Subcommittee:
Lisa I. Iezzoni, M.D., Chair
Hortensia, Amaro, Ph.D.
Richard K.
Harding, M.D.
Vincent Mor, Ph.D.
George H. Van Amburg
M. Elizabeth
Ward
Carolyn Rimes, HCFA
Olivia Carter-Pokras, Ph.D., OMH
Patricia
Golden, NCHS
Ronald Manderscheid, Ph.D., SAMHSA
DR. IEZZONI: I think we're going to get started. We're waiting for one more subcommittee member, Hortensia Amaro, but she will be here momentarily.
What I would do to parry just while we are waiting for Hortensia is just go around the room and have everybody introduce him and herself, and tell us where you are from. So those of you who are coming to speak to us, can hear who we are, because we are about to hear from you guys.
[Introductions were made.]
I thank folks for coming to speak to us this afternoon. I had the pleasure of chairing a presentation at the biannual National Center for Health Statistics meeting, the data user's conference. It was quite wonderful, but during this I learned something that I hadn't known before, and that was that our federal government has 45 different definitions of disability.
Well, we are interested in learning about at least four or possibly five today. This subcommittee does have an interest in people with disabilities and evaluating functional status, and how people are basically living their lives.
So Dr. Placek, I guess you are going to start off by maybe giving a sense of one of the major ways that people think about classifying this, the ICIDH.
DR. PLACEK: All right, thank you. I am always glad to have an opportunity to do that.
I'm going to keep my presentation short -- say about ten minutes -- because that will give you some time to really pick out what you really want to know more about. I'm going to really focus on three things: revision activities; current applications, seven in fact; and give you a feel for the new functional status codes.
The North American Collaborating Center housed at NCHS is working closely with the Canadian Institute for Health Information on the revision. Of course ICIDH was published in 1980, and has been under revision since about 1992. So for the last five years I have been working at least half-time or more on the revision effort, and in some cases on designing teaching tools on the old ICIDH, because it is still the only one we have, and will be for the next couple of years.
The most current revision title is up there, the International Classification of Impairments Activities, which by the way, used to be Disability, and Participation, which used to be Handicapped. So we have taken out two negative terms, et cetera, et cetera. For short we call it ICIDH-2, since it is our current draft of the revision.
I'll give you a handout at the end which has all of this overhead information on it, but you can access information about ICIDH and ICIDH-2 from the from World Health Organization Web, and you can also access the entire revision. This is it, ICIDH-2. If you would like your very own copy, send me an e-mail, give me your card today or whatever, and I will be glad to send you one.
I was actually planning to bring ten down, but as I started to check through them, either photocopying was out of order, or the secretary got them out of order. I didn't want to send a confused version out anywhere. So I'll get those straight tomorrow, and send those out to you, or you can download it yourself.
The other thing you can download is all the protocols for beta testing. All of it is up on the Web.
The NACC, the North American Collaborating Center, has established two North American testing centers to be testing this version. One is at Washington University. That is headed by Dr. David Gray. The other is at the Canadian Institute for Health Information. Linda Campbell and Janice Miller are working on that.
We are using World Health Organization protocols to test this in four ways. The first is linguistic analysis and translation from English into other languages. The current version is in 16 languages. We will have this probably in about a dozen languages by this December. Canada is working with the French on the French translation, a Quebece(?) translation version of the French. Then the U.S. is working on the Spanish with people at the Pan American Health Organization.
Then there are 12 basic questions, which sound very short, but they are short, but each question has five parts, and they are very large questions. There is an item evaluation of 50 items which are basically category headings. There are subparts to those 50 questions.
Then finally, options testing of 14 alternatives which have been proposed by the different collaborating centers.
So the protocols are every bit as thick as the classification document itself. These WHO protocols are being used everywhere in the world. Everyone is using the same protocols -- the Turkish, the Dutch, in India, in Japan, and so on and so on.
The process is about a seven or eight year process of revision, and we're well through it. We had initially divided up the territory worldwide for which collaborating center would work on which part of the revision. That has been done. There was a literature review. There was an alpha draft. There was a testing of the alpha draft.
There was a production of the beta draft, and the development of the beta testing protocols for the beta draft. That's the fifth step. So we are now really just starting the field trials using those protocols. We are testing the beta draft. Those tests will end in December of this year.
Then WHO consultants and employees will take all that information that they have collected worldwide, and produce another draft -- call it a beta-2 draft -- in time for the March 1998 WHO meeting in Tokyo, where we will have yet a much more refined beta-2 draft. Then in 1999, it will go to the World Health Assembly for approval.
The last NACC meeting was in St. Louis, Missouri. It was attended by about 40 persons. There were 36 recommendations. In all, North America produced about 200 pages of recommendations which went into making this up. Obviously not all the recommendations were followed, because many of the recommendations were in conflict with each other.
There has been really quite a bit of togetherness between the Canadians and the U.S. Representatives of both countries attended the WHO meeting in April, and they also met here in the Washington area in July to go over the beta testing protocols.
The NACC is financially supporting beta testing with small contracts to Washington University and CIHI. Another center in CDC, the National Center for Environmental Health is supporting several million dollars worth of research using ICIDH-2. There is an allusion there to $2.5 million in U.S. research, using three CDC grants. I'll refer to that later on another overhead.
The fourth annual NACC meeting will be in Ottawa in a couple of months, and the focus will be on where we are in the beta testing. It is primarily attended by North Americans, but other collaborative centers will be invited. The next WHO meeting is in March in Tokyo.
There are seven applications I'm going to talk about today on ICIDH and ICIDH-2. This overhead only refers to the old ICIDH, ICIDH 1980. Just to give you a feel for what has been done with it really in the last year, two or three, RADIUS is the Research Archive on Disability in the United States. It is a project funded by the National Center for Medical Rehabilitation Research. It uses all of the ICIDH classification headings for 16 major surveys.
Every variable in every survey, that is, disability variable, has been given an ICIDH heading wherever it fell, so that one can retrieve information from the 16,000 variables in those 16 surveys according to an ICIDH heading.
DR. MOR: Excuse me, I don't understand how. Could you just go through that again?
DR. PLACEK: Sure.
DR. IEZZONI: At some point are you going to describe to the committee what is in ICIDH so people can have a sense of what the content is? Then maybe how it would be coded might become more clear. Is that something that you are planning to do?
DR. PLACEK: Sure, that will be my last overhead in fact.
DR. IEZZONI: Can we switch?
DR. PLACEK: I'll tell you what. Leave this up for a second, and then I'll get right back to it. I want to finish with ICIDH before we go to ICIDH-2.
So there are 37 major impairment, disability, handicap categories which collect around 12,000 numerical codes for impairment, disability and handicap in the current ICIDH. Those 37 headings are used to categorize all of the variables in all of the surveys in RADIUS. It's just an application of the current -- we call it the old version, but it is still in use, and will be for a couple of years.
A second note, the United Nations Statistical Division only uses ICIDH. It is the world standard for classifying disability. The three products that are fairly recent that I can highlight are DISTAT, which is a database. They basically employed the data from 55 nations and back coded the disability data to ICIDH codes. Then they compared countries for all those data items, so that it is a first attempt really in trying to compare different countries.
Now the countries have used different questionnaires to collect that information, so comparability is a bit of stretch. At the same time, it does make a point that when you ask a question differently, you get different results. We need to standardize our data collection assessment instruments.
They also have a set of recommendations which they have made to all countries in the world regarding the disability questions that they should use in the future population censuses. They have also included table shells to show countries what the information should look like once it is tabulated and published. So they are doing their best to standardize disability information in the world. Then they have a manual for the collection of such information. It is a slick, published manual that sells for about $16.
I won't say anything about the last one, because Gretchen Swanson reported to you all in California.
The heart of the classification is often considered to be the disability section, now called activities. So ICIDH-2 has ten chapters. The ten chapters are listed there, seeing, hearing, recognizing, and so on, 1-10. Each category also has a first qualifier, which is level of difficulty, 0-9, and a second qualifier, whether personal or non-personal assistance is required. That is in your handout.
Now go ahead and put the overhead on that has the small print, that nobody has a copy of.
I just picked out movement, Chapter 4. This is just the beginning of movement. I just picked it as an illustration of an activity. So you can see how the coding structure is. A small (a) is for the term "activity." In the old ICIDH it would have been a (d) for disability.
Then the codes go 301, up into the high 300 numbers. I don't like this coding system, because in Chapter 4 we have the codes that start with a 300, but we might fix that next time or we might not. They did take some of my recommendations on this anyway.
You can see the level of detail of the coding. So a301, maintaining the changing body position is defined as including getting into a position and staying in that position for some time, as well as being able to change the position of a body as required. Then to that you add the first and second qualifiers if you wish.
Bear in mind you can also collapse, since coding is hierarchical. So that if you don't have very much detail, you could just have a very short code which says I3, which would be an impairment in a movement area. If you have only that gross level of information and no detail, you can have a very short code.
DR. MOR: Where did the "I" come from?
DR. PLACEK: It's "a." I'm sorry, "a." So a3 would be a movement disability. A movement difficulty in activity.
DR. AMARO: Limitation?
DR. PLACEK: Yes, the definition --
DR. MOR: Then a301 would be specific --
DR. PLACEK: It's a specific difficulty in a movement activity.
DR. MOR: Okay, then a301.20 would be something about suddenly you have difficulty standing on different types of surfaces?
DR. PLACEK: Right.
DR. MOR: So a plush carpet versus a non-carpet or trips on glossy surfaces or those kinds of things, due to movement, not due to visual acuity?
DR. PLACEK: Right. We don't have a "due to" yet here. What we have is a difficulty in a movement activity, and here is the difficulty, but the cause of the difficulty is most likely found in the impairment codes, where we identify an organ that doesn't work.
In this case, we would probably be looking at -- well, we could be looking at anything, but we might be looking at a leg or a foot problem, or it might be looking at a balance problem. If it would be an "i" code or an impairment code, that would probably lead you to understand why the person has difficulty in this activity.
DR. MOR: So we have the "a's" first; those are the activities. Then we have the "i's" which are impairments, which are the causes? No? I don't really understand. I apologize, because I don't know much about this.
DR. PLACEK: You know what? I think I have to give the apology, because I usually assume too much, that people have so often heard this presentation.
DR. MOR: I know very little at all.
