ability to command resources, and whether you have a college degree, et cetera. So that is what they are collecting. I think that you might want to look at that if you were to revisit that item.
I mentioned self-reported health status. Again, it can precipitate demand for health care. It can help determine the prognosis, and it is being collected in the national health interview survey, and that was -- that particular definition was recommended.
And I also mentioned functional status which is strongly related to Medical Care utilization rates, and particularly long-term care, nursing home need. And there the Committee talked quite a bit in the recommendations about the need for both self-report measures and clinical assessments. The recognized that there was no agreed-upon classification system and that there was no recognized instrument for children. So these were problem areas that needed to be addressed or that the committee felt was important.
In the area of encounter data, the Committee recommended selecting what is the type of encounter, the dates, the health care facility, and health care practitioner, very basic stuff that is in all of those encounter data sets.
The type of facility, the place of encounter, the patient's expected sources of payment, and I would say a very minimalist recommendation there. This is one of the biggest issues that people are interested in right now, not your source of payment so much, but what type of structure, what type of care a person is in, whether they are in a managed care, whether they are in a PPO, an HMO, point of service, and some work that the department is doing right now. This is probably one of the biggest gaps that there is. So I think that is an area that needs more work.
Diagnoses, health conditions, procedures, and services. That is basic. The Committee -- of course, that is always been in the data set. This time the Committee did add medications prescribed. This is the first time that the committee had actually recommended that as a standard core data elements.
I have got some stuff here on diagnoses and health conditions, but I think we are running a little long, so I will not show you all of that.
I just did want to mention this patient stated reason for visit or chief complaint, which the Committee did recommend in the ambulatory data set and recommended again. This really seemed to be very critical for primary care and other providers as to what really motivated the patient to seek care. It is very important to studying the process and the outcome of care. Again, there is no agreed upon classification system, although the National Center for Health Statistics does have a reason for visit classification scale that it uses in its ambulatory care surveys, and that could be used. I will only say that obviously the committee continued to recommended collection of the external cause of injury or poisoning, which, those of you who know me know that this has been one of my passions over the last 10 years I guess.
All right. Just to sum up -- challenges and future directions. Legislation I think is obviously going to be necessary for privacy and confidentiality. It probably will be necessary for a unique identifier. I will not talk about that anymore now, but it is sort of unlikely it seems that the standards process, the regulatory process alone will be able to solve that problem.
The standards for transactions, for claims attachments, which leads directly to the electronic medical record and work at the federal, state and local level.
This research agenda, as I said, is pretty much, for the most part, gone unaddressed so far. But this will be part of what I think the Planning and Implementation Team will be looking at in looking at past recommendations and unfinished business of the Committee.
Unique identifier is being addressed. The demographic and socioeconomic factors, race and ethnicity I know you are planning to address today. Occupation and employment, assets, other aspects of that need to be addressed and are not being addressed by anything that is going on right now.
Health status and functional status I mentioned as requiring more work. The work on classifications, some of that is definitely going on now with the HIPAA activities. Vocabularies and crosswalks between terminologies and classifications, sites of encounter and health plan characteristics. It is that last one, health plan characteristics, that I was referring to as being a tremendous data gap right now that there is a great deal of interest in to see how are these different types of care systems impacting on different populations.
So, that was kind of a whirl-wind tour of the core data elements recommendations. Do you have any questions that you would like to have answered? As I said, if anyone does not have a copy of the report and would like one, just let us know.
DR. IEZZONI: Thank you, Marjorie. Perhaps, before we have questions, we could just hear from Ron and Carolyn about these two kind of additional "core data element projects" just briefly. Ron?
DR. MANDERSCHEID: By way of background, in the SAMHSA and previously in ADAMHA, we have been engaged in the development of minimum data sets for mental health and substance abuse for a long time. The most recent iteration of these was in 1989 where we have a system that is based upon encounter information to which you attach characteristics of clients, characteristics of providers, financial characteristics, and organizational characteristics.
The current work is an effort to update that basic system from 1989, where we are looking at enrollment encounter data and how we bring enrollment into this picture, and population based information, and how we also bring outcome information into this.
We were undertaking this project, I think, before the National Committee got into the issue of standards in this particular area. So we were very interested in looking at what would be the minimum core data elements for mental health and substance abuse around enrollment and encounter data. So we went out and we looked at what was actually going on in the field and did a report on that. The initial report is what was passed around as part of this particular distribution.
We have continued that work that we were engaging in. We are almost near the final point of the recommendation now of what we would recommend to our field for enrollment and encounter data. There will be a subsequent report coming out in a couple of months that makes that recommendation.
As part of that, we did additional searching beyond the standard mental health and substance abuse providers to look at what managed care firms are also doing. Our second report will include that.
Now, how does this interface with the National Committee' work? The Subcommittee on Mental Health's Statistics took on the issue of minimum data sets on enrollment and encounter as it related particularly to what the general committee was doing on enrollment and encounter data. We took a lot of testimony on this, both for the issue of children in this area and adults in this area, and there really were two recommendations that came out of this; recommendation one being that any effort to look at the whole question of functional status needs to include social functioning as well as ADLs and IADLs, that you cannot characterize people who have mental illness or substance abuse problems simply by looking at their ADLs or IADLs. I believe there was a report put into the system to that effect during maybe the middle of last summer of 1996.
The second recommendation that came out of the Subcommittee was that we needed to have a high degree of sensitivity on enrollment and encounter data and how that data gets used once it becomes available. We wanted to prevent red-lining of our clients because they were high-risk people. Had we had the Health Security Act, that issue would have been addressed as part of the Health Security Act. In the absence of that, there are no protections on such data.
Therefore, the second recommendation was that we imbed these items that were of particular interest to mental health and substance abuse, including the social functioning, in the medical record, that we do not have a free-standing enrollment and encounter data set, so it at least has the basic protections of the medical record as that exists nationally and by state and so on.
To extend that feature of our work, we have also started a work group on the electronic patient record for behavioral health care. They now have a prototype of what that record should look like, including how we would link the various features of this that are needed.
We need to put together health data, mental health data, substance abuse data, social services data, housing data, and so on. The issue is how do we bring this all together in this electronic record? We do have a report -- an initial report of the prototype that we would be happy to share with that.
DR. IEZZONI: Ron, did a formal recommendation addressing these two items ever go forward?
DR. MANDERSCHEID: Yes. I believe it went forward last summer.
DR. IEZZONI: Okay. To the Secretary?
DR. MANDERSCHEID: Well, to the mega-committee.
DR. IEZZONI: To the full committee?
DR. MANDERSCHEID: The committee did reference it.
DR. GREENBERG: I think they were referenced in the report, the core data elements report. And, in fact, the first one on social functioning was definitely included in the recommendation for functional status. So that was just imbedded into the core data element there. The other one about how the data -- protecting the data, et cetera, it may have been a little more bleakly referenced, nothing other than the core data elements report went forward to the Secretary. But part of the reason for that is that those elements were not included in the core data elements recommendations because it was recognized that there was sensitivity. It was my understanding that that work was going to continue, as you are indicating it is.
DR. MANDERSCHEID: Yes, it is continuing.
DR. GREENBERG: So that it is premature to actually send forward a recommendation because we were not recommending a collection of those elements. But the first one got into the core data elements and the second one was supposed to come with a broader set of recommendations.
DR. IEZZONI: So, to be absolutely blunt about it, from your point of view, it sounds like SAMHSA is doing a lot of work in this area right now -- that this work has been going on independent of the fact that we have not had a Subcommittee from the NCVHS over the last, you know, nine months that it has been looking at this.
DR. MANDERSCHEID: What we would like is that our work have some consistency with what is going on in the health field so that we are not out in left field somewhere, or the health field is not out in left field somewhere in the mental health and substance abuse issues. It really needs to be coordinated work.
DR. IEZZONI: So, how might you envision doing that coordination?
DR. MANDERSCHEID: I think we need to bring back our own materials to the Committee, you know, from time to time, to the full Committee from time to time as we have these materials, and, as the Committee or the Subcommittees are working on these issues, then we need to be knowledgeable of that so that we can have these documents available and they can know what we are doing and the left hand knows what the right hand is doing here basically.
DR. IEZZONI: Jim, would you see that coming back to the full Committee or to the Subcommittee?
DR. SCANLON: Probably the Subcommittee, I would think.
DR. MANDERSCHEID: Initially.
DR. SCANLON: Initially, yes, yes. And, now, the other thing, Ron, you stressed was that whatever developments are underway for claims and encounter data, and enrollment data generally, you would want to mainstream with that to the extent that it could be done so as to not sort of have a --
DR. MANDERSCHEID: I think that we need ultimately to bring the mental health and substance abuse care system together with the primary care system at least in a virtual way.
DR. IEZZONI: I am just trying to be awfully concrete about it -- you know, what we can be doing specifically to be able to address this issue given that we only have one mental health expert on our subcommittee at this point. He is not here today -- you know, and how we can make this happen in an efficient but productive way. I mean, would you see perhaps once or twice a year, three times a year having Ron present what is going on right now within SAMHSA that is most of concern to you in terms of its coordination with the rest of the health care setting, so we could rely on you to bring those issues to us?
DR. AMARO: It sounds like in a couple of months you are going to have the final --
DR. MANDERSCHEID: We hope to finish the recommended enrollment encounter minimum data set. We can bring that as soon as we have that report.
DR. SCANLON: We would like to get that. I mean, I think you would want to get that to the Standards Committee as well here, and then you would also want to get it to the standards groups.
DR. IEZZONI: Maybe what we should do is have it combined with the Standards Work Group.
DR. SCANLON: I think so.
DR. IEZZONI: I think that actually that is the best idea, don't you?
DR. SCANLON: Yes. I mean, this is the way to feed directly into the system.
DR. IEZZONI: That is one thing that Hortensia has been very concerned about is making sure that the population-specific issues percolate back up.
DR. SCANLON: Yes.
DR. IEZZONI: Elizabeth, did you have a comment on it?
DR. WARD: Yes. I think that the idea of combining when we have something that concrete that we could look at, I think that combining the committees and subcommittees is a good idea. I am assuming that the SAMHSA group has what has already been put out in terms of the core data elements. Are there any there that are a problem now?
DR. MANDERSCHEID: No, we do not have any problems that we would say just do not go far enough for our populations.
DR. GREENBERG: I think actually that is what the Core Data Elements Report said, that both the long-term care -- the disability and long-term care statistics and mental health supported all of the elements that were recommended, but felt that they did not go far enough, and are going to be making subsequent recommendations. That was definitely something that was in the report.
DR. WARD: Yes. I remember the report was there. I did not know whether SAMHSA's response to that was that they agreed.
DR. MANDERSCHEID: We had an opportunity to review that in draft form before it was finalized.
DR. SCANLON: On the clinical data side, Ron, and then, again, the full committee, NCVHS, will probably be taking this up later again maybe in the fall right after sort of the first wave of privacy and standards. That, clearly, I would think we would want to have all of the considerations relating to mental health and substance abuse in that context as well.
DR. MANDERSCHEID: Because of the work that has been done then on the patient record.
DR. SCANLON: Right.
DR. GREENBERG: What I wonder is whether the extent to which you are involved in the HIPAA Group Implementation Team in any way might influence that work?
DR. MANDERSCHEID: No.
DR. GREENBERG: No.
DR. IEZZONI: Okay. So, again, being concrete, an actual plan for what we will do is, Ron, as soon as your group has your draft so that your enrollment and encounter data sets, we will try to bring that to -- would this be John Lumpkins' work group?
DR. SCANLON: Uh-huh, yes.
DR. IEZZONI: No, no, no. Our full Subcommittee and Barbara's full Subcommittee, okay. So we will arrange -- somehow staff will arrange for you to present at Barbara's Subcommittee with our Subcommittee a combined thing and be able to make some recommendations. Okay. Yes, Don?
DR. DETMER: I think it would be useful though if there would be a way you could interface with the work groups. Because I think that process is useful. It is useful to us to then respond responsibly to the secretary, because that is the basic process. So, if there is a way that you could interface with that too I think that that would be useful.
DR. SCANLON: We should have mental health and substance abuse represented directly on the enrollment and the counters from this thing. I thought we did, but I guess we do not. I mean, internally, yes. Well, our internal work groups as well. That is what Don is referring to.
DR. GREENBERG: You are not talking about John Lumpkins' work?
DR. SCANLON: No, I am talking about --
DR. GREENBERG: Yes, right. Because I know that they are finally addressing, I think to some degree, the architecture issues. And now the content issues are being addressed. Really, the only people out there who really are doing for much content beyond what is on the curved claims is CDC at the moment.
[Simultaneous discussion off the record.]
DR. DETMER: You and I can follow up on this.
DR. SCANLON: Let's try to get at least one person, Ron, on each of those workers anyway.
DR. IEZZONI: Patricia, you had a comment?
DR. GOLDEN: No, it was a question to Marjorie. I apologize, you may have said this. In the report is there a discussion of the rationale for the selection or discussion of the utility of those -- each item and also a discussion of why it was chosen to be a one-time or an ongoing collection. How detailed is this question?
DR. GREENBERG: There is a discussion. There is a rationale for every element, and then there is more of a general statement about how often it should be collected and the two types of categories. And on some specific ones it is also mentioned. But there is not a lot of detailed discussion about it.
I should mention, on race and ethnicity, that the report just recommended the current OMB director 15 and noted that it was being revised.
DR. CARTER-POKRAS: Actually, it is different from the current OMB directive.
DR. GREENBERG: Okay. It is very close, very close. It is a little different.
DR. IEZZONI: Carolyn, we should probably move on. Ron, is this okay with you, the plan that we have kind of worked out, very specific -- okay.
DR. AMARO: Can I just ask to be sent the two pieces that you mentioned, the recommendations and the prototype? You said that you have something.
DR. MANDERSCHEID: The prototype is what we are going to have in a couple of months. We do -- I probably have the document of the recommendations that went from the Subcommittee last summer to the Committee, and I can share those.
DR. GREENBERG: We have that report. Have we not provided that to the Subcommittee?
DR. MANDERSCHEID: No, no. That was the initial review of the field. There was a report from the Subcommittee itself making recommendations that I thin came in last summer.
PARTICIPANT: We did not send that out.
DR. IEZZONI: Is that on Internet? Is that on the page?
DR. GREENBERG: Well, I think it was done by the consultant.
DR. MANDERSCHEID: We will see if we can get a copy of that.
DR. IEZZONI: All right. Carolyn, we had this Long-term Care and "Disability" Report. Can you give us a little bit of context about that? It looks awfully different from the other two, the report that we got?
DR. RIMES: Well, it certainly --
DR. IEZZONI: Have you all seen this?
DR. RIMES: To give you an impression of what Ron has been describing, what you have before you is a fairly voluminous set of pieces of paper which tie together what are more likely long-term care data sets that are out in the field or being pilot-tested or being looked at now, with the Oasis which we will be discussing in more detail tomorrow, and there is the minimum data set.
