Room 505A, Hubert Humphrey Building
200 Independence
Avenue, SW
Washington, DC 20201
Protocol for Discussion of the July 14-15, Meeting on the Pacific Insular Areas, Puerto Rico, and the Virgin Islands
Overview of PRO Analysis and IPRO's Role in Analyzing Race and Ethnicity Data
Discussion of OMB Standard on Race and Ethnicity for Vital Statistics
Committee Deliberations
DR. IEZZONI: Shall we get started?
MS. ARAKI: Several of the representatives, at least from the Pacific Insular Areas are people who are Chief Public Health Officers. Some are health statisticians and the like, but I would say that the one from Guam is at a senior level. Peter John Comaco is the chief. He is in charge of the whole public health there. Do you want to go through the list? You know everybody.
DR. IEZZONI: What we should try to do is make sure that we are told how to pronounce the names of these countries, that we get phonetic ways to pronounce the countries and, also, phonetic ways to pronounce names because I think that that will be an issue for us.
Do you know who is coming from Puerto Rico?
MS. TUREK: That I don't yet. I have been talking with the Congressional Office, and she said that somebody may come from their office, somebody from the Health Department and somebody from the Governor's Office, but she hasn't given me the name of the official representative yet. We are in contact almost daily.
She is waiting to hear from home, but they are very interested, and they are going to have a large number of people here. I just don't have the exact names yet.
She is going to try to let me know this week. We, also, of course, at the request of Puerto Rico let the rest of the congressional delegations know we were having the meeting and, also, let the embassies know in case they wanted to send somebody. At this point we haven't heard back from any of them.
Of the people that are coming only one is cabinet level, and that is the Minister of Health Pilau(?). We talked to our International Office here and basically what we were told is we should treat them with respect and to, you know, not put them in the back of the room and you know, to essentially give them a position of honor at the table. We would have done that anyway basically.
I think Roylinne in her role with the Pacific Health Association probably knows the people who are coming better than any of us and we can let her kind of fill you in on what their roles are.
DR. IEZZONI: Great. Roylinne, do you want to give us some assistance?
MS. WADA: Do you have the names of the people who are coming?
DR. IEZZONI: No, we do not.
MS. WADA: From Guam rather than the Director of Health we have the Chief Public Health Office. He is like second in command. He is not deputy to the Director, but he runs all of public health.
DR. IEZZONI: And will we be told whether we should address people as doctor, mister, Ms., Mrs.?
MS. WADA: I will tell you right now, and then you guys can --
DR. IEZZONI: And you will have to give us that information.
MS. WADA: Peter John Comacho, the Chief Public Health Office has a master's in public health. He is not a doctor. He is quite young, quite informal, but the statistician is actually under him, and he is quite well informed.
Massau Ouada is the Minister of Health from the Republic of Pilau, and he is cabinet level in correspondence to people at this level go out as the honorable --
MS. ARAKI: Right, and I guess you call him Minister Ouada?
MS. WADA: We will call him minister, and you will give us notes about that?
MS. RIME: Also, we can do the name tags, and we can have everything set up and in front of us.
MS. ARAKI: He was a former health planner. So, he is very cognizant of the issues.
MS. WADA: Yes, he is, but he will speak at a ministerial level which is very different from the time that he was with the Trust Territories, very different.
MS. ARAKI: Right, sort of like the way it was at the IOM meeting. He speaks for the government.
MS. WADA: He wears the mantle of the government. So, he will tend to be a little bit more formal. I don't think you have to go out of your way to treat him like a minister or anything of that sort. He wouldn't expect that, but he may want a minute to just say how pleased he is to be here on behalf of the President of the Republic of Pilau.
From the Marshall Islands, unfortunately, there is an individual coming, I think who is attached to a special project with the Health and Populations Asian Development Bank Loan Program that the Marshalls has. I have no idea who this person is. He is new to health information systems.
They have had kind of a little merry-go-round of people coming on, getting off, never the right fit, and so he is new to it. I don't know why there is nobody coming from the ministry level, but I just got e-mailed that this is the person coming.
Amato Elimori is a nurse by training. He is from the Federated States of Micronesia. He is their vital statistician. He has been trained on the job. So, he does not have formal in-school training but a lot of OJT.
MS. ARAKI: And he will supposedly because he is coming from the FSM, he is representing the national government and the four states that are part of it presumably.
MS. WADA: Presumably. From the Commonwealth of the Northern Mariana Islands is Magdalena Sublon or Maggie Sublon. She, also, is not a PhD or DRPH. So, she is Mrs. Sublon, and she is the vital statistician or the health information person, also, trained on the job, no formal training, and I don't know how she will react in front of this type of group.
She tends to be a very shy individual and not very vocal.