DR. PLACEK: I'll start with ICIDH, which is International Classification of Impairments. An impairment is an organ that doesn't work. Disability is something you can or cannot do. A handicap is a societal disadvantage you experience because of this disability. So that's the conceptual scheme.
DR. MOR: The conceptual scheme I, at least, have some familiarity with. What I don't know is how it actually breaks down into coding. How those three sets of concepts are interrelated in a coding structure.
DR. PLACEK: They may or may not be interrelated. There are no automatic interrelationships. Any given impairment code may or may not result in any given disability code. That may or may not result in a given handicap code; all of our usual and typical interrelationships.
DR. MOR: Are there a whole series of impairment codes that correspond to each of these ten chapters?
DR. PLACEK: It is a different set of impairment codes, but they correspond fairly closely. In the old ICIDH there about 700 impairment codes, about 600 disability codes, about 100 handicap codes. There are twice as many codes in the new ICIDH, especially a lot more codes in the participation section, which replaced the old handicap. That is very much fleshed out. So there are probably 3,000 codes in the new ICIDH-2.
DR. MOR: So what is a 301.30? Is that a complete ICIDH code?
DR. PLACEK: Yes, that's a complete code.
MR. VAN AMBURG: What does it mean?
DR. PLACEK: If you have that code, it means you have difficulty in that area.
DR. MOR: But it doesn't tell you what the impairment is?
DR. PLACEK: This doesn't tell you what the impairment.
DR. MOR: Where do you get what the impairment is?
DR. PLACEK: One would have to do an assessment. It could be a clinical assessment or it could be done by a questionnaire.
DR. MOR: But is there a separate coding system for that?
DR. PLACEK: Yes, a whole other coding system for that.
DR. MOR: All right, so there are more numbers?
DR. PLACEK: Lots more.
DR. IEZZONI: There are 12,000 codes in ICIDH.
DR. PLACEK: No, 1,200.
DR. IEZZONI: How many are there in ICIDH-2?
DR. PLACEK: Pretty close to 3,000.
DR. MOR: So basically there are three separate dimensions?
DR. PLACEK: Yes.
DR. MOR: There is an impairment set of codes, about 300 and something.
DR. PLACEK: Yes.
DR. MOR: Then there is the activities.
DR. PLACEK: Call it disability.
DR. MOR: Then there is the handicapped or participation. There a whole set of other codes. They are not necessarily grouped into these ten chapters?
DR. PLACEK: No.
The ICIDH and ICIDH-2 can be viewed as if we take ICIDH, which has impairment, disability and handicap codes as three separate sets of codes, which are really independent of one another, it's three separate classification systems under that title of ICIDH.
DR. MOR: Can you just give us an example of a potential impairment code that might correspond to a 301.30?
DR. PLACEK: Maintaining a seated position. I'll give it a shot.
DR. MOR: What would be the impairment?
DR. PLACEK: It's a little hard to find it in a snap, but you can maybe --
DR. MOR: Well, neurological. Is there a neurological impairment?
DR. PLACEK: Oh, sure.
DR. MOR: I've not seen these. I know what the diagnostic things might be; I know what might have caused them, but how is it classified.
DR. IEZZONI: To what extent, Dr. Placek, do the impairment codes correspond to diagnoses that you would see in ICD-9-CM for example?
DR. MOR: That's better.
DR. PLACEK: It's better, because I can handle it a little better I think. The ICIDH-2 impairment codes are similar in many ways to some of the ICD-10 codes, the diagnostic codes, but they are not identical, and they do not use the same numbers. They use similar groupings though, similar chapter headings.
ICD-10 is a diagnosis or is a cause of death. ICIDH-2 is an identification of an organ system that does not work, but doesn't identify why it doesn't work. So if you were really going to do a really complete work up on an individual, then you would do an ICD-10-CM assessment to get at the diagnosis. Then that would lead you very quickly to which organ does or does not work and ICIDH-2.
You would need further assessment to get at which activity a person is or is not capable of, which disability it results in. For that we would need what I would call assessment tools, which is structured questions on what you can or cannot do.
DR. IEZZONI: Can I just ask also how much information on psychiatric disease and cognitive impairments is included?
DR. PLACEK: There is a terrific amount of information on that.
DR. IEZZONI: How does that interact with DSM-4?
DR. PLACEK: Again, the coding systems are not parallel or equal. Bridges can be built, but they have not been built.
DR. IEZZONI: Even though all this is being shepherded by the WHO?
DR. PLACEK: That's right.
MR. VAN AMBURG: Can we go back to the first grade here? I'm looking at this 301.20, maintaining a standing position. So this activity involves staying in a standing position for some time. What does that mean? How does one apply that? If someone has that code, what does it mean?
DR. PLACEK: It means you can't do that.
MR. VAN AMBURG: What is "some time."
DR. PLACEK: There is a 0-9 level of difficulty.
DR. IEZZONI: Would that correspond to minutes or hours?
DR. PLACEK: If you will put up the previous slide that has the ten chapter headings.
MR. VAN AMBURG: Yes, I understand the difficulty level, but how does one apply that? What is a five then?
DR. PLACEK: Five would be moderate difficulty. A zero would be no difficulty, and a nine is it's impossible to do.
DR. IEZZONI: Then I guess, George, what you're asking, which I would ask too, is what is "some time?" I can stand for two minutes, but I can't stand for two hours. So is there any definition of what you mean by "some time."
DR. PLACEK: I see your question. On that particular code I don't have a quick answer for you. I'll refer you all to personal and non-personal assistance.
DR. AMARO: Is there a standard definition of those? Of what the value from 0-9 represents, so that people across situations will be rating 5 for the same?
DR. PLACEK: Yes, there are short definitions for what is 0, 1, 2, 3-9.
DR. IEZZONI: It's more moderate, severe?
DR. PLACEK: Yes, of that sort.
DR. IEZZONI: So it's not specific --
DR. PLACEK: In duration of minutes, it is not.
MR. VAN AMBURG: I would think it would be a very, very big classification system. You get that code, you really don't know what it means is what you are saying. If someone had that code and they had a five, I still don't know what it means. Moderate difficulty standing on multiple surfaces. What does it mean?
DR. PLACEK: It could mean on a soft surface, on a regular surface. I'm not trying to --
MR. VAN AMBURG: I know. It's a difficult area. If you're trying to put statistics together and tabulate material based on the coding structure, then you have to know what the results mean when you are tabulating. I'm not sure how we would interpret that.
DR. MOR: Vince, thinking from your experience with like the minimum data set for nursing home patients, which has very explicit definitions for time and --
DR. MOR: Time and place and numbers of people, amount of help and so on and so forth. In general -- I don't know this literature at all -- but I would assume that there have been some kind of inter-relator reliability testing that has been done with this system, is that correct?
DR. PLACEK: Not with this. That's what the testing is for. With ICIDH-2 --
DR. MOR: But was there with ICIDH-1?
DR. PLACEK: Well, with 1, yes. The original version, there were about 1,500 articles and publications that pertain to its use; probably around 30 or 40 pertain to its reliability.
DR. MOR: My reading of the ICD-9/DSM-4 literature on reliability is basically for DSM-4 outside of the major classes, reliability deteriorates very rapidly, and ICD-9, the last two digits are basically highly erratic, without any kind of absolute precision. So I would assume the same thing would apply here. You get 3, maybe 30, but after that it would break down, and the other things would be just far grosser in their classification.
This approach is a nosological kind of approach, as opposed to an item-by-item characterization, although that does have clearly item-by-item characterization components in it, because it's got movement activities. So the movement activities, things like standing, moving, transferring, those kind of movements in it, and so is the depth associated with each of those activities.
I don't know the literature, but I would just imagine the reliability is probably not bad at the grossest levels of aggregation.
DR. PLACEK: Well, I'll address those points, because those are problems, not only with this, but with the whole ICD family. I know that the ICD-10 was tested -- Chapter 5, I know was tested extensively by NIMH, because they have field trial testing centers around the world. NIMH has been very involved from the very beginning, which is why the mental/cognitive aspects of ICIDH-2 are so much more developed.
I dare say that much of the rest of ICD-10 was not tested so thoroughly. Much of it is just out there. So I would say we're doing a pretty good job of testing this, but it won't be a perfect job, but it will be probably on a par with the other ICD family of classifications.
The other thing I can say is about training. We did some ICIDH training tools. We have a computerized training tool called Code IDH, where ten cases come up on the screen. You only work one case at a time. It's a narrative, and it might be written by a clinician. Then you code the impairment disability and handicap, typically 10 or 12 of those codes per case. You get a smiley face if you code it correctly, or a big hint if you do it wrong, and three chances to do it wrong before the right answer finally pops up.
It's a very user friendly thing, because it's on DOS and Windows and it's very cheap. It's $75. So people can practice that coding to get reliability. So we plan the same kind of tools with ICIDH-2 once we are a little bit firmer with the way that classification will finally end up.
I wouldn't be surprised in the next six months if we didn't have a contract in place for ICIDH-2 to develop standardized coding procedures, so that we can have two nosologists code the same thing on the same way.
DR. IEZZONI: We're really struggling to make sure that as our subcommittee colleagues on standardization and data sets think about what kind of information should be collected on each health care encounter, that there be information collected on how the patient is doing. So we are really struggling with trying to think about what we might recommend in terms of how to code that. So that is why hearing this presentation is very helpful to us.
We heard Gretchen Swanson in San Francisco. She was very enthusiastic. She, I guess, has a number of private clients who she serves as a consultant to, who use ICIDH-1. I asked her the question back in June about why ICIDH wasn't used more widely in the United States. She indicated that it is used widely in other countries. I think Europe, maybe France uses it widely.
DR. PLACEK: Do you want an answer to that?
DR. IEZZONI: Yes.
DR. PLACEK: That one's pretty easy. You said there were 45 disability definitions in the U.S. Just about all of them are associated with different programs that were funded by Congress. Those were politically designed programs, that when a pressure group decided that something needed to be done for veterans in certain categories, and so it goes.
The definition is really politically crafted by the AIDS. They scurried and looked for definitions, and trying to be sensitive to the political realities, and they often compromised on what categories are covered or not. This is how they devolve. This is why they are wedded to the definitions they are using.
Bear in mind that none of the definitions they are using are either U.S. standard, world standard, or any standard. They are just their own. They just created them, because they needed them for purposes of legislation.
DR. IEZZONI: So why, since ICIDH has been around since 1980, why haven't we seen it more in the U.S.?