This was done -- there is a cross-walk in the back. We had a one-day session of meetings where a series of people came in and presented. This is not as far along as Ron's. This never reached the recommendation stage. The Committee was sort of reorganizing at that point and things were changing. So what you have before you is sort of what I consider just the guts of the descriptive set of circumstances. It maintains a setting-specific issue as opposed to looking at it broadly across a variety of ways of looking at long-term care. And disability is vaguely mentioned.
A lot of the discussion around this that I can recall from the one-day session were similar to some of the issues that Ron raised on what the impact and overall effect of the ADLs, IADLs would have on people in terms of could they be punitively interpreted, plagued with potential discrimination, issues such as that. There was also a lot of discussion about what types of methodological studies should be done to really ensure that data collection specifically addressed these population groups for long-term care for the disabled.
It is still at the stage of -- I have a variety of comments on it, but we have not even really dealt with -- we have not commented back to Vince on. It is addressed at its edge. The Committee has not made recommendations to go forward. It is one of the things that Lisa annunciated that was still a pending task, and we have not even completed paying Vince for this.
I think there are real useful things in the crosswalk, and there are some -- having a collection of these things before us is probably useful. It is, as of a year ago, and things have somewhat changed.
DR. IEZZONI: George, Hortensia do you have any comments on this?
DR. AMARO: I thought what was important about Vincent's report is the idea that hopefully we could get away from the categorical types of measurement. I thought that to me that as the most important impression and what I was applauding when I read it is that we have got to stop creating a measurement for home care that could only measure people while they are in home care. And then we will go over here and measure how that person is doing and that whole idea that people do not move in and out. And that was that they do constantly do constantly move in and out was the part that was most impressive about that.
DR. RIMES: I think it needs to be picked up on a little bit more. I would like to see more of a emphasis on setting with care, which is what you are saying. The crosswalks, I think, are important in helping to look at that. Also, the way certain data are collected vary considerably across these instruments, by definition, by variable type, all of these things that are at the pretty beginning stages.
DR. SCANLON: Was Oasis included in the review?
DR. RIMES: Oh, yes, in combination with Vince's.
DR. SCANLON: Okay.
DR. IEZZONI: George, did you have a comment. Vince's report, in looking through it, made a recommendation on the selection of a functional status measure to be included. I have to assume that that was a personal recommendation of Vince's and not of the Committee's.
DR. RIMES: It is not a -- the Committee has not made any formal recommendation.
DR. VAN AMBURG: Can you comment on what you think of that recommendation?
DR. RIMES: I can comment personally. I have a real issue on some of the functional status measurements because I am not sure that they really are good measures for the long-term care population.
Doug Cane, at the University of Minnesota, and a variety of other people have started exploring that. It is probably one of the areas that really does need to be looked into from methodological standpoint before we make any real concrete recommendations. That is at least my feeling on it. It is touted too much for that population group without really understanding what it might mean.
DR. IEZZONI: Hortensia, do you have any comments?
Yes, Marjorie?
DR. GREENBERG: I just wondered if there was additional work to be done on the reports -- people commenting or things were modified or is it pretty much done or completed?
DR. RIMES: I suspect there will be a committee decision. I have -- [comment off microphone]. We have been somewhat reluctant to just kind of was them forward without having it as part of the discussion, meaning the last chair of this -- of the Disability and Long-term Care. I do not -- [comment off microphone].
DR. IEZZONI: Did you have a comment?
DR. SCANLON: It is another issue that I will just put on the table for the Committee's discussion. There, in the long form of the census, assuming that there will be one, there will be some disability questions, two or three, and those that are available now are -- I think that everybody agrees that they need work. So, probably over the next couple of months, the HHS will be working with Census and other groups to look at those. Then, maybe the time frame is such that the Subcommittee may not have a chance to get involved, but at least a briefing would be helpful.
DR. IEZZONI: Okay. Why don't I give you my comments now that we have heard from people.
I think that we ought to give Vince your feedback, Carolyn, and Judy Casper's feedback. Pat Henry will revise the draft accordingly, complete it and pay the man. Okay. It is not clear to me that the document is such that it would, without a lot of work, feed naturally into something such as Ron described SAMHSA as doing or Marjorie nicely described the full committee as having done with the national core data element project. It is basically a compendium of different information systems that tend to be setting-specific. And, without a huge amount of work, and without it guiding a conceptual paradigm that we do not even have, I think that it is unlikely that we are going to be able to progress that much further.
So, I think, let's put this document to bed, consider it a useful piece of background, but let's do what Jim just suggested.
I also understood that Social Security was about to start -- the Social Security Administration is about to start a big disability survey.
DR. SCANLON: They would like to -- I think they would encourage at least -- they would like to -- they are looking at the disability determination process and the validity of their measures. So, they have asked the National Institute of Medicine to look over their RFP I think it is, the design proposal for this study. I am pretty sure we could have them come in. I mean, they are not theoretically part of HHS, but they are health, and we could have them come in and at least talk about the design of the study. It actually will involve examinations.
DR. IEZZONI: Well, the SSA would, if they are thinking about using a disability determination. You know, basically, what I want to come up with is some concrete piece of something that we could do to not forget disability because obviously it is a critical issue, but I would like it to be setting-independent, rather than tied to long-term care, home health, hospice, et cetera.
Perhaps one of the ways that we can do that is by keeping track of how the field of data gathering is looking at functional status.
Again, going back to what Ron said earlier, that functional status is a multi-dimensional quantity. It is not a single attribute. And so, looking at functional status along the range of dimensions may be bringing in the SSA people and the Census people to present to us about what their -- how they are thinking about going about with these definitions, but also begin to think about the coding issues.
We heard a little bit in San Francisco -- on two different days, we heard two different things. It was very interesting. Mitchell LaPlant is a leading, leading disability researcher at UCSF. He was asked point-blank how he felt about ICIDH, which is the International Classification of Impairments, Disabilities, and Handicaps. He said it is a framework, but it is not in a usable state right now.
The next day we had this wonderful woman who presented. What was her name, Swensen?
PARTICIPANT: Gretchen Swanson.
DR. IEZZONI: Gretchen Swanson, who presented very eloquently the benefits that her clients have found from using the ICIDH. Obviously, in the core data elements that Marjorie described we have slots allotted for diagnostic coding. Should we also -- and e-coding, external causes of injury. Should we think about proposing slots for ICIDH coding or should we try to link that with our functional status measure?
So, maybe that should be -- those two things, you know, kind of monitoring what Census and SSA are doing on disability, on information, and maybe trying to have a hash- out of the benefits of the ICIDH and whether people would ever use it in this country.
Yes, Marjorie?
DR. GREENBERG: Yes. I think that what Mitch said in his written comments was that he suggested that the Committee look at the ICIDH, but he was not recommending everybody start collecting it. Of course, Gretchen is using it for rehabilitation patients. But I do know that NCHS would like to have an opportunity to brief the Subcommittee on the revision process that is going on, the International Revision Process that is going on right now on ICIDH, and its particular North American components of this work being done on mental health, behavioral health, with children, et cetera. I think that they would really welcome -- the people who are working on that would really welcome feedback from the committee on it, but, you know, whatever you would find most useful.
DR. IEZZONI: And both of these two activities on the Census and SSA, thinking about how to quantify disability and the ICIDH, I think, ought to be -- we ought to partner with -- again, I am not entirely clear as to whether it is John Lumpkins' working group or whether it should be Barbara Starfield's Subcommittee. But I think that that should be done also in conjunction with Barbara's Subcommittee.
DR. SCANLON: The other area is much closer to home, the National Health Interview Survey has completed the first and second wave of the National Disability Survey. We are at the point where we have some estimates now. If you remember, that got some population estimates, but also some estimates of service use in the community. So that might be worth seeing where the next direction for analysis might be. DR. IEZZONI: Well, I think the thing that is interesting, from our Committee's point of view about the NHISD is that it is noninstitutionalized civilians.
DR. SCANLON: That is right.
DR. IEZZONI: I was just talking to somebody who knows a lot about this who said that we have a boundary problem now figuring out who is noninstitutionalized. For example, are people who live in retirement communities where they own their own home --
DR. SCANLON: Group housing or something?
DR. IEZZONI: -- group housing? And so are we -- how are we finding these people? Are we not finding these people. Are we including them or are we not? I think that these might be survey issues. Again, is Barbara's committee going to be focusing on the survey issues?
PARTICIPANT: The Subcommittee might be also.
DR. IEZZONI: But the Subcommittee, yes. I mean, again, I think that it should be a partnering.
PARTICIPANT: For a population group.
DR. IEZZONI: Yes, okay. So, we have a couple of very concrete issues on having to do with the disability and functional status measurement. Again, Carolyn, put poor Vince out of his misery. Pay the man, get it done. Okay? Thank him kindly. Let's keep it as a reference document and let's move on.
Okay. We are not being as efficient as we started out with. But, I think, in fact, we have dealt with at least one item of the unfinished business, which was how to deal with the core data elements. Ron, are you and Carolyn satisfied with the resolution that we have kind of come up with? Okay.
OMB Directive 15. We are going to be talking about after the break. George? Your recommendations on state and community statistics. Can you tell us a little bit about what needs to be done to bring that to an appropriate kind of conclusion?
DR. VAN AMBURG: The remaining survivors of the Committee -- now I am the sole remainder --
DR. IEZZONI: Yes, true.
DR. VAN AMBURG: -- put together a short summary of the findings of our hearings over the last couple of years on the need for state and community statistics, and particularly community assessment. We put that into a series of recommendations. It is a couple of pages long. I think that Elizabeth has looked at it. You have looked at it.
DR. IEZZONI: I thought it was good.
DR. VAN AMBURG: And I think that all we need to do is to get the committee to approve that with a letter from the Secretary saying that these issues still need to be considered and thought about.
I think that the timing is very good because of the state surveys that are being considered right now and the changes in other surveys, that would focus and then on the need for community assessment is very critical.
DR. IEZZONI: Okay.
Well, Elizabeth has looked at it, I have looked at it, Hortensia could maybe look at it quickly. We could get Richard to look at it. I think, basically, the Subcommittee has looked at it. So, what we should plan to do is have George report this to the full Committee to get a letter drafted.
When can we do that? We could do it tomorrow, but we do not have time.
DR. DETMER: That sounds great. What we would like to do -- the layout for the meeting is to have, in the first day, a reading, if you will, of what we would like to approve, so that we would come back then the next day for action. So, as long as we have a draft that people could see --
DR. VAN AMBURG: I have a copy. Can the staff bring copies?
DR. GREENBERG: Well, let me think. If you have a copy here, we can make them down here.
DR. IEZZONI: All right. Then the resolution is --
DR. GREENBERG: I did not bring it with me. I think I have it back there. I could make someone a copy.
DR. IEZZONI: Maybe someone could copy it during the break so George does not have his only copy leave his hand. Okay?
And so what the resolution of George will be is that he will present this tomorrow for full Committee approval. And then does it need to be drafted in the form of a forma letter.
DR. DETMER: Before the full Committee discussion tomorrow. It can be approved the next day.
DR. IEZZONI: And does it need to be drafted in the form of a letter or can it just go the way it is?
DR. VAN AMBURG: With the recommendations, you would have a cover letter with those.
DR. DETMER: Or a transmittal memo, whatever you -- I mean, if the Committee agrees on this, we could move on.
DR. IEZZONI: Elizabeth, did you have a comment?
DR. WARD: I just had one question. Do you think there are entities out there that will be likely to help endorse and make those recommendations happen? I think that it is nice to send the Secretary a letter saying that these are our recommendations, but, in fact, states and local entities doing assessment are not usually looking at letters that are given to the Secretary to guide their work. There needs to be something that happens that actually -- organizations, those local -- particularly at the local level, that those groups are looking to to say we are about to do community assessment, what are your recommendations? How do we get it into those hands?
DR. VAN AMBURG: In the past, we have distributed recommendations out with materials to the state centers, letting the directors know at the state centers.
DR. SCANLON: We could send that stuff.
[Simultaneous discussion off the record.]
DR. WARD: I was thinking ASCO and NATO, and those kind of organizations.
DR. SCANLON: Send a letter from the Committee to them.
DR. VAN AMBURG: I think it is particularly important to get it to the state centers because they provide the statistical support to the locals.
DR. DETMER: We could be on the Internet broadcasting for the next two days.
PARTICIPANT: The next two days, right.
DR. IEZZONI: Okay. So, basically, we will submit George's recommendations through the Committee to the Secretary, as well as to the State Data Organization.
DR. GREENBERG: Or to the Data Council.
DR. SCANLON: Yes. We will probably make a presentation to the Data Council.
DR. IEZZONI: All right, George. Is that a reasonable conclusion for you?
DR. VAN AMBURG: That is great. We can get this wrapped up.
DR. IEZZONI: Okay. Let's get this done in the next couple of days and get that done. Because, George, this has been hanging around since last November or whatever.
[Simultaneous discussion off the record.]
DR. IEZZONI: Okay. One final kind of unfinished business at least on my list, and that was the Spanish translation. Hortensia, you saw that there was a letter about that from Ed Sondeck here. Are there any further steps that you feel that we should be taking right now to kind of tie up this loose-end? It is obviously one that is going to be continuing or that we are going to want to continually monitor.
DR. AMARO: I think that this was the last one, if I am correct, the last major need for translation in terms of the major data systems, right?
DR. CARTER-POKRAS: This was a gap that had been identified.
DR. AMARO: Right.
DR. CARTER-POKRAS: There may be some other data systems that we have translation issues with. But this -- it was because of the consolidation of survey effort and the fact that many of the surveys would be spinning off of the data question occurring within HIS, that this was deemed very important because the NHANES has, for the last couple of surveys, translated these questionnaires that they were doing to spin-off. But, now, with the change in the survey design, they are not going to have the same people. So a lot has changed. But it is still good that HIS is going to be doing this because that will improve a lot of the data that comes out of NCHS regarding Hispanic health.
DR. IEZZONI: Because, I think, since Dr. Sondeck wrote Dr. Detmer a letter, that Dr. Detmer should write him back. It is the polite thing to do.
[Laughter.]
DR. IEZZONI: Well, but it is really us, Don, that this letter was written to. You are just the head, you know, the titular head here. So the question is what should we say in our letter back to him based on that letter. George?
DR. VAN AMBURG: Dr. Saunders' letter did not address that David Williams brought up about bilingual interviewers at all, which was rather critical. As we move more towards telephone surveys, this is an even more serious issue.
DR. GOLDEN: Can I make a suggestion? That is -- perhaps if I had known you were going to create an item -- because the NHANES -- the health interview survey people have begun to work on this project, and they are going to be working with the Library of Congress which has a translation service. I think -- I am not sure whether they would use bilingual or not, but I would think that we would pretty much have to be involved with their approach if that is the service we are going to use. You might raise that issue. But it would be perhaps good if you would plan for us to update you on that.
DR. IEZZONI: Right. Before we state that in a letter.
DR. GOLDEN: We will give you a little more detail about what we know.
DR. IEZZONI: I do not know if, Pat, there was this letter -- I do not know if it was at your place. It kind of describes it.
DR. GOLDEN: I saw it before it went out of the center.