Falae Ouella is the representative from American Samoa, and he really works in the health planning unit. He really does not work with health data, and why he is coming I am not quite sure and have gotten no information.
MS. ARAKI: I think it is because he is Samoan.
MS. WADA: But he has worked with the health department for many years. I have met with him since 1984. So, those are your representatives at least from the Pacific islands.
DR. IEZZONI: Right, and we are still waiting to hear from Puerto Rico and the Virgin Islands.
MS. TUREK: Virgin Islands is Noreen Michael who is the chief epidemiologist, I believe, and I talked to her a couple of times on the phone, but I honestly don't know anything about what her role is, but I have seen her name on a lot of other correspondence about the state of health in the Virgin Islands and health care needs in the Virgin Islands.
So, my impression is she is knowledgeable. She certainly is well published. I think some of the stuff that you have gotten she has done.
DR. IEZZONI: You have done a wonderful job pulling together folks for us to hear from. So, I guess what we need is just the phonetic spellings, a little bit of background, you know, the sentences or two that you have told us, the correct word or whatever you call it, Mrs, Mr. or Dr., before the person's name, and then we will just be respectful as we would be anyway.
So, do any of the subcommittee members have any questions or concerns about this at this point?
Now, I just wanted to ask you a simple question that you don't need to elaborate on; are you happy with the people who are coming from our Federal Government? Were you able to get representatives from the agencies that you feel are at a level that they can speak authoritatively and interact well with these folks?
MS. TUREK: I am in particular happy about the people that are coming from the regions who have the most direct contact. We have Joe Izzer from San Francisco, and he actually has an office that is set up to deal with Pacific Insular Areas and we have the Deputy Health Administrator from New York, Robert Davidson.
I hope, and I am sure we will get somebody good from HRSA. We seem to have been playing a game of trying to find the right person.
MS. ARAKI: The CDC representative is Dennise Cougar(?) and she has often represented.
DR. AMARO: Hortensia, who is coming from SAMHSA?
MS. ARAKI: SAMHSA is Dr. Rouse, Beatrice Rouse
MS. TUREK: And we have Michael Montelpoli, if I pronounced that right from the Department of Energy, and they do have a program in the Marshall Islands on radiation. Richard Miller from the Department of Interior will be representing the Department of Interior, and he is from the Office of Insular Affairs, and Roylinne and, also, Rebecca Fower from Census will be here.
We have Maryelle Conda from the Administration on Children, Youth and Family, and as I said on the IOM we are still trying to find somebody because Dr. Lawrence couldn't make it. So, there are two people who are on the panel who are local, and we will go and ask if one of them can come.
DR. AMARO: Did we manage to get everyone to stay the entire time? I think there was some discussion.
MS. ARAKI: That was what the request has been in all the letters. So, we are assuming, oh, definitely I think the people that are coming from the insular areas, Puerto Rico and the Virgin Islands were planning to stay both days.
DR. IEZZONI: We need the federal representatives not to take off.
MS. ARAKI: We emphasized that the whole second day was a dialogue.
DR. AMARO: I was doing to suggest that we go out of our way to plan something for dinner because people are coming from abroad and there are usually traditions of doing that.
DR. IEZZONI: You are right. It would really help people feel welcome by doing that.
And I want to make sure that people don't get hung up getting into this building. I can speak personally about being hung up getting into this building. How are we going to make sure that that doesn't happen?
MS. ARAKI: Barbara, I guess we could probably make special effort to make sure that -- we will ask for relaxed security like we always do, and so, people will be able to sign in, but I guess we can give them special instructions, also, but our contractor has reserved a block of rooms at the State Inn which is right across from the Pan American Health Organization at the government rate, and it is four blocks away from the Foggy Bottom Metro.
So, I have given them all that information, and they are on the same line as this, but anyway they may, also, have their liaison offices embassy people here pick them up and take them Definitely Mesau will, just Mesau, probably not anybody else.
Anyway so they will all be staying, and you have gotten information from our contractors. So, we, also, encourage you guys to stay there, too. It is actually a nice setup. It is not just a room. It is a living room and a kitchen and everything. It is a suite.
June 22 is the last date you can call to make your reservation. It is called the State Inn.
PARTICIPANT: We got a letter from Lisboa about it, Hortensia.
MS. TUREK: I want to say that early on I ran into Roylinne and that made things really easy, and I want to thank her because in going through her and using her knowledge base it was a piece of cake.
DR. IEZZONI: I want to thank you as well.
MS. HITCHCOCK: Hortensia, do you think that since SAMHSA has no program in the insular area or the Virgin Islands or Puerto Rico to my knowledge would try to get somebody from either the OA or one of the centers where they do have block grants?