DR. PLACEK: Well, first of all, it was done by the WHO and published really in English, but by a British physician, so it was not an American product. I think we're very ethnocentric. The first translation was into Dutch. Naturally it was into French and so on. Of course, we didn't need to translate it, but then why should we pay any attention to it, because we didn't invent it? I think that is the quickest, simplest answer.
Another one is it is the standard of course, for all their programs, for all their policies. It's the gold standard for their programs, for their disability programs, for their employment programs.
It was picked up in Canada even before the United States. They did three national health surveys structured around ICIDH. Now federal agencies are beginning to pay more attention, especially as it is being revised, and especially as quite a few Americans are involved, because the Americans are in tune with U.S. data and policy needs.
DR. IEZZONI: When you say federal agencies, who seems to be the most interested?
DR. PLACEK: Let me answer you with a slide that I didn't put up yet. It will start out with National Center for Environmental Health on the top. The National Center for Environmental Health, which is a part of CDC, just funded as of August 1, three $800,000 projects to develop assessment tools that are really structured questionnaires and so on, so that the information can be directly coded to ICIDH-2. So these contracts are now in effect with Gale Whiteneck and David Gray. So there's one.
Also NCEH is building a new BRFSS disability surveillance system using ICIDH to orient its disability questions. The BRFSS is a short telephone survey done in all the states, but they are adding a disability module, 25 questions, and they have funded 15 states $350,000 a year for four years. Next year they hope to fund all the rest of the states. They will be codable and back codable to ICIDH- 2.
DR. MOR: You mentioned that a number of times, the codable and back codable into ICIDH. If would be very helpful to me -- is there a protocol for how one does that? Or is it specific to the given illness?
DR. PLACEK: If you look at a structured questionnaire which asks do you have difficulty in seeing? We might follow it through like seeing an objective clearly at 20 feet or something like that. Now it would be more specific than that, but if you get a yes to that, and specific information about how much difficulty of seeing, then that would be codable. You could be able to put a numerical code, an ICIDH code.
DR. MOR: The National Center did a disability survey.
DR. PLACEK: Right.
DR. MOR: Is that back codable?
DR. PLACEK: Yes.
DR. MOR: Was that designed for the back codable purpose?
DR. PLACEK: It wasn't designed that way. Any survey can be back coded. Any disability survey can be back coded to ICIDH. In some cases there are judgment calls, but the disability survey was not designed specifically to be back coded to ICIDH-2, although it will be, and is being back coded to it.
DR. MOR: So you're creating, somebody is creating a code of 70, 80, 90 different questions on disability, impairments and inability to function, what kind of assisted devices, all those kinds of things, and collapsing them into three separate codes?
DR. PLACEK: Yes, those 70, 80, 90 questions, if there are yes hits in some area of disability, it might convert to eight or nine codes. If the questions can give you enough specificity, then you can go to maybe third digit or fourth digit on an ICIDH-2 code.
DR. MOR: In some of the dimensions?
DR. PLACEK: Yes.
I'm taking up more than my share of time.
DR. IEZZONI: This has actually been very helpful I think. It has given a lot of folks much more information.
Are there any pressing questions for Paul right now before we move on to some of the other speakers?
MR. VAN AMBURG: I'm just wondering how they are going to structure the questions on the test to get any consistent and reliable data.
DR. PLACEK: The questions have been tested in the NCHS cognitive lab. Last week a preliminary report was given to the CDC on the BRFSS. We had all the state people for these 15 states that are going to be using them. Barbara Wilson is here today, and she works at the lab, and she was directly involved in testing the questions. So you might have more questions for Barbara.
MR. VAN AMBURG: I'm kind of interested in the back coding of this from a telephone survey. I'm looking at what you had up there on the movement. If they have a question on that, how would you back code it from one of those?
DR. PLACEK: Most likely it will be to a one or two digit level. You are just being able to say that there is an activity impairment at the moment level.
MR. VAN AMBURG: But you won't get the level of difficulty, the 0-9 codes?
DR. PLACEK: Probably not, no. You just get the fact that there is a difficulty, and what kind of difficulty of movement or whatever -- seeing, hearing.
MR. VAN AMBURG: That's almost dangerous from a policy perspective.
DR. PLACEK: Well, I believe it's four areas of disability, not all ten, that are being targeted in these 25 questions. It is not impairments, it is not participation codes, it's just four of the ten disability areas. They are being designed so that they can be collapsed into these areas to identify areas of difficulty or disability.
DR. MOR: So it just all of the sudden dawned on me, a very obvious thing. Somebody could clearly have multiple ICIDH coding sequences of nine or ten digits?
DR. PLACEK: Yes.
DR. IEZZONI: California now has 25 spaces for ICIDH codes.
DR. PLACEK: The parallels are exactly, in my view, the same as in mortality coding, where you have an intermediate cause, a contributing cause, and an immediate cause. You code all those things, up to 20 or so causes of death, but there is a sequentiality to it.
In my view, there is a sequentiality to disability coding, where impairments lead to difficulties in activity, which lead to restrictions in participation, barriers in societal participation, facilitators. It's the same sequence, except it focuses on those of us who are still living.
DR. IEZZONI: I suspect that ICD-10 would say that they code morbidity, as well as mortality. So we can argue that.
You're going to stay, aren't you?
DR. PLACEK: Sure.
DR. IEZZONI: Okay, great, so we can hit you again if we want to.
DR. PLACEK: Oh, sure.
DR. IEZZONI: This is all information, and very friendly. I hope that you understand that we have learned a lot from this presentation.
We would like to move on maybe in the order that is listed, although Michelle, have you gotten your breath? Census is very interesting, because the year 2000 census approach is being designed as we speak right now, isn't it? Yes, and you're going to be asking a little bit about disability, so let's hear what Census is thinking.
MS. ADLER: This handout that is clipped together is something from a decennial census notebook that has been passed out, which I thought you might find interesting. The decennial census, from a research perspective, is very important, because it is basically the only game in town where you can get state estimates, county estimates, estimates for MSAs, and even block depending on what you look at.
We just still on our national population-based surveys sometimes we are able to get estimates for large states or for medium-sized states to pool data. Basically, we just can't get local area estimates except from the decennial census, but that is not its purpose. It's the only survey in the Constitution, and the purpose of the decennially census is to come up with the apportion for the House of Representatives.
Every ten years, when we get involved with the Census Bureau and their various work groups, we are reminded incessantly that Thomas Jefferson was the first leader of the Census Bureau, and they pull out John Adams and all these people, the Founding Fathers compared with our secretaries. They are really into apportioning.
The first census in 1790, just basically got the name, rank, serial number, how many people in the household, men, women and there were the racial questions, but that was it. It was just for apportionment. Gradually they started adding on more social variables. In fact, in 1830, they started adding health and disability variables.
Every ten years we go through the exercise that the Census Bureau needs money from the House of Representatives in order to be conducted. The purpose of it is to find out who will be in the House of Representatives. So it becomes a very sticky political issue.
Their bottom line, if you go into this handout a bit, is that you need a statutory basis for whatever is on the decennial census. So we have come up with statutory uses and citations for disability data. Now some data on the decennial census, it will say directly in legislation that you need that data at local levels as collected by the decennial census.
For example, the Voting Rights Act for civil rights says that directly. Other legislation will say things like, you should get information at local levels, but it won't restrict you as to how you get it, or where you can get it, or perhaps even if you can get it.
Their first priority is any kind of language in legislation that directly says you have do this on the census. Unfortunately, there are many statutory uses of disability data through the ADA and from the Department of Transportation; EPA has some rules; the Older Americans Act. We have lots of legislation that refers to disability, but nothing says outright that we should get it at local levels from the decennial census. They call a program need, as opposed to a required need.
So we have kind of snuck in through the back door on health and disability data. In a decennial census 100 years ago we did actually ask some health questions, but have really not since then, which is unfortunate. That's my own personal opinion.
The second handout I have is the questions from 1970, 1980 and 1990, just to show you how far we have come. In 1970, we still only thought of disability as a work disability. If a disability, whatever that is -- and I'm glad that I'm following Paul, where he has talked about is disability -- but basically the disability was something that prevented the working age population from working, or limited the working age population from doing whatever it is that they were supposed to do, and that was it. There was nothing on the elderly, nothing on children. That was the only aspect that was looked into.
In 1980, we still had information on work disability, but there was this question, 19c, which has got to be one of the worst questions ever on a decennial census -- "Does this person have a physical, mental or other health condition that has lasted for six months, and which limits or prevents this person from using public transportation?"
I remember when that question came in, and DOT swore up and down we needed that. Of course the answers, people would say, yes, I live in Montana and we don't have public transportation. So no one understood the question. I've never seen anyone analyze it; that doesn't mean no one did, but that question fell by the wayside.
In 1990, which is the next page, this is the census that had the most disability data. We still in question 18 had the two part work disability question which turned up some time during the Depression. This time we wanted to get more of a flavor of disability, so question 19 is a proxy for IADLs, and question 19b is the proxy for ADLs.
We've got, "Because of a health condition that has lasted for six or more months, does this person have any difficulty going outside the home alone, for example, to shop or visit a doctor's office, or taking care of his or her own personal needs such as bathing, dressing, or getting around inside the home?"
Well, this was expanded more, and question 18 and 19 were asked of everyone age 15 or over. We know that most people who are 90 years old probably wouldn't make much sense of the work disability question, but nevertheless it was asked. In the tables that came out of the 1990 decennial census they are divided into ages 16-64 and 65 and above.
I have a paper on results from the 1990 census if anyone is interested. The estimates actually weren't that bad. This is where HHS did push radically on the Bureau of the Census. On the ADL question for the non-elderly we got 7.2 million, ADL or IADL question. SIP(?) got 7.6 million, and the disability survey got 6.7 million. For the elderly the 1990 census came up with 5.9 million; SIP had 8 million; and the disability survey had 6.8 million.
One of the things that happens with the year 2000 is that the Census Bureau judges all questions on index of reliability. They have a national content pretest. They have a variety of tests, including in the 1990 decennial census -- this was mailed out to about one-seventh of the population, which is 40 million households. About 1 million households were given the decennial census and questions from SIP to test the reliability and validity of the questions, and also to see if some of the interviewers were perhaps not curb stoning or answering the questions themselves.
When looking at that data, we found out that particularly for the ADL and IADL, the index of reliability was high, which is bad, which means that in their eyes that people at time one had said they have an ADL for example, and then at time two they had an IADL or vice versa.