DR. IEZZONI: Yes. It kind of describes what I know that Congress --
DR. GOLDEN: We have established a group. They have met with Census, they have met with the Library of Congress, and they really have made progress beyond this. I think that they are quite comfortable with the service that the library of Congress can give them. Maybe you might want to just ask for a continuous update for information.
DR. IEZZONI: Okay. Why don't we ask for that either, Don, through letter or verbal communication?
So, Hortensia, is that satisfactory for you?
DR. AMARO: Yes. I think that the bilingual/bicultural interview was the only thing that was not responded to that was a part of that letter.
DR. IEZZONI: Okay.
DR. CARTER-POKRAS: There was an issue that came out of the discussions that Hortensia Amaro had attended about NHANES that was a new issue that was not known at the time of David Williams' letter, and that was about which of the special populations would be universally sampled in the HIS -- excuse me -- in the NHANES.
DR. AMARO: Yes.
DR. CARTER-POKRAS: Because we saw a presentation of the Data Council also that was shared at the NHANES Meeting where they said Whites, Blacks, and the total population they would be able to provide estimates, but, for other populations, it would depend on the sample size. So it was kind of vague, which is a change from just a few months ago with the NHANES' design that was presented. So that -- I do not know, Pat, if you know. We were sending e-mails back about that.
DR. GOLDEN: One of the reasons I was late was that we had an extended staff meeting at NCH this morning, and I know that whole issue was going to be -- was part of the discussion. I think you have a meeting in the morning with Dr. Sondeck. I think, again, you will get some further discussion on that at that meeting. Some of these may be issues that are timed, but rest assured they remain less of an issue once you send them. And what we probably need to do is to have a formal way of updating you on these.
DR. IEZZONI: Okay. That is what we will do. So that sounds like a good solution to that.
Okay. We have not dealt with the agenda in terms of choosing broad substantive areas that we would like to look at or our nimble kind of environmental scanning procedures. We need to break right now, and then we will have -- yes, George?
DR. VAN AMBURG: I still have one more issue on this bilingual issue. As you move toward slates, and a telephone survey, what is NCHS doing about this? I know with the National Immunization Survey and the way we are doing it now is we are using the AT&T Language Line for translation.
PARTICIPANT: Using what?
DR. VAN AMBURG: The AT&T Language Line for translation. What is the plan for Slates?
DR. GOLDEN: I do not know, but I know Jennifer, who pretty much is involved with Slates, has suggested that as something to look into. I cannot give you the --
DR. SCANLON: We are using the same contractor for Slates as the National Immunization Survey, so we will have to -- if it has been resolved satisfactorily, the National Immunization Survey, we may want to follow up with that for Slates. Yes, we are definitely going to get the situation -- particularly with the over-sampling is being thought of where we are going to have language issues with Slates.
DR. IEZZONI: Let's keep monitoring that. Okay?
All right, great. Let's take a break and reconvene right at 3:00 and we will have our presentations. Okay.
[Recess.]
Agenda Item: Discussion on Federal Data Collection on Race and Ethnicity
DR. IEZZONI: Dr. Tucker, hi. We are going to get started to make sure that we give you all time for your presentations, and give us time for a discussion. Let me welcome you and thank you very much for coming this afternoon to talk with us.
Perhaps, Dr. Tucker, you could introduce your colleagues or tell us how you would like to organize this session or they could introduce themselves. How about introducing yourselves first? Maybe that would be best.
[Introductions were made.]
[Discussion off record.]
DR. IEZZONI: Dr. Tucker, please define all acronyms before you use them.
DR. TUCKER: I am going to cover the current population survey supplement on race and ethnicity. It was done in May of 1995. And then I guess that Manuel -- Claudette will discuss the 1996 National Content Survey. And, Manuel, you will discuss the race and ethnic targeted tests. And Roger will do the relationships to other research.
DR. IEZZONI: Do you all talk among each other?
DR. TUCKER: Yes.
[Discussion off record.]
[Laughter.]
DR. IEZZONI: Okay.
When OMB began in 1993 to review the 1977 directive number 15 on standards for reporting race and ethnicity results, they convened an Inter-agency Committee, which then turned around and convened a research working group to work on looking at investigating the proposed changes to the standards in terms of how it will affect the data quality and the data collection procedures that were in place for the different agencies.
And the first empirical study that was done to evaluate this was the -- the CPS now -- I can say CPS, I guess -- the Current Population Survey Supplement on Race and Ethnic Origin. This was conducted in May of 1995 as a supplement to the CPS.
The content of this survey was based upon what the Research Working Group had identified as essential research issues for this review which were the possibility of multi-racial category on illness of races, combining Hispanic origin of race into a single question, the concepts, the populations of concepts of race, ethnicity and ancestry, possible terminological preferences by the different racial groups, and new classifications or new categories that might be added to both ethnicity or race.
I will tell you a little bit about the May 1995 supplement. First of all, the CPS is the major labor force survey of this country which measures the unemployment every month. It consisted, at the time, of about 60,000 households interviewed each month. We took those 60,000 households and, at the end of the labor force series we asked -- we broke it up into four panels asking different questions to each of the 15,000 households, panels of 15,000 households. We examined the effects of adding a multiracial category to the list of races, including Hispanic in the list of races, and some alternative terminology preferences. The supplement was developed by a working group with the help of substantive experts on race and ethnicity, the questionnaire design experts and cognitive researchers.
Now, you have in front of you a press release which was released in October of 1995, I believe. I am going to go over the basic results that are contained in this press release.
First of all, we divided the survey sample into four different panels in which we tested out different versions of the race and ethnicity questions. One panel, the first panel has a separate race and Hispanic origin questions without a multiracial category.
The second panel had a separate race and Hispanic Origin question, with the race question included in the multiracial category.
Panel three had a combined race and Hispanic origin question without a multiracial category.
And the last panel had a combined race and Hispanic origin question with a multiracial category.
Let me go over exactly the question of wording for these for the panels. In panel one we asked, first, are you one of the following: Hispanic, Latino, of Spanish origin, followed then by which one of the following are you, and then we have this list of races.
Panel two included the same initial question on Hispanic origin, but now we added multiracial to the list of races and then asked those who identify themselves as multiracial to indicate which races they identified.
In panel three we just had one question in which we now included Hispanic origin as well as the other racial categories.
In panel four, it also just has the one question, which includes Hispanic, but also now includes multiracial with, again, a followup question for those who identify themselves as multiracial to say which race has been identified.
Now, one of the virtues of doing a supplement to the CPS is that in the first interview of the CPS survey is a survey in which you are interviewed once a month for the first four months, and then you are out of the survey for eight months, and then you come back in for four more months. That very first interview, which is done by a personal visit, we ask people the race and ethnicity of all members of the household. The current version is one which conforms to the current standard. So, we have a standard of comparison within the CPS survey itself.
These appear in this order on the CPS original survey asking the races, in this case, White, Black, American Indian, Aleut, Eskimo, Asian-Pacific Islander, and then an other category, followed by this origin or descent question where those who fall in the categories 10 to 17 are combined into a single Hispanic origin category.
In addition to looking at the combined race and ethnicity question and the multiracial category, we looked at those people identified as Hispanic to assess what they preferred, whether they preferred Hispanic to be a separate question or included as a race.
We also asked different racial groups to identify what terms or names they prefer to be identified by. We also asked some questions or reasons for identifying as multiracial which I will not go over at this point.
When we actually fielded the survey, we have an initial nonresponse to the labor force portion on 6.5 percent and then, on top of that, we had another 10 and a half percent. So, our overall nonresponse rate to the supplement was about 17 percent. It did not appear to be any different between the four panels, nor did it seem to be associated with the original race and ethnic identification from the original CPS.
We rated them for a nonresponse and for a supplement nonresponse but nor for population controls because the population controls are based upon the 1990 census and the original versions of race and ethnicity which we are trying to now evaluate new data, so we will destroy the comparisons.
the comparisons that we did make from the data include looking across treatment conditions, comparing them rather than to population totals. We had complex samplings which required special calculations, and we used a linear analysis for most of the analysis that you will see here today.
Okay. Let me get to the findings. When we compare the proportion of the population -- the proportion of the samples I should say that identify themselves as Hispanic, we find that, in the first two panels, those that have the separate Hispanic question identified the proportion identifying as Hispanic is higher than in panels three and four where we included Hispanic as a race but the instruction was that they should mark just one of those categories. We will go into that more, but there is definitely a difference here. It is a drop of about 20 to 25 percent between panels one and two versus panels three and four.
Now, in terms of the racial distribution based on these four panels, and this is just based on the first race question, not getting into the different -- we did talk about combinations of the different races, but just looking at the basic distribution, we find that, in panels one and two, we see about 80 percent of the population identifying as White.
When Hispanic is included in panels three and four, the proportion of the sample identifying as white drops by about four or five percent because some of the Hispanics who are identified as White on the race now move to Hispanic.
At the same time, you can see that when we use a multiracial category we do not affect -- it has no statistically significant effects on the proportions identifying as White or as Black. However, for the American Indians, we see a drop of about 20 percent or so of the count of American Indians when multiracial is included.
You can also see that when multiracial is included that the proportion identifying as other also drops. It drops the furthest when you have both an Hispanic and a multiracial category, because there are no Hispanics. They are clearly identifying as something else when there is no Hispanic category.
Now, we then asked those people whether -- however they defined themselves, in terms of which panel -- whoever defined themselves as Hispanic, whichever way they were able to do so, we asked those people what their preference was in terms of having Hispanic as a race. It turns out that, in panels one and two where there was a separate question for Hispanic origin, we find that something over 60 percent still identify that they would like to have Hispanic as a race on the race question. This jumps up to well over 70 percent in those cases where we have just a single question with Hispanic appearing on it.
As I said before, we also asked these different groups what their preferred terms or names that they wanted to be called. This might mean -- the results here might be influenced by the fact that we had to for the initial race and ethnicity questions use some terms. But, in this case, we see that a majority of Hispanics choose Hispanic as the designated term and that Blacks, the plurality of Blacks choose Black, but we see a very significant proportion, over 40 percent, choose either African-American or Afro-American as their term of choice.
American Indians. We see about a little over half choose either specifically American Indian or Alaskan-Native and a little over a third choose the more generic -- I guess about 40 percent choose the more generic name, Native-American or some other term.
Finally, we see that 28 percent of the multiracials actually choose that as their term. About an equal number say that they have no preference as to what the term is.
Now, we turn to looking at comparisons of the initial race and ethnicity questions that they answered in the first CPS interview, compared to what they did in these panels. You can see the, in the case of Hispanic Origin, you can see that there is an inconsistency particularly in -- not so much in panels one and two, although there is some inconsistency there in terms of identifying themselves as Hispanic, but we see a real inconsistency when you look down at Panels three and four where Hispanic now appears as a race, and you see that those people in the original survey who were identifying themselves as Hispanic, now you see that not as many of them in the supplement in panel three and four were identifying themselves as Hispanic because of the conflict that is raised when they are having to choose one category and choose between Hispanic and some of the other racial categories. We will talk more about that in a minute.
In the case of race, and comparing the first time they answer this question to the second time, we see that Whites and Blacks, the consistency between first and second time are over 90 percent. We also see that the -- all other -- the other category, which is the something else category, which has a large number of Hispanics using that in the original question in the CPS, the other category. We see that 65 percent and 55 percent chose that in these cases. It is still a dropping consistency, but not nearly as much as down here where they all tend to drop to Hispanics.
You can see that the other category was capturing a lot of Hispanics in the original CPS question. When they were given the opportunity they moved to Hispanic. Some also moved to multiracial.
Finally, you can see that for American Indians we have the least consistency of the four races. They tend to move around more than the other ones. The consistency is less, which historically has been the case.
Now, in terms of the conflict in Hispanics we are seeing when they are forced to choose Hispanics over one of the other races, we see this mirrored in looking at the response to these questions by the various Hispanic subgroups. You will note that almost all of the Mexican-Americans and three-quarters of the Puerto Ricans and Central or South Americans choose Hispanic when Hispanic is offered them in the race question. However, Puerto Ricans, who almost always -- I am sorry, Cubans, who almost always choose White when Hispanic is not given, now are still choosing White 50 percent of the time. So, Cubans do not move as readily to the Hispanic category as do some of the other Hispanic subgroups.
You can see what that does in terms of the Hispanic population is that, if you were to look at the Hispanic population as defined by these four panels, you would see that, in panels three and four, a larger proportion of them are Mexican-American, in particular, and a proportion then of Cuban drops quite a bit in those last ones. So the socioeconomic status of the Hispanic population could be affected depending upon what -- how you form the question -- which Hispanics tend to move to the other Hispanic category, which they are supposed to do.
Okay. Let me sum up here what we found in the CPS. The size of the Hispanic population will be larger if a separate question is used as opposed to a single question in which we only allow one -- selection of one category.
The size of the White population will be smaller if Hispanic is included as a racial category. It is likely that American Indian, and the Alaskan Native population will also be reduced if a multiracial category is included. Hispanics prefer to be in a racial category, although this was in the absence of knowing how it would affect the counts.
There is no discernable difference by panel in preferences for terminology, so that all of the panels were collapsed together when we looked at preferences because there really weren't many differences.
Mexican-Americans would be a larger proportion of the Hispanic population if race and ethnicity would be combined into a single question with only one response allowed.
The change in Hispanic identification when it is a race holds in comparison with the CPS control. So, you see that when there is a -- we saw that in panels three and four when Hispanic was included as a race that accounts of Hispanics dropped. We saw that mirrored in the inconsistency in comparing the original CPS interview results to the CPS supplement where you can see those people who originally chose to be -- were identified as Hispanic now. Some of them are now choosing not to be identified as Hispanic. We see a large portion of the other race category will move to Hispanic when that category is available on the race question.
American identification is less stable than other races when we compare the supplement to the CPS control card.
Again, Cubans behave differently on the race questions than other Hispanics when Hispanic is included.
The degree of assimilation is how long people have been in this country. It could interact with the treatment effects. The people who have been in this country for the least amount of time may be more sensitive to the different treatment effects.
DR. IEZZONI: Why don't we pause briefly for questions for clarification only? We will have our discussion after all of the presenters have talked. Does anybody have any -- are there any questions of clarification at this point?
[No response.]
DR. IEZZONI: Okay. Great.
DR. AMARO: I just had one question. Sorry.
DR. IEZZONI: Yes.
DR. AMARO: The option of American Indian -- you talked -- you presented so much data, I cannot remember which one it was -- but back when you discussed that some Hispanics go on to identify as American, indian and one of the options -- I cannot -- there seems to be an increase in this?
DR. TUCKER: No. American Indians -- people who identify themselves as American Indians are the most inconsistent subpopulation. So, in other words, you may go one day and get the man to identify as an American Indian, and the next day they may identify as something else. Whereas, most blacks and whites, in particular, they tend to remain White and Black over the course of several administrations of the survey. We found this inconsistency is probably magnified by the different treatment effects. But we also know from other questionnaires, even when you go with the same question both times, the American Indians are the group that is most likely to be inconsistent in terms of their identification as the original identification.