DR. AMARO: That is true. I think Beatrice would be good, in that she has data systems that might be helpful in thinking about you know for suggestions especially around times of recommendations on how those systems could be, but it might be good to have someone who is working with block grants. I think you are right.
PARTICIPANT: I just wanted to say that the initial letter for the representative from SAMHSA was sent to the Office of the Administrator and so the Deputy Administrator decided that Beatrice Rouse should represent SAMHSA, but it seems to me you are free to --
DR. AMARO: They may not have realized that it would be good to have somebody from the other office as well.
DR. IEZZONI: It sounds like she is a good person to have, but we might want to have somebody else in addition.
MR. TUREK: I think we need to go back to the Administrator though and ask permission because we asked for an official representative who could speak for the agency. So, anybody besides here would not have an official role without the Administrator's permission. So, I think you have to keep that in mind, Dale. I don't know how we handle that.
MS. ARAKI: But I think that the work groups that you set up initially, earlier we had some program staff who attended, remember? I think they are planning to come, and these are people who are actually writing the block grants.
DR. AMARO: Will there still be time to just write Dr. Chavez another note and say that we are delighted that Dr. Rouse is coming. We were thinking that because of the block grants that SAMHSA should really have two people at this meeting, the other person being someone who knows about block grants.
PARTICIPANT: Mary Han?
MS. TUREK: Perhaps we could call her and tell her that the committee has a special interest in learning about block grants and ask her if she could, also, bring somebody who could speak to that. I would almost rather do that than go back to the Administrator. Why don't you work it out and see if we can do that?
We should try to move on. Richard, you had --
MR. HARDING: Just a follow-up on Hortensia's mention about dinner that first night. Is that possible? Is there any way to have it be a kind of an occasion, maybe with --
MS. ARAKI: Are you talking about contract-wise?
MR. HARDING: I know it gets complicated, but is there any way to have it in a small room or a private room where we could ask the Secretary to come and say, "Hello," or Vice President Gore or somebody who would make it a little bit special for them coming all this way and helping with governmental issues? I am just asking. I don't know the ins and outs, but it would sure be nice.
MS. ARAKI: It would be nice, but I don't know how likely that would be. We would have to consider what we can do in a short notice. It is short notice. We did send letters, also, informing to the political offices, the delegation, the representatives as well as the embassies informing them and inviting them to send a representative to this meeting if they wanted to. So, we met that part of the protocol.
DR. IEZZONI: That is a good suggestion, Richard. So, think about it. See what is possible, and what would be appropriate.
MS. CARTER-PONGRAS: Off the top of my head there is a new Asian Pacific Islander initiative for the department and so, if we do try to secure the Secretary we may get Dennis Hiashi or someone else to attend. I don't know how hard we want to push for somebody that high level or do we just need a senior level.
MR. HARDING: I would like to have Vice President Gore or the Secretary.
MS. ARAKI: You have got more pull with the Vice President and with Tipper than we do. Richard, you are backing yourself into sending an invitation. Do you think she would host this at the Vice President's house?
DR. IEZZONI: All right. We will leave Richard to talk to Lynnette and Joan and Roylinne about that. So, why don't you talk to Richard.
Okay, that is great. I really want to thank all three of you for the wonderful job that you have done.
All the subcommittee members should have in their offices a packet of materials about 3 inches thick.
MS. ARAKI: Two sets. You had an earlier mailing, and then you recently should have gotten one. It was mailed Friday.
DR. IEZZONI: I haven't gotten it yet. So, we now have more than 3 inches of material that we can read to get background, and we are looking forward to this and thank you very, very much for all your hard work.
So, we will see you July 14, at 9 o'clock.
Okay, great. Why don't we in the interests of time move forward to our next item on the agenda and I see some of the subcommittee members looking at some of the material that I just had Barbara pass out.
I just thought that it was a good thing for you to get some articles that I have run across recently especially about the role of socioeconomic status in looking at mortality. I think that some of this will be very relevant as we talk about Healthy People 2010 tomorrow, and then something from the AM News about Medicaid managed care and something about what I will continue to call post-acute or what did I call it? Do they call it subacute care? Barbara might not be comfortable with that appellation either. That one we have to deal with.
Harriet, do you want to talk from there then?
MS. STARR: The only thing is I do have overheads.
DR. IEZZONI: All right. We need to have you miked if you are going to talk from overheads because we do transcribe these meetings for our minutes. So, maybe you could be miked right now, and you could just speak from there.
Harriet Starr is gracious enough to come talk to us from IPRO which is the peer review organization in New York which has done some studies of Medicare data for race and ethnicity reporting.
I will let Harriet tell us exactly what she has done.