Now at HHS we have carried on a dialogue for many years where we say that the ADL question and the IADL question are very subject to change over time. All of our population-based surveys, the longitudinal ones bear this out. The National Long-Term Care Survey conducted by Duke and the National Institute on Aging -- I don't know if you have looked at this -- and the MCBS, but people do change.
The chances of becoming disabled when you weren't disabled are not that high, but if you are disabled to start with, there is a big chance of going back and forth in the state of disability. Anyway, we couldn't get that point across, so they decided that these were bad questions.
DR. MOR: What was the time frame in between the SIP and the --
MS. ADLER: Eight months. No, I'm sorry -- yes, eight months was the first one. So they decided that this was not so good. They did another test around 1993, where it was five weeks in between, but the first interview was done as a mail out like the decennial census. The second one was the follow-up interview by telephone. They didn't have any provisions for getting the same person.
We also know from the National Medical Expenditure Survey data from an experiment, that if you change the mode from mail to telephone or telephone to personal, ADL and IADL questions are very sensitive to change in mode. They are also very sensitive to respondent. For example, if you ask a person with an ADL limitation the question, and then you ask the care giver, you are going to get a different answer. Usually the care giver will say the person has a more severe disability.
Well the upshot was that the Census Bureau designed a series of questions in January of this year, which is on the third page of this handout. The purpose of this was to expand it. By some miracle, we actually got more space on the decennial census.
It has taken me a while to realize this, but the decennial census, unlike any of our other surveys, is format driven, not content driven. We are reduced in the end to getting everything in one and three-quarters inches. It's a square root type of exercise. It's not a content exercise; all the reasoning and legislation in the world, it always comes down to that.
Nevertheless, we are slowly getting more area, which is an incredible victory; I can't tell you how much. In the 1990 decennial, even though it was so much better than 1980 when we had that silly public transportation question, we were criticized, and rightly so, because we didn't have any questions on children. We didn't have anything on total impairments.
So the Census had a version which has got a question on mental tasks such as learning, remembering and concentrating. They had an IADL question. Then there was a question with several parts -- that's the other things, people who design questionnaires are adding several parts -- picking out some disabilities from the American with Disabilities Act.
That's where it gets tricky. In the disability survey we had hundreds of questions on disability. I think there were more than 100 just on children alone. Depending on what you pick, you are going to get different estimates of disability. You could take an approach and say, well, let's pick the most popular disabilities and get a really big number, which would be something to do with lifting, bad back, and maybe depression, probably walking three blocks would be biggies.
If you did that though, you would get a big number, but you would probably overrepresent the elderly; people with arthritis, bad backs, and again, you would miss children. So it's a very difficult thing to try to figure out how to fit questions into a little space that would pick up as many aspects as possible, and make sense given the complex nature of disability.
On the last page, the federal agencies wanted the Census Bureau to re-look at the disability questions, so OMB convened a work group this summer which met two or three times every week at OMB. That is how many of us spent our summer.
That version and the Census version are being tested right now in a cognitive lab. In fact, tomorrow I'm going one block up the street to hear two interviews. The results of the cognitive tests will be done probably next week. They'll start writing them up.
There may be a dress rehearsal in the spring. The Census Bureau is currently -- they don't know if they are going to get funding for that. This little controversy is part of a larger controversy that the Census is facing in regard to sampling for the House of Representatives.
Should that go ahead, one of these versions or parts of the version would be in the draft for the 1998, and they will have to start printing up the forms.
DR. IEZZONI: Are the questions that are being tested in cognitive testing right now just the last page or the last two?
MS. ADLER: The last two. The Census Bureau an ADL question onto their version.
DR. MOR: Which ADL question?
MS. ADLER: It is on the 2000. They did add an ADL question, which comes out as 7b in this version.
DR. MOR: Dress, get around inside.
MS. ADLER: On the last page, what we have come up with and what is being tested for they call it the OMB question, but it was really derived by the work group composed of the federal agencies: HHS, the Department of Justice, the Department of Education, the Department of Veterans Affairs, the National Council on Disability, and the President's Committee on the Employment of People with Disabilities, and to some extent Paul's group, an interagency subcommittee on disability statistics.
What we did is given again this little space, is look at -- the first thing is we tried to have as many questions as possible which were relevant not only to adults, but to children so we could get some estimates on children. The first one is do you have any difficulty doing any of the activities. One is learn, remember or concentrate. In that question we hope to pick up mental impairments such as mental illness, Alzheimer's, learning disabilities. It's a broad mental impairment question.
The second one, talk, see with glasses or hear is a sensory impairment question. The third, walk three blocks or lift a bag of groceries of the two most common non-G(?) physical functioning items. We have the physical aspect.
The next one is work or keep house, which is a role way of looking at disability. What functions can you perform. Then E and F are the IADL and the ADL; together the ADL and the IADL, with those questions, we can come up with estimates of long-term care needs.
That is the disability part, and we found lots of legislation. In fact, the Older Americans Act refers specifically to basic activities of daily living. So we feel pretty good. It depends on how the question is worded, but we felt pretty good about the legislative underpinnings for these questions.
The next one is more controversial. The ADA does not have the data requirements which say we have to get data on the U.S. Census, but the Department of Justice is very interested in locating geographical pockets where there is a lot of discrimination in employment against people with disabilities.
There are no questions. The question is, does this person need special work arrangements for people with disabilities? We had quite a time taking some of the disability jargon -- we have used these words so often, we tend to forget what they mean -- work accommodations and accessibility and turning them into English. So we'll see if this question works, and basically get at a measure of how many people need these things in order to work.
You may notice that from this version and a Census version, we don't have the old historical work disability question. The reason we don't is that the disability community was very vocal in saying it wasn't the disability that kept them from working, it was the lack of accommodations or suitable environment that kept them from working.
So the work disability question is still used on the SIP for the NHIS and it is still there, but I can see that it is becoming fossilized.
DR. IEZZONI: Will this cognitive testing involve people from different languages?
MS. ADLER: No, in the cognitive lab we are limited to only ten interviews for each version. Some of the interviews will have like five or six people in the family with disabilities, so they assure me it isn't that bad. We may have a field test of these versions also.
DR. IEZZONI: So it will be ten tests in English.
MS. ADLER: Yes. That was a concern, especially from the Department of Justice, that not only these questions but the whole decennial census be available in languages, and also be available in big print, and be available on TVD or TYY for people with hearing impairments. There is someone at the Census Bureau who is working on that big, general issue. So we feel that that is being taken care of. Right now we thought we would concentrate on English.
DR. IEZZONI: Well, this has been a very eye opening presentation I think for all of us around the table. I am anxious to hear from the other speakers, but let me just pause and see whether any of the committee has a quick question.
DR. AMARO: You spoke earlier about the change of not just asking people about functioning at work, but in the home. Was there any consideration on the phrasing of that question 9 that it might be important to also look at whether people are in need of -- to get a sense of how many people are in need of special equipment, et cetera in the home to carry out their duties in their home?
MS. ADLER: Yes, but we just didn't really have the room for it. That is looked at in depth in the disability survey. It is sampled in some of the other surveys. There is so much that we did have here. We didn't have assisted devices in this section, and we really wanted it, but we were so limited by the space.
DR. AMARO: This is my area of work that I have done some work collaborating on some studies looking at functioning across different cultural and regional groups, and also with across gender and how that is interpreted. When we asked people about functioning and roles in the family, we got a lot of other information, especially for women that didn't come out around functioning at work.
I was just wondering whether, for example, women with children, you might be missing an important feature about women's ability to carry out their duties in the home.
MS. ADLER: It's mostly because the ADA has a very particular focus on discrimination in employment. Many of the suits on the ADA have to do with creating a barrier-free work environment. One of the things that we are using to track the progress of the ADA is the employment rate of people with disabilities.
We have very little information. The disability used to harass for several years ago, which was really not the first choice of those statisticians. It's just a Harris poll. I don't remember how they defined disability, but it had to be done on the telephone very quickly. Say, would you work if you could work? That is basically all the data that we have.
DR. IEZZONI: Is there any opportunity for public comment about the census questions?
MS. ADLER: Yes. I can't remember their schedule, but basic topics were presented to Congress last March. Then the questions will be presented, and I don't know their time.
DR. IEZZONI: We should probably we aware of when that would be. Can you find that out for us? Because we might want to comment on some of this. I know that you have like square inches that you have allotted to you, and it sounds like it really is territorially driven, but it might be our at least expressing some of our comments.
DR. MOR: Just one comment. So are you going to be there when they actually do the cognitive testing lab, and observing questions and responses?
MS. ADLER: Yes, two tomorrow.
DR. MOR: I have done a lot of work with a word called "need." Pay very careful attention to what the interpretive processes are of that word. One reading of this question is they are doing it for me. Another reading of this question is they are not doing it for me, and that's why I'm not getting it. Another reading of the question is that both of those apply, the union set of those, depending on that interpretation of the word "need" here.
It's a very complicated question because of the parenthetical four examples, but the root question is very open and ambiguous, because of the verb. Just be very careful in that. You could end up with worse than what the 1980s version was.
DR. IEZZONI: I know, that's actually how I would answer that question. While you were talking, I am not sure how I would answer that question, given that I work in a hospital, so there are ramps in most places.
DR. MOR: So if you are not incurring the barriers, and you're all set, then no, I don't need it. I'm okay.
MS. ADLER: That is one of the things we're testing, because we need child care in order to work. So it's a very difficult concept. One of the versions we came out with just had the language from the ADA, and it is a question in the disability survey which is do you consider yourself or do other people consider you to have a disability? On the disability survey there were 19 million people that answered yes to that. We didn't think they were quite ready to put that in a decennial census.
One last remark. We have taken this issue to the American Statistical Association and asked them to work with us to come up with another test besides the index of reliability that the Census uses that would be more responsive to health disability questions. I think that would help us not only in the future to get more health and disability questions on the decennial census. Think about how the health status question would fare in such an environment.
DR. IEZZONI: Hortensia has a quick question.
DR. AMARO: I was just thinking about how someone with a mental disorder or a psychiatric disorder or an addiction would answer this. I was just wondering, if you are trying to capture that whether the four examples can pertain other things besides physical environment, like maybe you have to have a program to assess and detect and support people with addiction in the work place.
It is makes you think only of physical. I wonder if you are going to miss all the people with psychiatric and substance abuse.