DR. AMARO: I will find it and ask you later. My question was along the lines of, for people who are of Hispanic background but of Latin-American indigenous background, you know,k what option they had in identifying?
DR. TUCKER: Okay. Those Hispanics from Central and South America, when provided an opportunity to check Hispanic in the racial question were much more likely to do so than Cubans are. That is what we were asking. Cubans are more likely to identify as white.
DR. IEZZONI: George?
DR. VAN AMBURG: What was your non-response testing methodology? How did you do it?
DR. TUCKER: It was a rating adjustment based upon our traditional nonresponse rating adjustment in the regular CPS that is defined by geographic cells.
DR. IEZZONI: Okay.
DR. BENNETT: What I am going to ask you to do was to take a look at this handout, finding on questions on race and Hispanic Origins tested in the 1996 National Content Survey. If you follow along with me, it will make it very easy in terms of understanding what is going on.
On the second page of that outline, it basically tells you what it is that I am going to do. I have background information on why this test was conducted, the objectives of the study, the major findings from the National Content Survey, and the detailed findings.
On page three, we talked about the fact that the 1996 National Content Survey and the 1996 Race and Ethnic Targeted Test are part of the Bureau of the Census 2000 research and testing program. Whether the Office of the Management and Budget was going to review staff directive 15 or not, the Bureau of the Census had an obligation that we needed to look at the race and ethnicity questions that were used on the 1990 census to find out whether or not they would be appropriate for the 2000 census.
This is something that we do after every decennial census. This is not new. It was new in the sense that we had some outside motivations to get it completed.
The 1996 National Contents Survey was a national sample of over 90,000 households. It represented about 95 percent of the country. It was less representative of the American-Indian and Alaskan-Native populations. One of the reasons as to why that occurred is because it occurred in mail-out/mail-back areas. American-Indians and Alaskan-Natives do not have telephones as often as other sectors of the population. So, when we are talking about doing a re-interview, we did not have a pool from which to draw. So, both of those things influenced the sample design to a large extent.
The National Content Survey was a survey to test the content for the 2000 census. It not only focused on the race and ethnic questions but other questions that were likely to appear on the 2000 census. So it is different from the race and ethnic targeted test that Manuel will describe because, in the race and ethnic targeted test it was designed specifically to test racial and ethnic populations. I will not say anything else in terms of the race and ethnic targeted test which is some of the targeted populations are described on page five, but Manuel will give you more information on that.
If I am talking too fast, I would say let me know. I have a tendency to talk very fast. Okay.
There were basically three test objectives for the National Contests Survey. We wanted to determine the effect of reporting in the race and Hispanic origin question if a multiracial category was added to the race question. We wanted to find out what would happen if we placed a Hispanic origin question immediately before the race question, and we also wanted to find the effect of the combined changes.
And what you have on pages seven through 11 are the actual questionnaires that we used in the 1996 National Contents Survey to test the effect of race and Hispanic origin reporting.
On page seven you see what we call our control panel, for the most part. On the control panel, it was a 1990 modified race question. It was modified in the sense that we made some changes in terms of instructions, and we made some changes having to do with terminology, things that we thought would improve reporting by persons using a self-identification context. Because, as I said earlier, this is a mail-out/mail-back survey. The re-interview was where we used the computer-assisted telephone reinterview after we administered the survey. In the months of May and June we conducted the reinterview with respondents. And our goal in the re-interview was to try to find out whether or not we were going to get consistent reporting of race and Hispanic origin.
So, in panel one, what you have is that you have the modified 1990 race question that does not include a separate multiracial category.
In panel two, you have the modified 1990 race question with a separate category that is entitled multiracial or biracial. We use the term multiracial or biracial as a result of cognitive research that was conducted. When we did the cognitive research and also for some of the results coming out of the May 1995 BLS supplement, told us that maybe what we needed to do was to include both the term multiracial and the term biracial so that persons would in fact understand what it was that we were trying to capture.
It was not sufficient for us to just do the category by itself in terms of the standalone category, but we also wanted to find out if respondents would provide the component pieces of their multiracial identity. But that is why you have on panel two where persons were instructed to write in the component pieces of their multiracial identity. On panel three, what we have is what we -- we switched the race question, the order of the race and Hispanic origin question. One of the reasons we did that was that in some of the research that was conducted since the 1990 census and some of the research prior to the 1990 census suggested that, if the Hispanic original question was placed before the race question that it would result in lower nonresponse to the Hispanic origin question as opposed to the race question. So, we wanted to see if, in fact, that was going to be confirmed.
So, in panels three and four, what we have is basically just an order change with the Hispanic origin question preceding the race question. And, on panel four, we have the addition of the multiracial category for persons to, in fact, write in their multiracial pieces.
What did we find out? Well, not to anybody's surprise, I hope, we found pretty much that about one percent of persons did in fact report as multiracial and that Hispanics were more likely than the total population to identify as multiracial. That was observed in panels one and two. The multiracial category had no statistically-significant effect on the Black population, the White population, the American Indian, Eskimo, and Aleut population, and the Asian or Pacific Islander population. However, let me qualify. With respect to the American Indian and Alaskan-Native population, we really did not get that many American-Indians and Alaskan-Native in the sample because, as I said earlier, it was not a targeted sample. It was a national sample. It was a test of the 2,000 questions to a large extent.
With respect to the Asian-Pacific Islander population, we did not detect a statistical difference at the 90 percent confidence level that the addition of the multiracial category effected reporting. But, there was some evidence when we looked at the write-ins that were provided, we found some indication that there was a drop, not statistically different, but still substantively different in terms of the proportion of persons in panels one and two who reported as Asian and Pacific Islanders. The drop was from four percent in panel one to 2.7 percent in panel two, and from 3.4 percent in panel three to 2.8 percent in panel four. So, there was something going on with the Asian and Pacific Islander population. It may very well be that we did not have our sufficient sample in order for it to test at the 90 percent confidence level, but it was something to suggest to us that something was going on in the Asian-Pacific Islander population when the multi-racial category was, in fact, added.
With respect to the sequencing, where we have Hispanic origin questions sequenced immediately before the race question -- and I say immediately because, if you look at panel one, you find pretty much that, in 1990, there were two intervening questions between the race and Hispanic origin question. The race question was question number four in 1990, and the Hispanic origin question was question number seven. So, we wanted to find out whether or not, if those questions were sequenced immediately behind one another if there would be an effect. What we found, to a large extent, was that the alternative sequencing of the Hispanic origin question first, with or without a multiracial category in the race question reduced the non-response to the Hispanic origin question, not that it eliminated it, but it did, in fact, reduce.
We also found that, without a multiracial category in the race question, there was reduction in the proportion of persons who reported in the other race category. That reduction was from about 3.3 percent in panel one to about 1.7 percent in panel two. So, what it seemed to suggest to us is that when the Hispanic origin question is placed immediately before the race questions, persons of hispanic origin will, in fact, report themselves as being hispanic and not necessarily report in the other race category on the race question.
We also found that it increased reporting of Hispanic in the White category. I think that this was something that was observed in the BLS supplements as well. With respect to the detailed findings, and I am on page 13, if you are following along, we found that the effect of the combined treatment, that is the multiracial category and Hispanic origin for sequence, that the amount of persons who identified as Hispanic on the re-interview, about six-tenths of a percent reported as Black on panel four where a multiracial category was available, and that compared to 2.7 percent on panel three where it is not.
So what you end up finding is that when you combine the effect of the multiracial category and the race question, the sequencing of the Hispanic origin question before the race question, there was a drop in the percent of Hispanics who reported as Black. Today, we still do not know why that occurs. It is an anomaly. We have done some multivariate analysis and still have not been able to figure out exactly what is driving that.
We also found that there was a drop in percent of Hispanics who reported as Mexican when there is a separate multiracial category and the Hispanic origin question is sequenced before the race question.
When the Hispanic origin question is sequenced first, we did not have a reduction in the high nonresponse rate to the race question by Hispanics. Hispanics still did not report in the race question.
If you will remember, in the 1990 census, there were about 10 million persons who reported in the other race category. Ninety-five percent of those persons wrote Hispanic origin entries. So what we had in 1990 was that we had Hispanics, what we think, in some way, trying to tell us that they wanted to identify in the race question or that they thought that Hispanic should, in fact, be a race as opposed to ethnicity.
Let me just go back a little bit and give you some detailed findings as it relates to the multiracial. We found that 98 percent of all persons who checked the multiracial category provided a write-in. All of those write-ins -- 55 percent involved two or more races, that is either entries such as White and Black or entries such as White and Asian-Pacific Islander or White or American-Indian.
Thirty-one percent involved an Hispanic Origin write-in and a race write-in. So, you might have Cuban, White, or Black, Puerto Rican, those kinds, or Central American, Dominican, American-Indian, or whatever combinations. But it involved at least a Hispanic origin write-in and a race.
Twelve percent reported only one of the directive 15 race and the directive 15 race are, as Clyde as already said, White, Black, American-Indian, or Alaskan Native, or Asian, or Pacific Islander.
This was not, I guess, new to us. I guess it was confirming to us to a large extent, in that we have done a 1990 study that looked at multiracial reporting and found to a large extent that multiracial tended to be reflective of persons involved in Asian-White unions, American-Indian-White unions, and Black-White unions. But to, I guess, some folks, surprising -- was not that that the union that most persons probably would have thought was common was Black-White, and it is not. If you really look at multiracial unions, the vast majority of the multiracial unions are Asian-White, which is 45 percent of those are Asian-White, 34 percent are American Indian-White, and about 20 percent are Black-White unions when you are looking at multiracial populations.
That is basically what we saw confirmed in the National Contents Survey. On page 15 is that with respect to the write-ins for the multiracial category. Thirty percent involved an Asian or Pacific Islander response; 25 percent involved a Black response; and seven percent involved an American Indian response. And the lower response, in terms of the seven percent for American Indian is somewhat consistent in that we did not have a large American-Indian sample referenced in our sample.
I think that pretty much, if we had to give you some summary kinds of comments, we would say that the multiracial category, in and of itself, are for the -- did not affect reporting for the White populations, the Black populations, the American-Indian or the Alaskan-Native populations. But there was some evidence to suggest that it did affect the reporting in the Asian and Pacific Islander population.
The combination of placing the Hispanic origin question immediately before the race question, led to some reduction in reporting in the other race categories, but it did not eliminate the non-response to the race question.
So that is a quick and dirty summary of what we found in the National Counting Survey.
DR. IEZZONI: Thank you. Are there any questions?
[No response.]
DR. IEZZONI: Okay. Who is next in the chronology?
DR. BENNETT: Manuel.
DR. DE LA PUENTE: I am going to talk about the race and ethnic targeted test, which is the third series of tests that we are talking about this afternoon.
Refer to the other handout, Assessing the Results of the 1996 Race and Ethnic Targeted Tests, which we call RETT. Please look at that and go to page one. Just to reiterate what Clyde stated, the apparent OMB directive has the racial categories listed on page one and ethnicity listed as Hispanic origin and not of Hispanic origin. These categories are used for data collections for all federal agencies and are also used for administrative reporting by federal agencies.
There are three studies that we talked about this afternoon. The first was the May supplement that Clyde spoke about. The second one was the 1996 National Conference Survey that Claudette just finished summarizing.
What I am going to talk to you about now is the 1996 Race and Ethnic Target Test. If you turn to page two, we can talk about the sample design. Now, the sample design of the Race and Ethnic Targeted Test was quite different from the sample design for the current population survey and for the National Content Survey. The first two surveys are national samples that are represented of the national population. The Race and Ethnic Targeted Test had a specific sample that we selected that focused on the targeted populations that are listed on page two. In order for us to get enough numbers of these population groups, we had to actually look at the census tract level data, just the 1990 census data, identify census tracts where there was a high proportion of these populations, and then take samples from those census tracts. So, we had a total of 112,000 households in the race and ethnic targeted tests, but, nonetheless, it is not a national sample. It is a sample of targeted areas in the United States, so you have to keep that in mind.
The advantages of selecting such a sample, like I mentioned earlier, was that we are able to get a sufficient number of African-Americans, of American-Indians, of Alaskan-Natives, of Asian-Pacific Islanders, of Hispanics, and of White Ethnics in order to do the analysis and comparison that I am going to describe.
The disadvantages of doing this type of sample is that we do not have a national sample, and also the results that we found in the Race and Ethnic Targeted Test cannot be generalized to individuals residing in areas where there is not a high concentration of these population groups. And conclusions about specific effects for specific population groups, again, are based and generalized for that particular targeted group. So, we have the six targeted samples that we just described earlier.
Now, when I discussed the findings, we are doing to discuss findings related to the specific targeted samples. So, when I talk about the Hispanic population, I am going to talk about the Hispanic targeted sample. When I talk about the Alaskan-Native population, we are talking about the Alaskan-Native targeted sample. So, it gets a little bit complicated because of the structure of the sample.
Now, if you turn to page three, we have in the red, there are basically four major purposes that we had in conducting the RETT. The first purpose was to determine the effect of allowing the respondents to report more than one race. And one way respondents can do that is to check a multiracial or biracial category. We saw that happen in the current population survey. We saw that happen in the NCS, the National Confidence Survey, and, again, we have tested it here in the Race and Ethnic Targeted Test, specifically because the sample was focused. So we wanted to know that that category would do on a focused or targeted sample.
Now, there are two other ways that individuals can represent the multiracial background, well, actually, one other way, which is to specify however many races they identify with. But, in order for us to collect that information -- and, by the way, the Race and Ethnic Targeted Test is a mail-out/mail-back survey. So, we provided instructions to the respondent in the race question. In some of the panels we told the respondent to mark one or more of the races that they most identify with as an instruction. On the other two panels, we put in mark all that apply -- to mark all of the races that apply. So, there are two sets of different instructions to respondents telling them that it is okay for them to mark more than one race. So, we tested the differences between those two sets of instructions.
The second purpose of the RETT was to determine the effect of collecting information about race, Hispanic origin, and ancestry in a combined question. The question to be used in the RETT are very similar to the questions that Claudette showed you earlier in the National Confidence Survey. But there were two panels where we actually had a combined race and Hispanic origin and ancestry question. So that was quite different than was done in the current population survey supplement because in the current population survey supplement, they had Hispanic origin as a race. Here we had Hispanic origin as a race, but we also had an ancestry question that is a follow-up to capture the detailed Asian and Pacific Islander subgroup, the detailed Hispanic origin subgroups.
The third major purpose of the RETT was to determine the effects of placing the Hispanic origin question immediately before the race question. Claudette went into great length about why we did that in order to improve reporting on both Hispanic origin and race. Again, we wanted to do it here in this sample because it is a different sample than the NCS and the CPS, and we wanted to test this again in this particular sample.
And then the fourth and final purpose or major purpose of the RETT was to test some alternative terminology that I will go into in a minute.