MS. ADLER: I know, that's a good point. All I can think of are changes in work schedule or assignment. You're right, that's a very good point.
DR. AMARO: Like having an employee assistance program might be a condition of a work place.
DR. IEZZONI: These are excellent comments. I guess the reflection is we are spending a lot of money on the census. We want to make sure that the information we get for our money can be interpreted, and that it doesn't do us a disservice.
DR. MOR: At least it is explicitly stated what it is, and what it isn't. That would be even more important.
MS. ADLER: Well question 1 may not work. I think there are probably more people who think it may not work than think it may work in a decennial environment. It is clear though, that in our national population surveys, we do need information like this, and we do need to put it on now as part of the regular employment questions.
One of the reason this is kind of awkward is we have it on the disability section. It doesn't belong there. It belongs in the employment section. We say, did you work? Why didn't you work and so on and so forth. This could be one of the reasons why somebody is not working.
DR. IEZZONI: Okay, well, thank you. This has been really fascinating. We'll continue to find out when will all the testing and public comment on that. We'll be very interested to hear what the results of the interviewing is.
Howard Bradley, you have been quietly sitting there taking copious notes. Now we can hear from you.
MR. BRADLEY: The disability evaluation study; that is not called survey anymore. It's study. The first page is basically the four purposes of the study. The contract for this has not been issues. The RFP will be out in the next six months we estimate. So it is still very much in the formation stage, pre-contract stage.
The four priorities on the bullets have shifted over the years. The study goes back to 1983. It was first proposed and then killed by funding considerations. In 1992-93, SSA and Congress were concerned about the growth of the disability rolls, which are funded in part by trust funds and in part by general revenues in terms of the Supplemental Security Income Program. Both programs were growing at such an alarming rate that the trustees and Congress ordered us to do research in these areas.
Several short-term projects were funded, and the longest term project was the DES. It has also been delayed the longest. The short-term projects located certain areas of growth that we had suspected related to the baby boomers, the growth in the children's programs, and the growth in mental illness, muscle-skeletal impairments over the period in question which was 1989 to 1992.
They could not quantify all of the growth. They quantified only about 50 percent using econometric analysis. So the larger question remains, and that is the first bullet here, how many adults of the ages indicated would meet our current eligibility criteria, and what are the characteristics of those adults?
We're looking at the adults primarily, because the children -- we knew that there was a great deal of push in Congress to do something about the children's SSI program, and they have done that. It was part of welfare reform, and we saw that coming. So we concentrated specifically on the adult population.
The ages, as you know, at age 65 Social Security beneficiaries can get their full benefit. Early in the next century the full retirement is due to increase to age 67, and several proposals that have come out from the recent advisory panels proposed raising full retirement to 70. So hence our interest goes up to the population through 66, because of those concerns.
So the first bullet is more or less a snapshot. What is going on out there in terms of our current eligibility criteria?
The second bullet is related to what we call the disability process redesign. Under the current definition of disability, keeping it the same in the law, how could we better evaluate adults in terms of their ability to work?
How can do we present programs better so we get more accurate decisions, more consistent, ones people can understand, and ones that are easier to administer from our perspective? We get several million claims a year, and frankly it is a huge headache. It is a huge administrative burden, and that drives of course the public discontent and congressional pressure on us to do something.
The third bullet is something that SSA has been interested in over the years. If you have a person with an SSA-level disability -- I'm coining that term -- someone who could come on our rolls, but who is still working, what are they doing to be able to work in terms of the combinations of interventions, and things like that.
The fourth is the long-term, what can we do in the future to continue to monitor progress? The Venn diagram is worth a few hundred words at least. The big circle is the adults in the population, the working adults. Then there is this kind of a more a mass of people that probably straddle the line, who could meet our current eligibility criteria.
So I'm going to talk about bullets 1, 2, 3 and 4 in that order in terms of our current disability definition. I have another handout I will give at the end -- it's very technical -- that you can take home with you and read.
The first priority was to figure out how many adults in our current population would, if they came in and filed, stopped working, came in and filed, how many of those would meet our current criteria? The definition up at the top, and that is taken right from the law, is again, the work definition. Defining disability for adults in terms of work in the national economy based on medical impairments and their duration of 12 months or terminal illness.
So that's what we are working with, and that, by the way, is also what the redesigns worked on. The redesign that you probably read about is a redesign of a process, not a redefinition of disability. We will talk about that in a moment.
When we do our current process we basically decide disability claims on some technical factors before we get into the medical evidence. These are typically, determination is made at the field office. The local Social Security Office decides whether the person has enough quarters of coverage to meet our Social Security disability insurance, or whether they meet the means/residency tests for SSI.
Of course aliens are in that mix too. As part of welfare reform there are certain citizenship requirements, and if they don't meet one of those, they are out. There is no medical development; it doesn't matter. They all just say who they are. They don't come into our program. They are weeded out, denied more quickly.
You have a non-working spouse of a working person. The husband was a school teacher. The non-working spouse had not enough work history to be sure under our program. She had ALS. She was disabled and she doesn't get anything. She doesn't get any benefits.
So we're not interested in those people. We are in the survey, to be able to identify how many people are in that group who would meet our criteria if they went back to work for example, or if their family went down to the SSI office; things like that.
So in response to this, the DES, in italics there in the middle, will collect work history, income level, current work activity. That is basically to be able to put the data we collect in this study in front of an adjudicator and decide whether that person would have met our criteria had they filed that day.
Is everybody on board?
The more important part of the study are the disability-related factors. In our current process we collect data from treating sources, and we also purchase exams as needed from specialists for certain tests that have not been by treating sources. We obviously get self-reports from the claimant as they go through their disability interview in writing.
The medical file and self-reports and third party reports are reviewed by a state agency which SSA funds to do this activity. It is done by a team of a physician or psychologist and examiner.
So to make estimates of whether people in the adult population are disabled or not for our criteria, we have to at least collect some of the same data that is done in our claims processing, whether the person is on the rolls or not. So to be able to have good estimates, the DES will have to collect the same interview data and will also have medical and mental status exams, and will also collect data from treating sources.
This data will be collected in a format basically, and put in front of a disability examiner or an SSA employee for research purposes again, to confirm that they can estimate of whether that person would be eligible under our criteria.
DR. MOR: Will only people who have filed a claim be in the sample?
MR. BRADLEY: No. It is a national sample of adults.
DR. MOR: So if you are going to be collecting data on me, and I haven't filed a claim, someone is actually going to do psychological and medical tests and consultative exams?
MR. BRADLEY: Well, probably not consultative, but we will ask you to undergo a medical or mental status exam.
DR. IEZZONI: How many respondents?
MR. BRADLEY: Right now that is kind of in flux, but 5,500.
DR. IEZZONI: And will you pay these people?
MR. BRADLEY: Yes, probably, with OMB approval. Again, we get to the issue of the size of the survey -- and Michelle's comments on the census also apply to us --
DR. MOR: Well, no, it is more than just the size of the survey. This is an issue of a particular kind of bias, and the value of the time of the individual to subject themselves to these kinds of tests will directly relate to the economics of the value of their time as a person with an impairment or a disability who chooses to continue working.
DR. IEZZONI: Yes, for example, would you agree to this?
DR. MOR: How much is the value of your time for going to some kind of such and such test given the opportunity costs, when you are exactly the kind of person who, because of what's between your ears as opposed to other parts of your body --
DR. IEZZONI: Right, exactly.
MR. BRADLEY: Do you have any ideas in this area?
DR. IEZZONI: Pay well.
DR. MOR: As an economist would say, you would have to have a differential payment source as a function of income in order to have non-differential attrition or refusals. So you've got a big problem. Given if you have a straight random sample, the number of events of interest will be very rare in the population of 5,500.
DR. IEZZONI: So you are going to have to oversample.
MS. RIMES: Right. We're not just talking about oversampling. You have to target or do what HIS had done, or screen the sample in an initial process.
DR. MOR: Like the disability survey, but you might want to use a more targeted question on the disability survey.
MS. RIMES: Yes.
DR. AMARO: Is this going to be done also in English only, or other languages?
MR. BRADLEY: I don't think we've talked about it. Do you recommend Spanish?
DR. AMARO: Well, I think you need to look at the issue of language, because you are in a mental population that might have a very different profile. The answer to all these questions that you had earlier might be different for those different groups.
DR. IEZZONI: And geographic issues I think are going to be very important, because people's ability to work may be very dependent upon what kind of employment is available. If you are living in a place where most people are laborers on a farm, you are going to have very different attitudes about whether you can work, as opposed to if you are in a city where there are a variety of different job types that you could get.
So this is going to be a tremendously challenging project to do well.
DR. MOR: There is very substantial geographic state level variation in the disability rates precisely because of the politics of the contracts, and the disability assignment process, both in its interrelationship to state disability policies, as well as federal disability policies. Which means that you have got big sampling issues related to the state overlay.
DR. IEZZONI: So Howard, I feel as if we are being very professorial in your direction, appropriately so, but I am a little bit concerned. At what point are decisions going to be made about what your sample is going to be? What your sampling strategy is going to be? Who is going to be making those decisions? When is this going to happen?
MR. BRADLEY: Well, we are aware of these issues. We are attempting right now to hire a project officer for this survey. I can't address them. I'm not in a position to address them. I don't have the background to do that.
The Institute of Medicine has a committee to review the research related to the disability process redesign. They included in that DES, so they have just come out with a report raising some of the same issues.
DR. IEZZONI: Is there any way we can help you on this? I think it's obvious that a number of us around the table have some serious concerns. Given that you are at the very earliest level, I know you have the Institute of Medicine looking over your shoulder, but we are another body that could be potentially useful.
MR. BRADLEY: I think the best way you, with your time constraints that we heard about this morning, would be to send me your comments. I'll give you my card and you can send it by e-mail or whatever, and I will feed them into the process. That's all I can suggest.
DR. IEZZONI: Okay, well, we will think about that.
DR. AMARO: Can I just add, is there any more detailed information about the study, purpose and design, et cetera, that we could respond to? That might help us.
MR. BRADLEY: I'll get that for you. We have published as sources, a request for industry review about a year and a half ago. We put in the Federal Register, plans related to the redesign, to include a synopsis of the DES. I'll give you the dates of those.
Let me just go on and talk about the other aims of DES. The first bullet on the page you just have seen, again we have the statutory definition of disability. Under the redesign we committed to research on changing the disability decision process, with greater emphasis on direct evaluation of functions related to work.
So the other major purpose of the DES will be provide a national sample of adults for testing of functional tests when the DES goes in the field in a few years from now. Long range we expect to have functional tests to administer for the purpose of testing it for our disability claims process.
The third purpose, and that is in the middle of the page, the SSA has proposed legislation and there are other bills coming up; two last year that were introduced or re-introduced or will be re-introduced this year on basically testing claimants in our process with vocational assessments at the time of in-take.
We expect probably in the next year's congressional session for there to be some kind of amalgam or compromise. There are going to be a lot of negotiations on how SSA would offer job placement services to job claimants or beneficiaries. SSA has a proposal which involves offering Medicaid coverage to working beneficiaries. You will probably hear more about that. Obviously, the DES could provide data on what enables or what people use to work in today's economy.
Finally, the last purpose, which is the last bullet on your first page, is the linking of DES, so that we don't have to do this every five years. It is likely we will have to go with linking of the data with other surveys, other administrative data that we already collect via claims of beneficiaries to monitor the levels of disability in the future.
Now as promised, one of the things that I do at SSA is I'm on a work group within the agency, and we are basically conducting research on the disability process. I talked to the issue of the redesign. How do we make our determinations better, more accurately? I have summarized basically some of the projects from 1995 and 1996, up to the present.
The last bullet contains the Institute of Medicine interim report on the DES, which came out about a month and a half ago.
The next handout is just an overview description of the IOM contract to review the research involved in the disability decision process.
As promised, I will have another handout with our definition of disability.
DR. IEZZONI: That's great. Is there anything more?
Are there any pressing questions? Otherwise, I would like to move on. Please go ahead.
DR. RUSSELL: I was asked here today to help out Gerry Hendershot, who was supposed to be here to talk about the disability software. I will be giving some introductory background information on the National Health Interview Survey Disability Supplement.
Please feel free to slow me down, stop me and ask questions if there is any need for clarification on any issues.
Just a brief introduction. It is a household survey. We take all families that are within the household. Usually there is one family. We only focus on the civilian, non-institutionalized population, so we don't do nursing homes, and we don't go into prisons.
It is the nationally representative probability sample. This particular supplement was conducted in two years, 1994 and 1995. This big, massive questionnaire here is just phase one. It was a lot.
The great thing about this questionnaire, the supplement is the wide range of issues that are dealt with, and we will cover those in the next slide. The other idea behind the supplement too is that there is no one definition of disability in the supplement. It is left up to the analyst to figure that out. Put all different kinds of questions together and make up your own definition.
In terms of a little bit of background on it, cost-wise, who knows? Twenty million more or less, who knows? I don't know how much, but that's a rough estimate.
MS. ADLER: Thirteen point seven.
DR. RUSSELL: Plus the two bases of the data, and all the time it takes to clean up the data and get it out to the public. So there is a lot to it.
It is amazing that 18 federal agencies and one private foundation were able to get together and coordinate this, and get all their ideas together into one survey. That kind of gives you the breadth or the complexity of this.
The Census Bureau conducted the field work for us, two years, two phases, and we will talk about the two phases. We released the data on an IBM mainframe CD-ROM.
Now let's talk about just phase one. It's a base interview within the household. This big questionnaire was administered after the core questionnaire that the HIS fields every year, which is the same questions every year. All family members -- the whole sample was asked all the questions. We did allow for proxies, for instance children were not allowed to respond for themselves. The most knowledgeable adult answered for the child. That is the HIS always conducts their interviews.
It is possible that one adult answered for all family members, if that is all that was home at the time of the interview. We also allowed field representatives from Census to finish up the interview; if they couldn't finish it all in one setting, to call them back on the phone, and finish it up on the phone.
The interesting thing about the phase one questionnaire is that it was the major screener for phase two, for how we selected respondents for our disability follow back survey, which is the phase two.
Well, I can't list all the hundreds of questions in phase one, but I can give you a broad topic of what kind of questions you might find -- sensory, communication, mobility impairment. The list is lengthy.
We do focus quite a bit on children; their special needs, details and education.
Two questions towards the end of phase one survey I find quite interesting was the questions of do you yourself perceive yourself to be disabled? Number two is, do you think other people perceive you to be disabled? Those are very good sociological questions. Then we have a whole litany of demographic questions. That was phase one, how we measured and what content was.
Let's talk about phase two and some background on that. We refer to the phase two as the disability follow back survey, DFS. In order to have received this questionnaire, you would have had to have answered certain questions on the core questionnaire, certain questions on the phase one questionnaire, and in another supplement called the interasset supplement. For various reasons, we picked certain questions. Depending on how you responded to those, you were picked for phase two. So not the whole sample got phase two.
Phase two was also face-to-face. We interviewed the selected respondent based on their answers previously given, and again, proxies for kids. Sometimes if we picked an adult that wasn't home, or we couldn't find them, we would ask another eligible adult about them, so we did proxies for them too.
Phase two is a little bit more complex in that there are four different types of questionnaires for phase two, and it really breaks down by age. So there is a questionnaire just for kids, one for adults, and one for older people, and then for polio. In general, if you want to know the content of the phase two questionnaire, it was more specific and refined questions on other topics.
In general these were the topics that we covered in the phase two questions. So we asked more about serious impediments in transportation.
Now that we all know the content and how it was administered, don't we all just want to go analyze the data? So let's talk about this 1994 data. Well, 1994 phase one data is out.
We captured 107,000 people, and our response rate was 87 percent, a little bit less than our core, because by the time you go through the core questionnaire and you get to this, did have some people drop out. In general, there really was no bias for the people that were dropping out. There is no one category of people who are always dropping out all the time. The data was released in July 1996.
The question on all of your minds is, oh, please more data. When can I have more data? Well, phase two we suspect will come out in the winter of 1998.
MS. ADLER: Does that mean like 1997 and 1998 winter, or the following year?
DR. RUSSELL: Well, if you really want to pin me down on a date, the reason I said winter is because the data is always changing, so it could be December-January- February, somewhere in there.
DR. IEZZONI: But it's not December 1998, it's more like January 1998?
DR. RUSSELL: Right. So it's closer to January- February 1998.
The 1995 data estimated date of release -- I'm not going to promise any dates, because these are always changeable, but we think that October 1997, phase one of the 1995 data will be out, and phase two more spring of 1998.
DR. MOR: Will the 1994-1995 have a direct linkage so that they can be added, or can you just add to those weights?
DR. RUSSELL: Good question. We have two questions. In terms of linking, it is just a matter of merging. They are not the same people, so you don't have to link people. So it's just a matter of merging data by your weights by two if you want an estimate of one year.
The other tricky part of your question is this survey straddles two sampling design meters. Our sampling design changed in 1995. So we have yet to work out the process whereby --
DR. MOR: The adjudicated weights.
DR. RUSSELL: Yes. In terms of those who used to get again, the variance estimations, that is going to be tricky too, because of the sampling design changes. That is still to be worked out. We are well aware of the problem.
Other questions before I deluge you with some data? You're all tired by now, so I'm safe.
Earlier I mentioned that there is no one definition of disability. This little graph -- what we did was we took the SIP definition of disability and severe disability, looked at their definition, and then went back to similar questions on the HIS and tried to recreate that with the data, which you can do fairly well.
If you look at John McNeil's publication for 1991- 92 SIP data, it kind of follows the same trends. We are kind of picking up the same type of thing by age. So that's one way you could do it.
If you were interested in modeling the ADA's definition of disability, there are plenty of questions in there to do that too. You could probably do a pretty good job.
Other ways of maybe looking at disability would be maybe estimates of well, how many people have an ADL problem? How many people had an IADL problem or a functional limitation? How many people perceive themselves as disabled? How did they see other people perceiving themselves? So that's another way of doing it.
One last little data slide here is we have assisted devices information too. Hopefully, in the next couple of months we will be having a new publication coming out at NCHS looking at assisted device usage.
DR. MOR: Vision?
DR. RUSSELL: Any type of device used for a vision impairment.
DR. MOR: Not glasses?
DR. RUSSELL: No, different from glasses.
Another strength of this survey is since HIS fielded so many supplements in 1994, and you could use the phase one 1994 data, and link it up by person to all the other supplements that we have too. So if you are interested in some year 2000 objectives, and you want to link it back to some disability issues, you could do that. All of these are supplements that we fielded in 1994, and they are all linkable by person. You can do it.
One important one that I would point out is the income and asset supplement. That has a lot of employment information. It has SSI, SSDI information, which also may be very important. That is not released yet, but it soon will be, because it has completed being imputed at the University of Michigan in terms of income. That is the only thing we imputed.
DR. MOR: Is the health insurance module out yet?
DR. RUSSELL: Yes.
DR. MOR: Just as an aside, how did you link them?
DR. RUSSELL: Each person in the data file has a unique number, and that's how you link them up. We do not have Social Security number on our data file. It is one of the highest non-response items in our questionnaire.
In terms of what you can do analytically, these are things that I just came up with. You may have other areas of interest. You could analyze it by age groups if you wanted to, or you could take a look at just specific maybe policy groups. That's not an official word, it's just a way of looking at certain groups of people and how you may want to assess it.
If you wanted to look at SSDI recipients or applicants in any particular way, you can do that too. If you want to just look at people that are receiving some kind of disability pension, you can do it that way too. So there is plenty of material combined here in things to do.
In terms of the data being out, having only been out for about a year yet, it is still very rich in what can be done with it. There is more data to come.
It is a complex survey in terms of just its magnitude, so I would encourage you, if you have further questions for clarification, or you want some documentation of the actual data, the questionnaire, anything of that nature, I would encourage you to contact me. I would be happy to help you out.
DR. IEZZONI: Are there any quick questions for Dr. Russell?
MS. WARD: Why polio?
MR. RUSSELL: Well, I think to give you a political answer, I think someone had some money in the right places, and they wanted to do that, but Michelle can probably answer that better.
MS. ADLER: It was to look at the effects of post- polio syndrome. The sample size and age was big enough to come up with people like that. So in phase one there were many questions to identify people who had ever had polio, and the questionnaire for people who had polio is very good. You should probably look at it. It includes not only medical aspects, but social and rehabilitative aspects of polio.
MS. WARD: I was just thinking as a condition it has been eliminated.
DR. IEZZONI: It's the post-polio syndrome.
Any other questions? Great, thank you, Dr. Russell. The clarity of your presentation was great. It was very helpful to hear about the NHIS since we heard about it in its infancy or its prenatal state too. I congratulate you on that.