Now, if you turn to page four, we will summarize here for you the percent or the range of percent of people who identify with more than one race. And now, before we go over the percentages, actually, reporting on more than one race includes people who check the multiracial category or persons who responded to the questions mark one or more and mark more than one races, or the persons who had their questions mark all that apply and mark all of the races that have applied, and also the persons who checked more than one racial category when they were instructed not to, they were instructed to only check one category -- the proportion of persons who saw that questions and did not follow the instructions and checked more than one. So, those are the people who we are going to talk about, the people who have identified with more than one race. That is the universe.
Now, when we look at the flat targeted sample on page four, we see that between one and three percent of respondents in that particular targeted sample identify as multiracial. The American-Indian ranged between two and seven percent. The Alaskan-Native and Asian-Pacific Islander were actually the highest -- the population that identified as multiracial the most often. The Alaskan Native was between five and 13 percent, and the Asian-Pacific Islander was between four and 12 percent.
Now, the Hispanic targeted sample, again, that is a separate sample. When we had a separate race and Hispanic origin question, we found that between 25 percent of respondents in that particular targeted sample said that they were multiracial.
Now, however, when we had a combined question, we had a jump to eight and to 19 percent of persons in that targeted sample who said they were multiracial. And it is interesting because the 18 percent came from panel E in the sample, in the survey, but instructed the respondents to only mark one or select one, and 18 percent did not see or did not follow that instruction.
And then panel F, which was the 19 percent comes from panel F, that is where we gave people a choice, I mean, the option of marking more than one. So, there, we have basically the same response, whether you instructed them to do it or not.
And the White-Ethnic sample, targeted sample, we have between one and two percent identified as multiracial.
Now, also on page four, as we just finished saying, the unrequested multiracial responses, that is to say the presence of the respondents who checked more than one when they were instructed not to, that proportion of responses occurred more often in the Alaskan-Native targeted sample and the Asian-Pacific Islander targeted sample, and the Hispanic targeted sample.
Now, I would like you to turn to page five. In order for us to make sense and make these data meaningful, specifically the data where persons checked more than one when they were instructed to check more than one, we had to come up with a couple of approaches of tabulating the data. And we came up with these approaches merely as an illustration to the reader of how these responses can be tabulated and presented. So, we used these approaches as illustrative approaches.
So, we have three different approaches that we used. The first approach was to take a single race, that is to say, that we only recorded the responses for only one of the four racial groups in the record number 15. And then responses -- persons who said more than one race or said multiracial were put into a multiple race category. So, that was one way of relating the data that we collected. Another way of relating the information that we collected was to do an historical series approach. Basically, the historical series approach -- responses that were provided for more than one race were tabulated in one of the OMB race categories. The is one and only one of these responses and the following works fell in the following categories. They were either Black, American-Indian, Alaskan-Native, Asian-Pacific Islander. So, for example, if we had a respondent who checked Black and White in the check more than one question, we classified that respondent as Black. If a respondent said that they were American Indian and White, we classified that person as an American Indian.
So, this is what we called the historical series approach. This approach reflects the use of the data that we have for monitoring civil rights legislation.
Now, if a respondent said that they were Black and American-Indian, for example, we would collect them with some other race. If a respondent provided us with three races or more than three races, we would code them as some other race. So, it is a way of allocating the information and coding the information. That is what we call the historical series approach.
And then the last one was the all-inclusive approach. That approach simply says that we have coded the persons whatever race they gave us. If the person said that they were American Indian and Black, they were coded as having -- they were counted twice, once for American Indian and once as Black. So, we had the total added one to 100 percent. Because one respondent could have more than one race.
So these are the basic three approaches that we used to illustrate the way that we can tabulate the information. In the report we go in more detail into that approach, into those approaches.
On page six, if you turn to page six, we can talk about the major findings. First of all, I would like to talk about the findings regarding the multiracial category or the effects that he multiracial category had on reporting.
In the Asian and Pacific-Islander targeted sample, the multiracial category, we do the reporting solely as an Asian and Pacific-Islander. Now, however, when we use the historical series approach that I described earlier, we eliminated this effect. So, when we tabulated the data according to the historical series approach, they reduced reporting -- the Asian and Pacific-Islander was eliminated.
In the Alaskan-Native targeted sample, the multiracial category, we do see reporting solely as American-Indian and Alaskan-Native. However, when we are using the historical series approach to tabulate the data, the effect was eliminated by tabulating the data in the historical series approach.
And the multiracial category had no effect or no statistically-significant effect on the Black, American Indian, White Ethnic, and hispanic targeted samples.
Now, I would like to talk about the findings regarding the mark one or more instruction. That is when we asked respondents to mark one or more races if they so desire to do that. We found that the mark one or more instruction on the race question had no effect whatsoever on any of the targeted samples, no statistically-significant effect.
However, the mark all that apply instruction that we had did have an effect on the Asian and Pacific Islander targeted sample. Specifically, the mark all that apply instruction, we do the reporting solely as Asian and Pacific-Islander. But, again, when we use the historical series approach to tabulate the data, this effect was eliminated by using that approach.
And we found that the mark all that apply instruction had no statistically-significant effect on the Black, American-Indian, White-Ethnic, or Hispanic Targeted sample.
Now, if we turn to page seven, I will talk about the findings of the combined question on race and Hispanic origin. Now, the questions that Claudette reviewed earlier on the National Confidence Survey are very similar to the questions that we had on the Race and Ethnic Targeted Test, but the combined portion was quite different. So, for that reason, I would like you to turn to page 15 of the handout so that you can see what the combined question looks like. The combined question on page 15 is question 5A and 5B. In that question, we ask respondents to mark one box for the racial origin that they identify with. We have the OMB racial category along with Hispanic origin as a category and multiracial or biracial as a category, and then we had a followup question, which is 5B, and we asked what is this person's ancestry or ethnic origin? So, we expected or hoped that the respondents who, for example, said that they were Hispanic would put in their Hispanic origin subgroup in part B, will be Cuban, Mexican, Puerto Rican, Columbian, or whatever. The same thing is with Asian and Pacific Islanders. We would expect them to put their subgroup identification in 5B, whether it be Chinese, Korean, or whatever. And, of course, the ancestry of the other racial groups will also be captured in 5B.
Now, we can turn back to page seven. Now, let me see the question. We can talk about the findings. We found that when we had a combined question, the nonresponse rate to the combined question was much lower than the nonresponse rate to the race question, to the separate race question. For example, when we had a separate race question between 13 to 14 percent of Hispanics would leave the race question blank. However, when we had a combined question, only about 1.5 percent of Hispanics or persons in the Hispanic targeted sample I should say left that question blank. So, we reduced the item response to the question by doing that.
Now, in the Hispanic targeted sample, we had a very high-percentage of multiple responses. As I mentioned earlier, it was between 18 and 19 percent of respondents in the Hispanic targeted sample checked more than one race. However, when we had the all-inclusive approach way of tabulating the information, there was no effect on the total reporting of Hispanics. That is to say that when we took into account more than one of the responses to the combined question, we eliminated the difference between the Hispanic proportion obtained in the separate Hispanic origin question with the Hispanic that we obtained in the combined question. So we eliminated that effect as we used the all-inclusive approach.
Now, with the Asian and Pacific-Islander in the combined question, both the combined question reduced reporting solely as Asian and Pacific-Islander in the combined question. But, again, the all-inclusive approach we eliminated that effect.
Now, the next part of this is what we found in the ancestry write-ins in 5B. In the Asian and Pacific-Islander targeted sample, the combined question with the multiracial category resulted in lower reporting of several API categories than in a separate question. That is to say that the information that we obtained in the combined question for the Asian and Pacific Islander was not comparable to the information that we obtained in the race question alone. That is the question that had the multiracial category.
However, on the combined question that had the instructions mark one or more, resulted in the reporting of all detailed API categories, except for the other API. So, again, there as a high reporting of other Asian and Pacific Islander in the combined question than in the separate race question.
By the way, there was also a higher reporting of other Hispanic in the combined question as opposed to the separate Hispanic origin question.
Now, in the Hispanic targeted sample, we found that we could not recapture the subgroup information that we obtained in the separate Hispanic origin question. In the separate Hispanic origin question, we had the Hispanic subgroups, Mexican, Puerto Rican, Cuban, and so on. We found that in the combined question, we could not recreate that distribution because, number one, about 10 percent of persons in the Hispanic targeted sample, left that part 5B blank, so we do not know what the subgroup was. They may have checked Hispanic, but we do not know what Hispanic subgroup they were. So, 10 percent left that blank. And then when we tried to code that information and do the all-inclusive approach, we could not achieve the same distribution that we did with the separate Hispanic origin question. So the Hispanic origin data collected in the combined question is not comparable to the Hispanic origin data that we collected in the separate question. The separate question gave us better data than the combined question in terms of the Hispanic.
If you would turn to the next page, which is page eight, we have the sequencing of the findings of the race and Hispanic origin question, one placed before the other. Basically, we found the same type of thing that Claudette reported in the National Content Survey, that placing the Hispanic origin question first reduced the nonresponse to the Hispanic origin question mainly from non-Hispanics. Also, placing the Hispanic origin question before the race question reduced reporting in the other race category in the race question mainly from Hispanics who checked the other race category.
And we found that placing the Hispanic origin question before the race question did not eliminate or significantly reduce a nonresponse rate that would traditionally or historically have had to the race question among Hispanics.
Now, the findings on the terminology very briefly. We tested the spelling out of American-Indian, instead of using the abbreviation Indian (Amer.), A-m-e-r. And we found that the use of American-Indian, really, there was not significant statistical effect when we used that term, as opposed to using Indian-American. So, there was not statistically-significant effect in the American-Indian targeted sample nor in the other targeted samples that we spoke about.
When we used Native-Hawaiian versus Hawaiian, in combination with changing of the placement of the category, we found that there was a higher reporting of Native Hawaiian than of Hawaiian in the Asian and Pacific-Islander targeted sample. I think that the percentages were six percent compared to seven percent. So it was a very, very small increase in the reporting of Native Hawaiian. And the Guatemalan, or Chomorro, versus the Guatemalan category had no statistically-significant effect in the Asian and Pacific-Islander targeted sample.
Basically, that is all I would like to go over. I would like to leave the rest of the time for questions. Roderick has a couple of things that he wants to say about the things that we have been talking about this afternoon.
DR. IEZZONI: Okay. Thank you. Are there any clarifying questions?
[No response.]
DR. IEZZONI: No. Okay. Let's move on.
DR. HARRISON: I will just try to briefly summarize some of the highlights that we think emerged from these three studies, and some of the key findings. The first is that across the three studies, particularly the CPS and NCS where we have national samples, one would estimate that less than two percent of the population will exercise options to report more than one race, create a category or checking more than one. We have seen much higher rates in the RETT due to the targeted nature of the samples.
The second is that it is consistent in all three studies that even with the targeted samples that the options for reporting more than one race do not have effects on estimates of the White or of the Black populations. It seems that it would not -- reasonable that it would not have statistically-significant effects on these populations.
On the other hand, the surveys indicate that the Asian-Pacific Islander population and the American-Indian population are the ones within which, at least some segments are most likely to be affected.
Clyde cited the effects of multiracial category on reporting as an American-Indian, and Alaskan-Native in the CPS. Some of these results need to be contextualized within the inconsistent reporting of American Indian from the CPS. The RETT -- the NCS had too few American Indians in the sample to -- probably to detect any effects. The RETT did not find any effects in the American-Indian targeted sample. it did find effects in the Alaskan-Native targeted sample where reporting as American-Indian, Alaskan-Native dropped, and where efforts to edit the data using the historical series still led to fewer, a lower percentage reporting as American-Indians.
I might note here that the CPS sample is more likely to draw American Indians who live in urban areas who might be more likely to be engaged in inter-racial marriages or families. The rest targeted areas on and near reservations and might, therefore -- the lack of effect in the American-Indian target sample there might reflect a lower likelihood of American-Indians residing on or near reservations to report multiracial or to exercise the multiple check option.
The CPS did not find but the NCS found a drop in reporting as Asian-Pacific Islander. That was not significantly significant. The RETT found differences -- drops in the reporting of Asian-Pacific Islander in that targeted sample that were statistically-significant. So, this populations seems like one of the most likely to be affected by any option to report more than one race.
So, I think, a key thing here is that, as one hypothesis coming into this research that the populations that were most likely to be affected were those that had the highest inter-marriage rates, you would have identified the American-Indian and the Alaskan-Native and then the Asian-Pacific Islander populations as the populations that might -- the statistics for whom might be most affected by this. In fact, it seemed to be one of the key suggestions coming out of this study.
This and most research conducted to date consistently shows that sequencing the hispanic origin question from the race question produces a nonresponse to the Hispanic origin question and leads to lower reporting by Hispanics as other race and higher reporting as White. It would thus seem that for non-Hispanics who tend to be the bulk of nonrespondents to the Hispanic origin question, having that question first with the instruction to answer both increases your likelihood of, in fact, answering that no, they are not Hispanic. And also that was bolded and highlighted.
And it does seem that Hispanics, having been able to express their Hispanic origin on the Hispanic origin question then feel freer on a race question to identify with one of the OMB race groups rather than to check other race and write that in to try to indicate that they see themselves primarily as Hispanic.
On the combined questions, we see across all three studies that reporting as Hispanic alone would drop substantially in a combined question. In the RETT, where most of the reporting in the Hispanic targeted sample, most of the drop was accounted for by people who checked a box in addition to the Hispanic box, and in that multiple-check kind of context. It is important to recognize that the percentages who checked more than one box were comparable on the panel where we instructed them to check only one in those where we did omit this.
So, clearly, Hispanics confronted with a multiple -- with a combined question, many clearly want to express their racial as well as Hispanic identity and solve the problem whether instructed to or not by checking more than one. Since they did not have the opportunity to do this as readily in the telephone interview, we saw the statistically-significant drop that could not be recaptured. It seems that, in the self-reporting context of the mail returns of the RETT, we would be able to get comparable numbers using the multiple checks and any entry that reported Hispanic, along with whatever else is reported we can get counts of the Hispanic, total percentage in the Hispanic targeted sample that would be comparable to that from the separate questions. However, we found that we were not able to get complete counts using the separate questions as a standard of the specific Hispanic origin groups. This is largely attributable to the fact that 10 percent of the people who checked the Hispanic origin box did not provide a write-in to the second part of that question.
So, the combined question, at least in the self-administered mail return context, seems to work reasonably well for capturing the total Hispanic population, but not as well for capturing specific Hispanic origins.
It worked well for capturing the specific Asian and Pacific-Islander groups who tended to -- whose write-in responses -- who tended to provide write-in responses at a very high rate.
I should also note that the combined question did reduce nonresponse, so one has a trade-off between an improved response rate to the combined over the separate questions, but with the trade-off on the detailed -- the information -- [comment off microphone].
I think that those are the main findings across the three surveys. There are some terminology findings consistently. Black was preferred over African-American, with a substantial percentage indicating preference for each, and substantial percentage that said either term is acceptable or preferable.
I think, by contrast, the preference for Hispanic over Latino was much stronger. We are finding in these studies only 12 percent or so of the population prefers the term Hispanic over the -- prefers Latino over Hispanic. So the gap there is considerably greater.