Gerry, the Medicare Current Beneficiary Survey?
MR. ADLER: Sure. I'll try and be as quick as possible, as everybody else I know has a lot to cover.
I want to talk the survey the Health Care Financing Administration conducts called the Medicare Current Beneficiary Survey, the MCBS. Medicare is one of our two programs, the other being Medicaid. We don't yet have a survey of Medicaid beneficiaries.
Since 1991 we have had an ongoing survey of Medicare beneficiaries, a random sample. First, I'll tell you about it, and then I will show you some of the questions and answers from it.
The MCBS originated in the Actuary's Office of HCFA. The idea came from the need to estimate the sources and the objects of health care spending. We have always had claims data available for analysis, but we have not had the kind of detail, nor the information that isn't reflected in the claims about what health care people utilize, and how they pay for it. The Actuary is also frequently to model the consequences of alternative program policies, and thus needs a comprehensive source of information on a sample basis.
As I said, we began interviewing in 1991. Since then MCBS has been a multi-purpose resource for HCFA and the department and other health care researchers in government, foundations and universities. We have one of our customers here.
I just want to mention that since the recent reorganization of HCFA, you will find us in the Office of Strategic Planning, not in the Actuary's Office, but it doesn't matter a whole lot; we are still doing the same thing.
The MCBS is a continuous survey of a nationally representative sample of Medicare beneficiaries, sampled from HCFA's enrollment files. The sample size ranges from 12,000 to 16,000, depending on the round.
The MCBS is longitudinal in that sample persons are interviewed every four months over a substantial period of time. The periodicity is because we want to collect an ongoing record of the services they use, their coverage, and the payments.
The interview is conducted in the household or the long-term care facility in person, using computer assisted personal interviewing or CAPI. I do not have a copy of the questionnaire here, because it is in the computer. It is a computer program which automatically guides the interviewer through the questions.
The questionnaire focuses on financial issues, insurance coverage, the utilization of services, a detailed account of charges and who pays, plus characteristics of the respondent, his or her situation, health status and functioning, access and satisfaction, and a long list of personal characteristics.
It also permits the insertion of supplementary sections on timely topics. All we have do to is get OMB clearance. It's simple. We have developed a process for incorporating supplements on a routine basis.
The data also, when they come back from the survey, are linked to the respondents' Medicare claims files, so we have overlapping and linked sources of information.
The MCBS sample represents all Medicare beneficiaries, both aged and disabled, and is over and under 65, whether they reside in the community or in long-term care facilities, whether they are in fee-for-service or managed care. It is a single sample. There are no dual samples for household and nursing home. We sample from the rolls, stratifying by age, and sampling in PSUs. That's it; wherever the respondent is, we interview them with the appropriate instrument.
In sampling we use a strata, under 45, 45-64, five year intervals for 65 to 84, and then 85 and over. We oversample for the under 65 and the over 85, so that they are adequately represented in the survey.
In the first four month round -- I call each four months a round here -- the sample of 14,530 people yielded 12,674 completed interviews or 87 percent. Once you get people past the first round of interviewing, the response rate goes up, because who drop out, drop out in the first contact first round. So later rounds are in the 97-98 percent range of response rate.
Every year we augment the sample to bring in new enrollees and replenish sample cells that are depleted by refusals and death.
In the last few years we began rotating the sample after about four years. That is, we retire a portion of the sample each year, and replace them with new respondents, so that each respondent is in the survey about four years, to give us about a three year observation period, allowing for an initial interview to get the person into the sample, and the close out interviews at the end.
I won't go into the advantages of the rotating sample at this point unless you ask about it.
Data for the MCBS are collected through a contract with Westat, Incorporated. They have a permanent staff of about 200 interviewers around the country assigned to the survey.
As I say, the community interviews are conducted face-to-face using computer assistance personal interviewing or CAPI. Since the current round, we are also using CAPI for the long-term care facility interviews. The MCBS is one of the first major surveys to use CAPI extensively. It makes the interview easier to administer. It makes the data more accurate, and makes it faster.
As of now, the MCBS has done about 300,000 interviews in CAPI. A lot of surveys have gone to CAPI since we first did. They think they invented it, but I believe we were the pioneers.
Proxy respondents may answer the questionnaire for the respondent, but they must be designated; they are usually a family member. The interview typically lasts an hour, but there is considerable variation.
The topics of the interview I will run through quickly. I mentioned some of them -- demographic information, household composition, insurance coverage are asked at the beginning, and updated every round. Annually we ask about health status and functioning, access to medical care, satisfaction.
Each round we collect utilization and payment data by service and type -- dental, hospital, emergency room, inpatient hospital, institutional use, home health medical providers, other services, and prescribed medicines. So we get very complete information; we are HCFA after all. We get a full accounting in each of these areas of the charges and payments from all sources.
DR. IEZZONI: Even for your managed care patients?
MR. ADLER: Yes, we ask everybody. The problem with managed care patients is they don't have any good claims. So we are doing some pilot work to compare the fee- for-service and managed care. It is a tough area, and it is going to get to be more of a tough problem, because more and more people are going into managed care. Without the interview data, we have nothing at all.
So much for the descriptions. Any questions on that?
I have part of the health status functioning questionnaire here.
DR. AMARO: So the data utilization and cost are gathered from the respondents?
MR. ADLER: Yes. They are compared with the same spouse that are not in the survey or vice versa. We make some accommodation between the two sources.
DR. AMARO: I know if you ask me, if I'm not paying it, I don't remember, if my insurance is covering it. So I'm just wondering how reliable and accurate that information is.
MR. ADLER: It's possible to use amalgamated data on utilization. Again, with regard to the managed care, that is still an open issue of how well people can report their utilization in managed care.
Does everybody have a copy of this? I just want to run through it, because it I think it's much more fun to look at questionnaires and data than talk about survey design, even though both are necessary. Our initial question in the health status and function questions is fairly standard. I would like to ask you about your health in general compared to other people your age. Would you say that your health is: excellent, very good, good, fair or poor? The only difference there is not all surveys ask, compared to other people your age.
If you look at the next page, there is tabulation by age of the responses to the questionnaire question. It is remarkable steady by age. Roughly the same percent of people regard themselves, compared to their own age group, as excellent, very good, good, fair or poor.
Now there is one other factor beside the compared to your own age question that may be influencing these data.
DR. IEZZONI: Gerry, you don't have the younger than 65 on this page. They might have a slightly different attitude about that.
MR. ADLER: They do. Most of these tables here are about the over 65, and about those living in the community. The other factor is that people with very poor health status who go into the nursing home may drop off of this chart more and more as they age, so there may be a couple of other things working here.
I was frankly amazed that either people are reading the questionnaire, or listening to the questions and answering them precisely, or else something I don't understand is going on. We'll compare that with some objective indicators of health status in a little bit.
The next page asks questions about cataracts and hearing aids, difficulty eating solid foods. I left that in just to show the continuity of the question on health status, as with the preventive care measures -- mammogram, pap smear, hysterectomy and flu shot on page 3.
The first function status questions are the non-G items starting at the bottom of page 4. How much difficulty, if any, do you have stooping, crouching or kneeling, continues with lifting or carrying objects as heavy as 10 pounds, reaching or extending arms above the shoulder level, writing or handling or grasping small objects, and walking a quarter mile.
The next five pages have the results of those items by age group for the age of living in the community. It is a somewhat different picture than the general health status question in that there is a decline in the ability to stoop, crouch or kneel with age.
There are some interesting differences between the different functions. That is, if you look at the next page, more people have no difficulty lifting an object as heavy as 10 pounds than in the stooping, crouching or kneeling. More people have no difficulty reaching or extending arms above shoulder level, handling or grasping small objects. Then when you get back to walking a quarter of a mile, there is increasing difficulty. So you have to look at the differences between the items that are on the scales.
Has a doctor ever told you, you have -- I left that in just as an example of what conditions we ask about, and what results we get. This chart shows both the aged and the disabled. What amazed me about it was that the disabled, that is, those under 65 have about the same rates of these different conditions as the aged. It's not quite intuitive. They have a lot more psychiatric disorder and mental retardation, but when you get to pulmonary disease, diabetes, arthritis, they are right up there.
DR. CARTER-POKRAS: How do the estimates compare with other data sources. We were just looking at the hypertension. Has a doctor ever told you that you had hypertension, and that seemed really high.
MR. ADLER: Which data source did you compare it with?
DR. CARTER-POKRAS: I was thinking like NHANES.
MR. ADLER: Yes.
DR. CARTER-POKRAS: I ask because it's a self- report here.
MR. ADLER: Okay, this was over 65.
DR. MOR: It should be around 34-35 percent.
DR. CARTER-POKRAS: For 65 years of age.
DR. MOR: For 65, yes.
DR. CARTER-POKRAS: Well, this is high.
DR. IEZZONI: Yes, that's high. So these could be people who controlled it. It is self-report.
MR. ADLER: We'll have to look at self-report versus proxy report.
The next page says the questions about ADLs and IADLs. These are fairly typical questions, comparable to other surveys. Because of a health or physical problem, do you have any difficulty bathing or showering, dressing, eating, getting out of bed or chairs, walking?
Then we ask about is this because of the health, that there are problems? Do you receive help? Does someone usually stay in the room? Do you use special equipment? The IADLs are also fairly difficult on survey questions.
The following page shows the percent of elderly in the community who do report difficulties with ADLs, and which ADLs. Then the one after that is the percent who report difficulty with IADL items.
Finally, this computer program here is how we define our scale of ADL. If someone has difficulty -- if to the difficulty item on the ADLs they answered I don't do that, if they receive help or use equipment, then they get a score for that ADL, and we add those ADLs together. Similarly, if they have a problem within the high ADL, or don't do because of health, they get a score for the high ADL.
The next to the last page shows a table of results for ADLs and IADLs for that scale that is defined by that SAS(?) variable. What is interesting to me is that persistence of function, that is the percent non-disabled in the 65-69 is about 75 percent, and the percent in the 85 and over group is still about 31 percent. If you look at people who have no ADLs, that is combining non-disabled with IADL only in the over 85 group, you have still go about 55 percent who do not have ADLs.
DR. MOR: This is the community population?
MR. ADLER: Right.
DR. MOR: Would the overall aggregate look much different?
MR. ADLER: Yes, you may have the spill off into institutions of people with the most disability, but I'm looking at the people with the least disability at this point.