Some of the other terminology questions, like Chomorro and Guatemalan (sic), are kind of census-specific.
DR. IEZZONI: Thank you, Dr. Harrison for summarizing the exhaustive presentations. I am truly impressed by how the three of you who presented numbers know those numbers like the back of your hand. It really suggests that you have been living with them for years intimately.
Does the Committee have any comments or questions at this point or thoughts?
DR. MANDERSCHEID: One of the questions that occurs to me in each of these areas there is error in any methodology we use. Is there any way of stepping out of these measurement systems to get more to a gold standard of this against which we could compare any of the actual methodologies we are using here, whether we are using paper or by telephone or whatever?
DR. HARRISON: There is some research. I cannot cite it specifically, cite it in the draft which is soon to be published at the OMB Federal Register Notice. Claudette might remember. The Census Bureau conducted some research a couple of decades ago when it was moving from a numerator to self-identification. That research showed that self-identification provided more reliable, consistent identification.
There is much research that Olivia has and has been working very hard with a lot of it in one of the draft chapters by Robert Han in the CDC which, on the other hand, shows the great unreliability of observer-identified race particularly for the American-Indians and for Asian Pacific-Islanders. So, I think that the research would tend to suggest that self-identification is the preferred or gold standard.
Now, whether self-administered or telephone, I am not sure that we have yet definitive research on the advantage of self-administered. We saw, for example, in the almost complete -- 98 percent of the people who chose a multiracial category in the NCS provided write-ins, whereas, the non-response rate, the lack of specific or -- about 12 percent provided an entry that would collapse to one race under current coded screens. The comparable number in the CPS was 33 percent, I think, which probably suggests that that works better in a self-administered mode.
On the other hand, you are much less likely to have non-risk -- [comment off microphone] -- in a caddie or person interview. I think that, once you get beyond self-identification versus observer identification, you would be into a set of trade-offs which would have to be examined closely and where decisions might be driven more, in the end, by the purpose and cost of the survey than by --
DR. TUCKER: The reason, I think, as Roderick pointed out, that, you know, in the CPS, about 60 to 70 percent of the interviews that we do in the supplement were done on the telephone, and the rest were done by personal visit. They were all interview-administered. So, in our case, we do not have the problem with people checking more than one when we only ask for one because we can just read a list and ask them to tell us which one. So, there were lots of differences.
We also have a problem on the telephone administering of more complicated -- like, for instance, in the -- when they did this combined item, that combined race and ethnicity, and ancestry, it would be a very difficult thing to try to do over the telephone. It could probably be done in person because you could have flash cards. But it is just, you know, if -- there were just a lot of these little things that make it difficult. In fact, by doing something in person with this combined question you might be able to insert an interview approach that would allow you to get this kind of detail. Unfortunately, we cannot administer the census anymore through personal interviews.
DR. CARTER-POKRAS: Our office -- [comment off microphone] -- bibliography, and so that goes beyond a couple of the articles by Robert Han that talk about some of these issues.
But the one important point to make is that people sometimes are -- do not realize the difference between reliability and validity. When we are talking about reliability, we are talking about consistency in reporting, asking the same question over time, on multiple occasions over time. We really cannot talk about validity when we are talking about race and ethnicity data collection because who is to say that that is not the correct information that we are gathering through self-identification. So, that is why we usually say self-identification is the gold standard. But, even with self-identification, you may have some problems in not having consistent responses, which is one of the reasons why the National Committee has gone along with the Health Care Financing Administration's position to collect race and ethnic data at enrollment and eligibility determination, and not at every encounter, because then you might have inconsistent responses, and trying to figure out what to do.
One of the things, when we go back to talk about the work plan for the committee -- one of the issues that is kind of left hanging is the fact that, even though that is the official position, the linkage is not occurring. In some cases it is because the technical ability is not there at the state-level. Sometimes it is because the encounter level data is not being collected by the managed care organizations, so there is nothing to link at an individual-level basis. So that is something that we are going to have to -- hopefully somebody will deal with.
DR. IEZZONI: George, you had a comment?
DR. VAN AMBURG: On the content survey, you said that the race and ethnicity is one of a set of content that you looked at. What was the item along those lines for the various race and ethnicity types of questions that you had? Did you do any imputation on that at all and what was the imputation technique?
DR. HARRISON: While Claudette finds the nonresponse rates, we did not do any -- we did not do the typical census imputation procedures on this data set. We treated non-item/non-response items as that. All of the percentages that we report are then the percentages of people who did respond to the item.
The striking thing about nonresponse rates to the rates question in the NCS was that they tended to be about five percent, four to five percent for the population as a whole. The 1990 Census was something like two percent. But for the Hispanic people who indicated on the re-interview that they were Hispanics, the nonresponse to the race item was 25-30 percent for some of these.
So, the key thing there might have been not so much the absolute levels as the striking contrast between the Hispanic and the non-Hispanic population.
DR. IEZZONI: A comment?
DR. AMARO: Do you have any recommendations that you are forwarding regarding changes in how we collect?
[Laughter.]
DR. IEZZONI: I think we have some suggestions.
[Laughter.]
[Simultaneous discussion off the record.]
DR. IEZZONI: Catherine Walman is going to be here on Wednesday.
DR. AMARO: Okay. Is there a date when you are going to be making recommendations that we will --
DR. IEZZONI: Yes, there will be a date, but it is not clear yet.
DR. HARRISON: All of these findings were made available to a research subcommittee from all three surveys, plus additional research and bibliographical literature reviews and public comments.
The OMB has established a research subcommittee which Clyde co-chaired that -- and appointed three people to offer chapters on the multi-racial issue, the combined Hispanic origin, which Olivia chaired, and the various terminology and classification issues. And drafts based on those were completed and we are now in the process of reviewing stuff for publication in the Federal Register Notice the first week of July. These will be the recommendations of the inter-agency committee to the OMB. There will be a 60-day public comment period, and then OMB taking these recommendations and the comments on them in October sometime.
DR. IEZZONI: That is what we will be planning as a committee to comment on.
DR. HARRISON: Right.
DR. IEZZONI: Yes.
DR. CARTER-POKRAS: Actually, it is much more than just the recommendations. It is not just a single sheet of paper.
DR. HARRISON: Yes.
DR. CARTER-POKRAS: Because it has a summary of all of the background work, the research results, literature review, public comments, and all of that have been summarized and the salient issues have been identified and laid out so that, hopefully, people can understand where the recommendations came from, and they were not just grabbed out of air. Some of them were extremely painful.
DR. AMARO: So, there will be a 60-day comment period starting in the beginning of July?
DR. IEZZONI: Approximately, yes.
Can I just ask a question? In some communities and cities in the United States there are substantial populations from Pakistan, from India, from the Middle East. In your cognitive interviewing or other studies, how did those people tend to self-identify? Because White versus Black, versus -- it is just -- I think that there might be --
DR. HARRISON: There is on the census a check box for Asian-Indian. There is not one for Pakistani. Most Pakistanis would check -- there is a box for other Asian Pacific-Islander, and they would check the other Pacific -- other Asian Pacific-Islander box than might be in the Pakistani.
DR. IEZZONI: They would? And the people from the Middle East, people from Iran?
DR. HARRISON: This is an issue that was presented by some of these communities to the OMB. Most people in the middle east would -- most of the Middle Eastern people with origins in the Middle Eastern countries would, in the current directive, be classified as White.
There was a request to OMB that a separate classification be allowed or created on the argument that these populations are distinct and are neither White nor Black. There were other arguments that perhaps the analogue is really to Hispanic origin, where you have people who can be of any race, but who have a common identity, and a common language, history, et cetera.
So, there will be -- and one of the issues that we were trying to test with the combined race and Hispanic origin question of the RETT was whether people in these populations would report one of -- their specific Middle-eastern origins in this.
We are still in the process of -- we -- we are -- what Claudette summarized today is the research that we reported in a preliminary report on the results of the RETT. We are trying to produce a sole report, a more detailed report sometime in the next couple of months. And this is one of the issues where we are trying to examine the reporting on those questions, the -- they are questions based both on the numbers of people in the sample and the lack of a kind of standard for the Asian Pacific-Islander or Hispanic comparison to try to see what assessment we can make of how well those questions work. So that is something that we are going to need a few more weeks to analyze. I would like to add that since the 1980 census we have had ancestry question on the long form. That is where we get also people of the origin that you noted and other people from other origins report in that question. The Bureau did provide Congress with a list of items that we would like to include in the 2000 census, and ancestry was one of those items that we listed.
DR. CARTER-POKRAS: Lisa, Jennifer Mattes, from the National Center for Health Statistics, I do not know if she gave you this charge. She wanted to make sure that the HIS --
DR. GOLDEN: We decided that we would not bring it up, I guess, with the schedule being -- that we would move on. But, I think, perhaps, if we could pass it out, it would be good.
DR. CARTER-POKRAS: Okay. I do not have anything to pass it out. It is just that Roderick was talking about the percentage of people who would select a multiracial option of it was offered. We do have some additional information besides the large-scale test results. The National Health Interview Survey, since the early 1980s, has been including mark more than one race option. They found approximately 1.4 percent, over the years, have selected to mark more than one race.
In addition, there is an adolescent health study that was sponsored by the National Institute of Child Health and Human Development at NIH, and these are 7th to 12th graders across the country. They also did a check all that apply kind of race question and about eight percent of those kids checked more than one race.
Several of the states have passed legislation requiring a multiracial category. Some of the states have also made administrative action decisions to make sure that all of the school systems, for instance, North Carolina and Florida, including a multiracial category.
In those cases where we do have this information in the school systems, it still looks like it is less than two percent. So, the upper limit so far is that eight percent from that study are 7th to 12th graders, but we do not know. It could be, as this becomes more commonly used and more popular, people may decide to select that option if it is offered.
DR. AMARO: And I think that an important point too is that it does seem to differ across populations, you know, the estimates. So that is something important to remember.
DR. BENNETT: Of paper I did a while back, we were looking at how children in interracial households that were identified in the 1970, 1980, 1990 census, we found pretty much -- we were basically using the NCHS rules whether they would identify with the mother or identify with the race of the father or the minority parent. What we found is that consistently, for children who are in black/white union, is that about two-thirds of them will identify with the minority parent, whether that minority parent is black or white or whatever, but that you do not find that consistent parent for the American Indians, or for the Asian and Pacific Islander population -- that it varies in terms of from census to census as to whether they are going to identify with the mother or the father, or the minority parent. There is no clear rule or no consistent pattern coming out or that has emerged thus far for the American Indians and Asian Pacific Islanders.
DR. GOLDEN: These standards are still going to be issued as minimum, and they will still not preclude people from going beyond that. I think, as a construct of what one wants to do, it is to think about the use and the uses. That has been part of the whole issue is that the uses have been varied for different statistical agencies even. And so getting it in on the one hand and using it on the other is really where the real thought has to be.
I think, again, most of the research has been done in terms of the collection end, and it also has been done in terms of a survey instrument that has the ability to have interview and self-identification. The problem comes obviously with the administrative records and where we do not have the flexibility.
We have done some looking at for the birth certificate and the impact of just even the suggestion of a multiracial problem -- who will consent on some of that information (sic). But how to consolidate all of this information, each study is being very enlightening, but now what does it mean at least in the health context? It is one thing, in my mind, to report on the fact of the matter that there are x-number of people in these categories. Thinking about the utility of that as a dependent variable takes on a slightly different dimension. I think that is part of what we have been -- that is part of Han's position, what does it all mean? I think it means -- it is probably a more critical question with the multiracial category and how to use that information in a health context.
DR. SCANLON: It looks like the Census Bureau will be publishing a Federal Register Notice as well for the dress rehearsal just after the --
DR. BENNETT: We issued what last month, April -- we issued April -- was it April? May. Within the last couple of months, we issued the Federal Register Notice that basically said that, if persons wanted copies of the questions for the dress rehearsal, that they could contact -- the Bureau will be issuing in July our -- a similar notice as it relates to the race and ethnic questions that will be tested in the 2000 census. We have not made that decision pending the decision coming out of the OMB.
DR. GOLDEN: What is your date for finalizing your questionnaire?
DR. BENNETT: Our date for finalizing the dress rehearsal is October 15th.
DR. SCANLON: That is when OMB is going to be announcing its decision?
[Laughter.]
DR. SCANLON: That is smart.
DR. BENNETT: It is not just a coincidence. We had to split blood just to get that extension.
DR. SCANLON: All right.
DR. IEZZONI: So, basically, this will becoming out on July 1, and we have 60 days to prepare a response. So, I guess the question for our committee is how do we want to do that? Let me not stress our guests any longer. Do people have additional questions or more discussion about this?
DR. CARTER-POKRAS: I just wanted to announce before we -- I do not know -- everybody leaves -- there is a bill on the Hill to add the multiracial category. So, who knows, if that gets passed, that may be superseding some of the decisions.
DR. IEZZONI: It may become a moot issue.
DR. CARTER-POKRAS: That is Petrie -- right. In addition to that, there is the third hearing in a series of hearings by the House Subcommittee on Government Management that has been moved to July 25th.
DR. IEZZONI: Yes.
Well, why don't we -- our committee stay here and talk about how we would want to address this issue. We are not due to adjourn until 5:30 I hate to tell you guys. We actually have work to do. This has been a very, very helpful presentation, with very clear summaries. We thank you and look forward to seeing the notices of the Federal Register.
All right. Let's stand at our place and take a break.
[Pause.]
Agenda Item: Committee Discussion
DR. IEZZONI: We are supposed to finish our work plan today. So far, we have figured out what to do with the core data set, but we have not figured out a work plan. I think that the first issue is obviously how we want to organize ourselves to respond to the OMB Directive 15 and the July 1 Federal Register.
Let ms just say quite honestly that I do not feel competent to lead a response because this is not an area that I know about. I think that we do need to have somebody on this Subcommittee or somebody who can work closely with this Subcommittee to help sort through the voluminous amounts of information and really come up with a very considered response.
I am not going to type-cast anybody in deciding who that maybe should be. Maybe there is nobody on our Subcommittee who could serve that role. If so, we need to come up with a plan B for how to come up with a thoughtful and productive response to the OMB Directive 15.
Let me just leave it open to the Subcommittee right now. Is there anybody for whom this is an important issue that they know something about?
DR. AMARO: I would be happy to help out with that, especially if I could work with Olivia, and Marjorie could pitch in a little bit.
DR. IEZZONI: Okay.
DR. CARTER-POKRAS: They are also in my section. I could not believe he named who was responsible for that. I am going, oh, no.
DR. IEZZONI: Okay. Then this has been very easy. Hortensia -- let's put this on the record. Hortensia has agreed to work with Olivia and Marjorie when the OMB Federal Notice is posted.
Basically, what I think we should so is e-mail to try to get a consensus. I think that it might be unrealistic to think about holding a meeting this summer.
DR. CARTER-POKRAS: I also want to say that I am really available in July.
DR. IEZZONI: Okay. Olivia.
DR. CARTER-POKRAS: So I really would like to do it in July and get it done because I am not available in August.