The final chart is just converting the data to a table, into a spatial diagram. It is the same information, function status by age. You can see the shrinkage of the non-disabled population or the proportion of the population doesn't disappear.
I have one more thing for you. This is for your later perusal. This is an excerpt from a volume that Westat has prepared, and will prepare on an annual basis called "The Health and Health Care of the Medicare Population."
DR. IEZZONI: It is nice.
MR. ADLER: It wasn't as slick. I think if we sold it, it would have to cost about $20 or $30.
This is just the introductory chapter giving the highlights and findings, and the technical appendix, which describes much more thoroughly the methodology of the survey.
DR. IEZZONI: Gerry, thank you. That was a great presentation. I think we have had four wonderful presentations, actually five that have showed us the diversity of the way people think about disability. Frankly, I'm not sure how to deal with this, because there is just a tremendous amount of diversity.
I think that we might like to try to get out early, but I would like to hear from Paul and Michelle, your final thoughts, because you have been in the trenches of this for many years, and have a lot of experience dealing with the diversity of definitions within the federal government and its implications.
Can you just give us some summary thoughts that you have about this issue, and how our committee might be able to help think about it?
MS. ADLER: I guess it was five or six years ago when I came to the Subcommittee on Mental Health Statistics and Long-Term Care. At that time the Social Security rolls were beginning to take off. The ADA had been enacted. We were getting a lot of rumblings on SSI for kids and we had basically no information on disability. The only thing we had was the 1978 survey on disability in work by Social Security, plus some elderly surveys here and there.
At that time, the committee was very helpful in supporting the notion of the disability supplement, and also the committee took a very active role in developing the mental health questions on the disability survey. That was one of the big steps. We basically had to start from the ground up.
The entire committee was very helpful during those early years. I would quite like to thank you now, because that was just a tremendous effort, and I don't know if we could have gone forward without the committee's help.
We are now at a different point. We do have disability data. It may not all be in on the CD-ROM, but we do have it. It has been collected. At least the consciousness raising of disability is now to the point where we have got a disability exam study is becoming real, the MCBS has functioning questions, SIP has much better functioning questions than they had in the past, and we have more square inches on the decennial census.
Now we have all this data. What should we do with it? Particularly if you look at children, one of the best things about the disability survey is it is really the first and only comprehensive survey of disability in children ever attempted in the U.S. Particularly for children under five, the whole concept of disability, what does it mean if a three year old needs help with whatever it is they are supposed to do, because we are kind of unclear about what it is they are supposed to do.
There are many research issues that can come out of the disability survey in regard to children, and there are also a number of methodological issues. We went for the overkill approach. Someone suggested the question and we put it on basically, but there have got to be many overlaps between these questions.
There has got to be a core set of questions on children's disability, and also on the working age and the elderly that we need to look into. We need to do it in a stricter, more scientific way certainly than is being done in the year 2000 census. We can take this data and look at it logically and rationally, and come up with some recommendations on data elements that work, data elements that don't work, and approach this thing so that the 2010 census and all these other surveys that will happen in the next few years, we won't be scrambling quite so much.
The other comment I might just say that was brought up by the disability evaluation study, but what has happened in regard to the ADA and to the work accommodations, and particularly with computers is that the whole work environment is changing. What does it mean if you have a disability if you can work? It is good to do a physical exam. It is good to do a mental exam. It is good to examine the local economy for employment and look at the person's education and so on and so forth, but in our work with children in special education, they are now saying that anyone, regardless of their disability, can use a computer and I believe that, so that they can function.
My boss as of last Monday, the deputy assistant secretary for disability and long-term care is Bob Williams, and he has severe mobility impairments. He's in a scooter. He has the guide dog, and he has speaking impairments. He's difficult to understand, but he has a machine called a Liberator, and it is like a big keyboard. He goes like this and he presses keys, and it speaks for him.
He was able to get through graduate school and be a political appointee and do lots of very good work. So the assisted devices are taking off on their own, and the whole notion of work environment and work accommodation is rapidly changing while we sit here and argue about disability questions.
The other thing that you can help us with are these changes in disability. We were doing good a few years ago to have any information on disability. A point in time, even if it was a long time ago, if we had one point, we would have been so happy to have that information, but the MCBS is longitudinal. SIF is longitudinal.
Looking at changes in disability and separating out changes due to mode of interviewing, maybe some questionnaire and non-sampling and sampling errors of true changes due to rehabilitation or deterioration of condition or disability would be something that we would need a lot of help with in the future.
DR. PLACEK: My comment would be five years ago next month the National Committee helped the ICIDH with a boost with hearings on ICIDH. It turned out to be the largest gathering ever in the United States on that topic. About 25 people around the U.S. who had been using it one way or another, testified to the committee. So that was kind of a jumping off point for getting involved in the revision process.
So I hope the National Committee will keep one thing in mind. This kind of comes up today. Whenever you talk about disability, we have heard about the Medicare study and the SSA survey and the census questions and the HIS questions. What we have is a lot of different methods of assessing disability, and as a result we have a lot of non-comparable definitions.
It is likely that many of the data can't be compared, and so the programs can't talk to each other statistically. The same is true with the U.S. and the world. The U.S. can't talk to any other country in terms of disability definitions, because so many different definitions are used. What we have is a Tower of Babel.
So to me, my work in the ICIDH has reinforced that we need to look for the gold standard, and try and use that framework in framing up our disability questions and analysis. That's my parting shot.
DR. IEZZONI: That's very helpful. We should talk tomorrow about whether the committee that is looking at coding data and transmission standards might be able to join us in helping us think about how to code disability information.
DR. PLACEK: The key is standardized assessment tools, which are being developed with the new revisions.
DR. IEZZONI: Do any of the committee members have any final thoughts or questions?
MS. WARD: I have one. We're very focused now on the disabled. Is there a need, or is that being met through other -- are employers meeting any of these demands? What is the employer's response to the ADA? Because of this enormous technology, are schools responding by allowing children to bring computers?
I'm thinking of the response to the disabled. If we haven't done this particularly well, and we can't standardize this, I also don't want to go over and --
DR. PLACEK: ADA has got another definition of disability in law. There are no good data on it except complaints.
I started to say, I know the HISD as much as any other analyst, but the HISD has every definition of disability, because there were four or five major agencies, and they were all funding, so they all wanted to make sure this or that question was in it. So we collected them all, and in a sense it gets at something of everything.
You're asking an ADA question, and is there a survey designed around ADA? I think the answer is no.
MS. WARD: I'm also thinking of schools. If you are in health care, you are also dealing with children who have special health care needs, whose families are trying to get these kids in schools. You could call it the bad case anecdotes about yes, my child has been told by all the people who evaluated her, that if you can only have the lap top in the classroom, her problems would be solved, but I can't get my school to let her take it.
That's the other environment that we work in, is the disabled encountering the environments, and coming to somebody in the health care field and saying, we're not matching.
MS. ADLER: That is the essence of the debate around the 2000 census questions. The disability community wants to look at the environment, not just medical aspects. The disability survey was designed -- by the time the questionnaires were put together, that was four or five years ago, which is almost a generation.
In the phase two interview there is an extensive section written by David Piper on work accommodations and employment accommodations. It is very good for its time. There is also a really detailed section on assisted devices for adults and children. That will help particularly with employment. There is some information on the environment for children.
You are right, we do need to develop measures of the environment whenever we talk about disability. In the future, just having the questions sensory impairment and the ADLs is not going to mean a whole lot without the accompanying questions not only on the environment, but on the particular individual's accessibility to the environment.
DR. AMARO: In one of our major topics that we chose to deal with Medicaid managed care, I was wondering if there was any information available about the availability and actual use of all this technology that improves functioning and participation in the world by people across socio-economic status, geographic areas?
It's sort of theoretically available, but is it really available? How many people's lives are really being improved by all this technology? I think that would be a relevant question.
DR. MOR: The data are available on the HDIS, because you do have, because of the phase one supplement, in particular for 1994 -- it wasn't done in 1995 as well, the health insurance component -- which otherwise is not possible to get from the HIS what short-term insurance coverage is for the overall population, but you can for this 1994 group. That's 70,000 individuals?
MS. ADLER: One hundred and seven thousand.
DR. MOR: So for 107,000, some proportion of those people will be Medicaid recipients.
DR. AMARO: Medicaid managed care?
DR. MOR: No, it won't deal with Medicaid managed care, but the question you are asking is really what's the accessibility I think of these kinds of devices.
DR. AMARO: That, and what's the impact of managed care going to be, the accessibility.
MS. ADLER: Probably the easiest way to look at it is a phase one in the ADL questions, and I can't remember if we did it for the IADL, but to find out if someone can do the activity for example, bathing, by themselves, with special equipment, or they need to be reminded or supervised and all these combinations. Then how difficult is it with these various combinations, and what assisted devices are used. So you are going to measure for example, with bathing, how many use an assisted device for that.
In phase two, the ADL questions are repeated, and in different kinds of therapy and services, and the devices grouped together to find out how much did the person pay out of pocket. Find out if Medicaid paid for it. Are there any unmet needs. For example, if you didn't have it, did you need it? It's not direct. You have to kind of go around and about.
DR. IEZZONI: I think Hortensia's question is prospectively. Going forward from this point, are we going to have data systems that will allow us to monitor on an ongoing basis, whether persons in Medicaid managed care get the assisted technologies that they need? Especially as you described very eloquently, Michelle, that the assisted technologies now can do many, many things that even a year or two or three ago they couldn't do.
So that is something that we should be thinking about. This is just a reminder that we will be having that meeting the last week in September to try to finalize what we want to focus on in the Medicaid managed care project.
Any final questions? I really want to thank our speakers. This has been extremely informative. I think that there are a few things that I would like to follow-up on. I would like to make sure that we get the Federal Register notice when the census questions come out, because I think that we would like to be able to respond to those.
We should definitely at least look at the IOM findings about the disability evaluation study, and see whether we could add our voice in just confirmation in some of the issues that have been raised.
I think we will help Paul about whether Barbara Starfield's committee on data standards might want to think jointly about looking further into coding for functional status issues.
So this has been very helpful. Does anybody else have any final comments? Wisdom for us from the panelists? No.
Well, thank you very much for spending Monday after with us. Thank you.
[Whereupon the meeting was recessed at 4:45 p.m., to reconvene in full session the following day, Tuesday, September 8, 1997 at 9:00 a.m.]