DR. IEZZONI: Well, what I would like to do actually is to ask all of our Subcommittee members to tell Carolyn about their summer vacation schedules. I know that that sound silly, but I think that we just need to have a sense of who is going to be able to respond and who cannot. DR. CARTER-POKRAS: Yes.
DR. IEZZONI: Will July be bad for Elizabeth? Are you available in July?
DR. WARD: Yes.
DR. IEZZONI: George?
DR. VAN AMBURG: Yes.
DR. IEZZONI: You are available and I am available in July. So, I think that, if you can get a response to us in July that we will be able to comment and get that out. But then we want this to come from he full committee. So that adds another layer of complexity. I think that we will have to go to e-mail, and we will have to rely on staff at that point to really make sure that it gets disseminated throughout the entire committee, that we come up with responses, edit appropriately, and get this out before the deadline.
DR. CARTER-POKRAS: They will also need to see the recommendations together.
DR. IEZZONI: That is exactly right. I think that the Federal Register -- that could be disseminated immediately. I mean, basically, what I am hoping is that the session on Wednesday will alert the full committee and get them interested in this topic so that when they see this it will not be a new thing to them, it will be something that they know about.
Olivia, could I also ask you and Hortensia to work on this linkage issue that you raised? That certainly can be something that would go on into the fall. But it sounds like something that maybe should be addressed with Barbara Starfield's committee as well, you know, since it goes back to the core data set issues and so on.
Okay.
DR. AMARO: I have a clarification question in terms of our task to come up with recommendations or to respond to the recommendations.
DR. IEZZONI: Yes. Right, to respond to the recommendations.
DR. AMARO: That would be on our behalf.
DR. IEZZONI: Yes.
DR. AMARO: So we are not looking here for volumes.
DR. IEZZONI: No. It should be a letter that would --
DR. AMARO: A response to each of their recommendations.
DR. IEZZONI: Yes, right. And it would come across under Dr. Detmer's signature.
DR. GREENBERG: Do you have a copy of the letter that David Williams wrote for the Subcommittee previously?
DR. IEZZONI: Oh, okay. Okay.
DR. GREENBERG: Do you have a copy of that?
DR. IEZZONI: Well, this would come out under Don's signature.
[Simultaneous discussion off the record.]
DR. IEZZONI: George had a comment.
DR. VAN AMBURG: The Director's 15 recommendations will be essentially, the way I understand it, and you can correct me -- recommendations on what categories of racial ethnic groups should be collected for tabulation.
DR. CARTER-POKRAS: Right. Well, the way they have it structured -- I have not seen the latest version because there are different people who have been responsible for different sessions. Someone at OMB is pulling it all together. But, the way that it is laid out is they have all of the introduction -- you know, these are the kinds of issues, the summary, the research results, public comments, et cetera, and then they have a final chapter that lays out the recommendations, and it goes issue by issue. For instance, as far as multi-racial, you know, what is the response of the committee, and what is the recommendation as far as the Middle Easterner? What is the response of the Subcommittee -- [comment off microphone]? What is the response to terminology and what is the response? So it goes issue by issue.
Then, finally, they have summarized sort of a draft version, a mock-up of what OMB directive number 16 or whatever the number would be. And so most people probably will switch to that to split to that one to get sort of a gut feeling. But it may be that there is an issue that you want to learn more about so that you go deeper into the report.
DR. VAN AMBURG: Okay. But this is really a set of classifications and a way to tabulate data that is a minimum.
DR. CARTER-POKRAS: Right.
DR. VAN AMBURG: The Census Bureau, on the other hand, is going to be publishing something in the Register about how they plan to collect the data.
DR. CARTER-POKRAS: It is going to be almost verbatim as far as the race and ethnicity.
DR. SCANLON: But what about the subcategories?
DR. CARTER-POKRAS: The subcategories, right. They will show about the subcategories.
DR. SCANLON: Okay.
DR. CARTER-POKRAS: But the basic categories are going to be the same.
DR. VAN AMBURG: They are going to tell us or they are going to have the standard question first or combined?
DR. CARTER-POKRAS: Right. Right. They will lay out -- they will lay out like --
DR. VAN AMBURG: That is different from the old directive 15.
DR. CARTER-POKRAS: Well, there are going to be some changes. What they tried to do in the mock-up is they took an old version of OMB directive 15 and they bolded where they would make the changes. Okay? So they laid out what OMB directive 15 looks like and then they had what the mock-up would look like and they bolded the sections.
DR. VAN AMBURG: So, the Census Bureau is essentially going to follow exactly directive 15 on their budget?
DR. CARTER-POKRAS: Right. The new -- directive -- right, 16. The new recommendation, the new recommendation. Okay.
DR. VAN AMBURG: That is interesting. Because the old system was not set up that way.
DR. CARTER-POKRAS: There will be an ability to collect subgroup data which will not be required. Because, you see, we have got a large variety of agencies. We have some agencies that are involved in civil rights monitoring and enforcement. They do not use self-identification. They use observer identification that is provided say by employers.
[Simultaneous discussion off the record.]
DR. VAN AMBURG: The issue that I am getting to though was that we need to have comments on --
DR. IEZZONI: Yes. We can we reach closure on this?
DR. VAN AMBURG: -- the Census Bureau's Federal Register Notice as well. That is what I was getting to.
DR. RIMES: I think that it may be that you could just copy it. You know, in other words, use exactly the same because the comments will be very similar. It is exactly the same letter.
DR. IEZZONI: Okay. Let's have Olivia and Hortensia look at both Federal Register Reports and move on here.
Okay. We have until 5:30. I see faces looking pretty tired around the table. But, Don, out in San Francisco, was a real task-driver and he said that by the end of these meetings this week we have to have an official work plan. So, let's go back to my comments earlier. I will confess that I misspoke. There are five things that I had on the table rather than six. But let's talk again about these two types of areas that are new business, the first being issues that arise that we need to nimbly respond to because they are right on the radar screen, OMB, or HCFA, or somebody has produced a Federal Register Notice that we need to respond to. Then the second issue of coming up with one or two kind of broad content areas that will actually try to have as a major body of work that we would write a report on by the end of the year or some other suitable document.
Okay. The first type is kind of a nimble response to immediate crises. We are going to have a discussion tomorrow about the HCFA and the Home Health Agency. The question that we are going to have to deal with as a Subcommittee at the end of tomorrow's meeting is do we need to do anything more? Will tomorrow's meeting -- I guess Carolyn, we will have some representatives from HCFA there tomorrow.
DR. RIMES: Only.
DR. IEZZONI: Only. We will not have industry represented.
DR. RIMES: Nor will there be anything on rehab or anything on nursing homes or anything else. It is just on the Oasis.
DR. IEZZONI: Okay. Now, this is a big issue for one of the mandates of our committee, which is long-term care, home health care, post-acute care basically, and date of burden, simplification, you know, stressing people out.
DR. RIMES: ADLs.
DR. IEZZONI: ADLs, yes. It has got a lot of ramifications. So, what I would like you all to be thinking about tomorrow as we hear the testimony is is this something that you would be willing to commit another day to? Okay. That other day would be bringing in people from the industry, bringing in people from patient advocacy groups. Okay. So that will be the question at the end of tomorrow is do we want to have another day on that? Okay.
Now, the data release policies. Don, have you been tracking this at all, HCFA's data release policies?
DR. DETMER: Well, just that I heard a little bit about it actually from Paul Cleary.
DR. IEZZONI: Yes. Paul has been on the hot seat on that one.
DR. DETMER: So, I think it is a very good -- a very timely and appropriate thing. I see you have the question up there.
DR. IEZZONI: Yes. I think that this ought to be done in conjunction with Barbara's Subcommittee though, don't you think? Or do you think --
DR. DETMER: Whatever.
DR. IEZZONI: Because I think that this goes beyond concerns about research on --
DR. DETMER: Yes, I guess.
DR. IEZZONI: -- specific populations. I think that it is a general research issue. So, we will put that on the agenda to maybe talk with Barbara about that.
DR. DETMER: Do others know about what she is talking about?
DR. IEZZONI: Yes.
DR. DETMER: Apparently, in the wake of the Website SSA --
DR. IEZZONI: Yes, it is more detailed. It is more health services research than that.
DR. DETMER: One shot across the bow to chill the blood apparently of the folks in HCFA that allow data to go out for research purposes was that -- but, apparently, as a result of that, there has been almost a, seemingly to the research community, a real shutting down or almost a sense of effort to apparently just avoid any criticism as near as one can tell on the research community side after years of obviously having access to such data and without problems. This is seen as not being really in the public's interest and probably one that the public may not even be aware of.
DR. IEZZONI: Well, some in the public are aware of it because that is where some of the complaints actually came from --
DR. DETMER: Yes.
DR. IEZZONI: -- from the Health Services Research.
DR. RIMES: Well, that is a different group from the general public.
DR. IEZZONI: No, no, no, no. The public responding to a Health Services Researcher contacting them by name.
DR. RIMES: Okay. All right. Okay.
DR. IEZZONI: Okay.
DR. DETMER: Okay.
DR. SCANLON: HCFA is reviewing -- actually, to give due credit, HCFA has probably been one of the more broad-minded agencies in terms of the release of administrative data. Data released for NCHS and OCPUR and Census is easy. They basically cannot give away identifiable information except under very strict circumstances.
HCFA has actually been quite forward-thinking in all of this. There is some worry that perhaps they may have gotten just a little bit out too far particularly with these privacy recommendations that everybody is worried about as well. So, they are not conducting a review, as I understand. Carolyn, with her data released policies within HCFA governing both public use data files and what would be termed special arrangement files at either content identifiers, or contain enough information for identification would be likely. So it might be a good idea to have the folks brief.
But to be fair, HCFA is trying to walk a fairly thin line between what they see. You know, to be honest, I should probably go off of the record here.
DR. IEZZONI: Okay. Can we go off of the record? Why don't I talk to Barbara Starfield tomorrow about -- will she be here tomorrow? Do people know?
PARTICIPANT: Yes.
DR. IEZZONI: Yes. Why don't I talk to her? Hortensia?
DR. AMARO: I was just wondering whether this is not really sort of a broader issue for the committee versus just for this committee --
DR. IEZZONI: Right. That is what I was thinking -- it was broader.
DR. AMARO: -- and whether it should be something that gets --
DR. DETMER: A full commitment?
DR. AMARO: -- yes --
DR. IEZZONI: Yes.
DR. AMARO: -- rather than us because it is a larger issue.
DR. IEZZONI: Right.
DR. DETMER: Well, it relates to certainly our recommendations for the Secretary as well.
DR. IEZZONI: Yes. I think -- why don't we raise the issue because we are talking about it? But I agree that it ought to be the full committee because I think that it does have to do with what Bob Gelman is doing.
DR. DETMER: Well, it is already mentioned in the draft document for our recommendation to the Secretary.
DR. IEZZONI: But, trying to do --
DR. DETMER: We come down pretty hard on the side of research.
DR. IEZZONI: Yes. I do not know to what extent we want to be aware of what HCFA is doing in terms of their data release policies right now. Because they are apparently dropping language.
DR. AMARO: They are in transit --
DR. IEZZONI: Right.
DR. AMARO: -- as everything else in HCFA is.
DR. IEZZONI: So, okay. I think that this needs to be full committee.
DR. AMARO: Okay.
DR. IEZZONI: Data for Puerto Rico. Hortensia, talk to us about that a little bit.
DR. AMARO: Well, I think that this is an issue. I know that Marjorie and you have been -- sort of have the history of different times that it has been brought up. There have been a number of times that this recommendation has come up in the past. I think -- I cannot remember whether it was the data castle or previous committees, but the argument has been that there is a very important and sort of ongoing bridge of travel sometimes related to health issues and sometimes just related to back and forth migration from the island and especially to the northeast. And there is an impact in our health care system on both sides. On the one side of that we know nothing, have no information on the health status of individuals, on a whole series of issues related to all of the information that is gathered in the continent. That has important financial implications, as well as implications for our ability to assess the health status of the population on the island, to target public health programs. They are basically out of the loop of our health information systems. This is a population that is very much impacted by the policies in the continent, financing, and visa-versa, and that we really should move toward being informed about the health status of people on the island so that we can provide a better assessment, and in their planning for public health programs, and so forth. So, really, that is all I have to say about it right now.
I would like to think through how -- look a little bit more at the history of where the recommendations have gone, how they have been responded to and maybe some new avenues of introducing this issue again. This issue came up at several other meeting at the Hispanic People 2000 meeting we had, at the NHANES meeting that I went to. It has been coming up and being suggested probably for the last five years or something.
DR. IEZZONI: Does anybody have a sense of where the Department is on this?
DR. DETMER: I guess that the question I have got is where are we on other non-state entities in general.
DR. SCANLON: I do not think we are getting very much. I think that -- as Hortensia knows, on the population in the U.S., the previous NHANES -- there was an Hispanic NHANES that focused on subgroups. I think it would be a possibility even for the next NHANES to focus on Puerto Ricans in the United States. I do not know about -- the Health Interview Survey, for example, I am not sure they have interviews in Puerto Rico either, or in Guam or in other places. It is just sort of a policy I think.
DR. RIMES: The other issue is like some agencies, HCFA, as an example, as part of their Medicare program beneficiary survey, they do interview in Puerto Rico. So there is a real variation across the department. It may be relative to which contractor they end up with, because it is the --
DR. GOLDEN: Probably initially what the mission was -- which was simply starting with a National Health Care Survey, and it just did not move beyond that. I think we have just never revisited from a change initially.
DR. IEZZONI: Do you have a comment on this?
DR. MANDERSCHEID: I was just looking. I have a meeting tomorrow afternoon at the Interior Department with some other people around the Pacific Territories. Some of those get block grant money and some do not. Some refuse to be part of the United States, and now they are independent, and we have relationships with them until the year 2001 and 2009.
I also met with all of those people about a month ago to see exactly where they were because they do provide grants to them as we do to the states. They have several very basic problems, much more basic even than you are talking about -- communication. If you are going to have viable data systems there, they have to be able to communicate with each other and interior to their own area. They cannot do that. They do not have e-mail. They are not on the Internet. In some cases, they do not have decent telephone systems unless there is a large military base in that area. For example, Guam has excellent communications, but American Samoa has nothing. There are some very basic issues here.
So, what we decided we would do is we are going to query them regarding their communications structure, and the current status of their data systems to begin working with them in a systematic way.
But, on the issue of Puerto Rico, there was an interesting thing that occurred to us just recently. Puerto Rico gets block grant money from us. Yet, when we produced the estimation structure for SMI and SCD, we did not have the basic demographic data on Puerto Rico to produce estimates for them, so, they are kind of now worried. They are caught between, on one hand, being required by us to give us estimates of this population, and we have not given them the tools to work with.
We had to go separately to the Census Bureau to get the basic demographic data for Puerto Rico because it is not given out with the Continental U.S. data. So, we now have acquired that. We have to do a special buy. That data literally came in this morning.
But I am a year behind in doing my work with Puerto Rico because I did not have the material to work with. So it is a serious problem, and it is one that will be difficult to get attention to because it does not produce lots of votes in any of these. In Puerto Rico or in the Territories it does not produce a lot of votes, hence it is difficult to get anybody to have interest in it.
DR. SCANLON: The issue would be the consistency of policy as well. I think it has to -- I mean, the implications of looking at Puerto Rico as a territory -- the consistency would be to look at other territories as well in the Pacific and what the precedent or what the practices would be.
DR. GOLDEN: I would think that would probably be more of a starting point in trying to get some sense of what is done in similarly situated situations and also some of the helps and hinderances for doing that. I know we do -- that we have a biostatistics unit in there, which is a totally different system. That is a civil system. I know that CDC has a public health entity where they get some data through their systems there. But, for a national type of survey or some other type of survey, I think we would have to maybe get an approach to looking at it. That might be something that we could do as a part of a work plan.
DR. CARTER-POKRAS: I think there is a general issue about data for trusts and territories. There is definitely a problem with funding in that you cannot demonstrate you have a problem. I remember CDC finding out about a tuberculosis grant where Congress had specified certain locations would receive these funds. And tuberculosis, considering their high HIV and AIDS rates, you would have expected them to be in the top 10. Well, they were not able to demonstrate it because they only had one laboratory that does the tuberculosis test. So, they were not able to demonstrate that they had the high rates. But I would still think that Puerto Rico is unusual among the trusts and territories. One is they are U.S. citizens, and that is not the case with all of the trusts and territories. The other is the proximity. We have got -- they are very close to a tremendous transmigration which, when you talk about the Pacific Islands you do not have that issue. And the third issue is that we have national goals and objectives for Puerto Rico. And so we cannot possibly address infant mortality for Puerto Rico unless we also do something about infant mortality on the Island of Puerto Rico. The mainland and the Island have to be addressed. So, I mean, that -- I think that is the reason why Puerto Rico was singled out. But it is a general issue about trusts and territories that we need to address.
DR. SCANLON: Yes. I think the point is if you go -- you are not going to get away with looking at Puerto Rico alone before you receive other people -- other people who want to take a look at those. You might as well look at them.
DR. DETMER: Yes. But then you can make recommendations on Puerto Rico. I mean, I think you could actually speak to the whole thing, but then you could come down and say --
DR. GOLDEN: We have the other island of -- don't we have the same issue in some of the Asian Pacific Islander populations?
DR. ARAKI: They migrate there.
DR. GOLDEN: Yes.
DR. ARAKI: They migrate to Hawaii. I mean, they have the highest incidence of tuberculosis. There is an epidemic in leprosy that has come up. I mean, they have diseases there. It is not separate. The thing is that Puerto Rico is a Commonwealth Status as is America Samoa. American Samoans are American citizens, similar to Puerto Ricans.
DR. IEZZONI: I was just going to actually propose -- I am sorry, Lynette, but I was just going to propose that we spend a day on this and really kind of learn what the issues are, learn just about the geography in trusts versus territories, and epidemics, and all of the other, tuberculosis, and, I mean, we need to know --
DR. DETMER: That is surveillance reporting. We want to hear that too.
DR. IEZZONI: We need to know that too exactly. That was what I was going to say.
DR. GOLDEN: And the other issue is the very real one that you brought up and that is the census data and how they report.
DR. IEZZONI: Right.
DR. GOLDEN: Usually it is for the Continental United States -- that is if they do not get data for them.
DR. IEZZONI: Well, no, they collect census data out of there.
DR. GOLDEN: They collect it but they do not --
DR. MANDERSCHEID: They do not require it as part of the U.S. They do not ever consider that to be part of the United States.
DR. ARAKI: But they actually do collect U.S. data.
DR. MANDERSCHEID: Yes, they do, but they do not make it available to any one.
DR. IEZZONI: Don wanted to comment.
DR. DETMER: It seems like this, as well as some of the other things that you have on the list may be appropriate also for you to talk over with Ed Sondeck, because he is talking about having this retreat really across the major process exercise and topic exercise for the Center.
DR. SCANLON: You are talking about the workshop.
DR. DETMER: I know, yes. But, I mean, there is a draft of these that seem to me like they are right in the middle of getting this to work. At least, after you made your list, it would probably be smart to interface with him on which things really ought to be on that agenda and then which are not. Then you would deal anyway.
DR. IEZZONI: Is there a staff person who can help us pull together a day on this?
DR. CARTER-POKRAS: Actually, I was going to suggest, as far as timing, Lynette, you are probably involved with the Healthy People 2000 Asian Pacific Islander Progress Report.
DR. GOLDEN: I am.
DR. CARTER-POKRAS: Okay. Pat is. I am sorry. Because they are planning to have that in September, and so they will have a much better idea as to the availability of data from the Pacific Islands at that time. So either we can plan for the September meeting of the National Committee or the one afterwards and take advantage of the fact that we will have a better assessment as to where we are.
DR. GOLDEN: I guess the issue I see is not so much what is available but what is not available and how to get what is not available. We know that very little is available. That is 10-15 minutes worth of discussion. I think it would be important to get a handle on who is doing what and get us up on this same level of discussion. Separate from the issue of why, I am thinking more along the feasibility and the issues around doing it.
DR. IEZZONI: Right. Let's be proactive.
DR. CARTER-POKRAS: It is also very timely because the Department is in the process of developing an Asian Pacific Islander Initiative. We have got the Hispanic Agenda for Action also which we use. We have got a couple of initiatives. This is a new one that is under development. So, we have an opportunity to influence the process or help flesh out that aspect of it.
DR. IEZZONI: Okay. Let's do it. We will have a day in the fall sometime. It might be a little longer, but let's say a day. At the end of that, hopefully, we will have enough to be able to write some recommendations.
DR. SCANLON: We can have a meeting in San Juan.
DR. IEZZONI: Right. Or Hawaii.
[Laughter.]
DR. IEZZONI: Okay.
[Simultaneous discussion off the record.]
DR. GOLDEN: To get back to some of the things that I heard earlier on on standards, in whose purview will it be to look at, for example, the standards and make commentary on their particular relevance? One of the areas I am particularly interested in is socioeconomic status. Most of what we have done has been retrospectively. We have looked at socioeconomic status from what we know about the past. For example, as we have new systems, I am thinking of managed care, I do not know if the trends that we have seen in terms of utilization and all will hold up vis-a-vis new systems and other things impacting upon the same populations that we have studied.
DR. IEZZONI: That is actually a great segue. Because I think before you came in I was kind of trying to spend a little time waiting for Marjorie to start with her talk and talking about some big -- well, no. What we were talking about is finding some big issues like how to deal with the under and uninsured, how to look at the impact of Medicaid managed care, looking at child health. So that is what I would like to spend the last 20 minutes that we have today is thinking about, for the next year, what issue -- let's just do one, okay. Let's do one. What issue really, really grabs people and is politically hot and is something that we can actually make a statement about that could be important and deals with people who are left out in our society in some way?
I gave you some suggestions earlier. We certainly do not need to take those. Does anybody -- Elizabeth, do you have any thoughts about what you might just be really excited to look at for the next year?
DR. WARD: I guess I -- looking at your list, for me the issue is versus what we want to say about something across the country and what are we going to say that is so bland. Because what is really important is what is happening in either regions or states. That is where my dilemma is in terms of looking at your list. That is where I was still having a debate with myself here looking at what would be the issue. What could we influence by doing something that would be looking --
DR. IEZZONI: It is a generic question about everything we do frankly. Don is smiling. It is true.
DR. WARD: Or a directive. Perhaps what we want to do is pick an area and direct it to be done regionally or have some sort of --
DR. IEZZONI: Right, yes.
DR. WARD: -- analysis. I mean, that would be --
DR. VAN AMBURG: I think of more of setting a framework that others can use.
DR. IEZZONI: Right, exactly.
DR. VAN AMBURG: Politically hot in the states anyway is Medicaid Managed Care, and monitoring the effect of that.
DR. IEZZONI: Monitoring the effect of Medicaid Managed Care. Is that something --
DR. VAN AMBURG: There is a lot of concern out there.
DR. WARD: That will tell you -- that will bleed into the health of children, because most of those in the Medicaid Program are children.
DR. IEZZONI: Okay. It seems like there is an immediate kind of interest in that.
DR. CARTER-POKRAS: And also somewhat for the health of uninsured and under-insured because it is up to the states about how they want to handle that issue.
DR. IEZZONI: Okay. So what we will look at is the impact of Medicaid Managed Care, a state issue that leads into children as you said, deals with the under and uninsured.
DR. CARTER-POKRAS: And can I add something? When the amendment -- and as the committee -- what is our title?
DR. GOLDEN: Population-specific issues.
DR. CARTER-POKRAS: That report will pay special attention to the various populations.
DR. IEZZONI: It is kind of implicit. It is implicit in this I think, the interest in this.
DR. GOLDEN: Of the populations that I think is getting a lot of attention in the Department that we have not traditionally looked at is the rural populations. I know that there is an initiative to look at the rural health area. So, I would like to at least add that to our thinking.
DR. IEZZONI: Okay. What I think we are going to need to do -- this is our issue then, is we are going to need to come up with what are the questions? What are we concerned about? What are the states concerned about? Then we will try to develop a work plan around what the major questions are based on who we need to talk to, who we need information from. Do we have a capability to have a subcontractor? If we find something, we would have the capability of having a subcontractor do something special, for example, case studies at different states or looking at policies in different states and that kind of thing. Okay.
DR. VAN AMBURG: Can I add something? It is not necessarily -- the states are going to ask some questions. But it is the questions that the states are going to have to respond to that is the issue.
DR. IEZZONI: Well, that might be what we need to hear, George.
You know, I actually think that we need time as a Subcommittee to talk about this. Tomorrow we do not have time. Waiting until September seems -- well, maybe that is what we are going to have to do. Maybe what we could try to do -- do we have a full committee meeting in September? We do on September 8th and 9th.
DR. ARAKI: There is also a Subcommittee on Health Data Needs Standards.
DR. IEZZONI: Right.
DR. VAN AMBURG: It is in August.
DR. IEZZONI: In August, early August.
DR. ARAKI: August 5th to 7th.
DR. IEZZONI: No, I know. I am teaching that week and I cannot come to that.
DR. GREENBERG: But there is nothing that stops you from having that meeting in September like for one day, if you want to.
DR. IEZZONI: How do people -- I mean, what we could do is maybe have it also to be able to discuss what Olivia and Hortensia have been coming up with.
So, I know that --
DR. CARTER-POKRAS: Actually, there is something else that you may want to consider, and that is requesting a briefing regarding this new report that is going to come out being released by the President around July 1st on children's health.
DR. IEZZONI: Well, that could be part of a --
DR. CARTER-POKRAS: It is something that we went over for a couple of years. Laura McGovern has been working on that. It will have a lot of data that pertains to this Medicaid issue.
DR. IEZZONI: That could be part of a day where we are trying to think about what the questions are. That would be a great way to start out. So, do people think that maybe one day in July you would be willing to fly in again? Maybe we could do -- poor Elizabeth, oh, no. That is starting at 10:00 and ending at 5:00, that kind of thing.
DR. GREENBERG: Is there any way to tie Elizabeth in, in vision, the way we did with John London?
DR. IEZZONI: Yes. I found it to be very useful.
DR. GREENBERG: If we had that capability, it would be good.
DR. DETMER: Yes.
DR. IEZZONI: Okay. Tomorrow during our breakout session, because we need Richard Harding, he is a child psychiatrist, and I think that we need him to be involved in this -- tomorrow afternoon, we will try to schedule a day in July, and what we will try to do is maybe have a briefing from that White House group, maybe spend a little bit of time talking about what Olivia and Hortensia have come up with on the OMB and the Census Response for the Federal Register, and then just basically spend that day fleshing out what the issues are, the questions are and how we as a Subcommittee want to go about addressing them. Does that sound okay?
DR. GOLDEN: As I understand it, that report is just replete with data again and information.
DR. IEZZONI: Well, we should get copies of it.
DR. GOLDEN: I was going to say could we maybe get something before then? Otherwise, we will find that we are just perhaps going to be inundated with information and we really will not be able to absorb it and make a good discussion.
DR. IEZZONI: And what we should probably do is ask whoever presents to us to really focus on the Medicaid and poor children, and ethnic minorities.
DR. GOLDEN: And some summary, yes.
DR. RIMES: There is an existing Presidential Initiative going on that at least HCFA is participating in some stuff there too. So we can do some break-aways.
DR. IEZZONI: That is why this is hot topic.
DR. SCANLON: Now, along with Medicaid Managed Care, I guess, as long as we are in the living room, I think, Ron, you would have issues about the -- if it is hard to get data on encounters for Medicaid in managed care settings, it is almost impossible to bet them on behavioral health care.
DR. IEZZONI: Sure.
DR. SCANLON: So, maybe that can be part of yours.
DR. IEZZONI: That needs to be part of the issue.
DR. MANDERSCHEID: The specific evaluation projects that SAMHSA is supporting on Medicaid Managed Care. DR. IEZZONI: Excellent.
DR. MANDERSCHEID: Some of them were doing it independently, and some were doing it collaboratively with the Medicaid Office.
DR. IEZZONI: Okay, well, we need to -- absolutely, we need to --
DR. GOLDEN: And the House Subcommittee did look at children's issues.
DR. IEZZONI: Carolyn, we do not know where you are going to be.
DR. RIMES: Yes, we do.
DR. IEZZONI: Okay.
DR. SCANLON: We know generally where she is going to be.
DR. IEZZONI: Okay. Tomorrow afternoon we will -- the Medicare Managed Care office, we will be talking. But tomorrow afternoon, again, we will try to come up with a day that we can do this.
DR. AMARO: Along with sort of this packet of background reading, I know, for me, it will be helpful, and there is so much out there now about Medicaid Managed Care, but I was wondering, SAMHSA, I know, has put out some things right in the studies that are being done. Whether there is anything you could send us on that and maybe if there is another document that kind of reviews the questions that are being asked, if there is some kind of background information on that that we could get.
DR. SCANLON: I think that HCFA has a whole -- you have a packet, Shirley, that you will ultimately get waivers on what the characteristics are.
DR. RIMES: And there is an evaluation criteria too.
DR. SCANLON: There is a national evaluation on this.
DR. IEZZONI: Well, we are actually done 10 minutes early.
DR. GREENBERG: Can I just ask, from the point of view of a date, the last week is --
DR. IEZZONI: Right, is the PHC.
DR. GREENBERG: -- is the Public Health Conference.
DR. IEZZONI: Yes, I know. I am going to be there two days.
DR. GREENBERG: We cannot really have it that week.
DR. IEZZONI: Right.
DR. CARTER-POKRAS: We are writing our dates down to make it easier for you.
DR. IEZZONI: All right. Great. So, Carolyn, you will be there tomorrow afternoon with us. Okay. Great, everybody. Are there any final comments?
[No response.]
DR. IEZZONI: Thank you for a terrific meeting.
DR. AMARO: Thank you.
[Whereupon, the meeting was adjourned at 5:20 p.m.]