MS. STARR: Good afternoon, everybody. As Dr. Iezzoni said, I am Harriet Starr from IPRO, and I am Vice President for Project Development and Evaluation in our Quality Improvement Program.
What you heard Dr. Elstein speak about this morning was our HEDIS audit program which is part of IPRO but not the part I am going to address this morning. I am really going to address some analyses that we have done related to race data and some of our quality improvement studies and Ms. Rimes, also, asked me to give you an overview about IPRO, some of the things that we are and some of the things that we do.
So, I am just going to do that briefly, and as I said I did bring some overheads here. Basically I don't know if you can see that. Anyway IPRO is an independent company. We are a not-for-profit organization, and we are committed to assessing and approving the value of health care services received by consumers through the use of innovative methods and technologies.
Just a little bit about our history. We were actually incorporated in 1983, as a result of the merger of the PSRO program for those of you who go back all that far. So we are actually quite an old organization.
Our headquarters are in Lake Success, New York. We, also, have offices in Albany and in Vermont, Burlington, Vermont. Our annual revenues are approximately $25 million.
We are ranked amongst the nation's top health care review firms, that is in terms of the volume of review activity that we conduct, and we have a staff of over 250 which includes physicians, epidemiologists, data analysts, programmers, systems analysts, coders, etc.
As an independent company we do work both in the public and private sector. In the public sector our large customer is HCFA. We are the peer review organization for the State of New York, and for those of you who are not familiar with the peer review program, basically the mission is to assure that the health care received by Medicare beneficiaries is appropriate at the right level of care and at professionally recognized standards.
The PRO scope of work has changed considerably over time from one that was based largely in case review to now one that looks at patterns of care, and the patterns of care is something I am going to talk about in terms of our quality improvement projects a little later on.
We, also, do work for state government. We serve as the Medicaid managed care external quality review organization for New York, for Rhode Island and New Mexico as well.
For New York we, also, do a lot of fee-for-service utilization management work, and we do, also, work for AIDS patients in terms of review of AIDS designated units in New York State.
In Vermont we, also, do fee-for-service Medicaid utilization review work as well, and in the private sector our work is with some corporations of labor unions, third-party administrators and so forth.
This is basically a list of our core services. We are very heavily in the data area. We are involved with doing quality measurement and improvement projects which is what I am going to address a little later on.
We do, as you heard this morning HEDIS validation. We, also, on the commercial side contract with some of the managed care organizations to do validation and auditing of their commercial HEDIS data apart from the Medicare data.
We do external quality review, utilization management as sort of the traditional case review and hospital bill audits for private customers.
I thought I would mention just a little bit about our major data sources. We do get the claims database routinely from HCFA as part of our Medicare contract.
For Medicare it is basically Part A data which we work with most extensively. We, also, get Part B data on request, and we have supplemental databases, a beneficiary file which interestingly on the issue of race may very well be different from the claims data because apparently there is no cross check between those.
So, we have the beneficiary file which would give us information that is used for population-based analyses that we do, but we, also, have other information that comes from the claim, of course, and we have a physician file as well.
That is from Medicare. Since we do extensive Medicaid work we, also, have very similar in New York State Medicaid claims files, mostly inpatient, some outpatient data.
New York State, also, has an all-payer database called SPARKS which we request and get, and that gives us information on all payers.
So, we have quite extensive data to work from, and we use that data to do routine sort of pattern analysis, sample selection for health care quality improvement activities in our routine review activities.
We do have, and we are building as we speak clinical data from abstraction of medical records which we use in our health care quality improvement projects.
In addition we do have more traditional review results like on a case-by-case basis was the care appropriate; was the length of stay appropriate and so forth, and managed care plans supply data to us for things like HEDIS validation and counts of data validation and certain focused clinical studies that we do in the managed care arena.
I am just going to say a few words about what the health care quality improvement program is, describe a little bit how that process works, and then I am going to go into the projects that I am discussing today.
The health care quality improvement program is part of the PRO, that is the Medicare scope of work. It was implemented in 1994, basically with the goal of improving the processes and outcomes of care for Medicare beneficiaries.
It was a movement away from case-by-case review where the intent is to identify aberrant providers and instead look at patterns of care with a kind of CQI underlying philosophy which is that improvement of processes will result in more widespread and through voluntary cooperation longer lasting improvements for a larger proportion of the Medicare beneficiaries than will identifying aberrant providers on a one-by-one basis.
Okay, just a little bit about our methodology so you can understand where this comes from. Basically we develop ideas through the literature, through established guidelines like those from AHCPR, through speaking with the physician community but for our Medicare projects we pick ideas that are widespread and prevalent amongst Medicare beneficiaries and then have an established evidence base behind them.
We don't really invent new guidelines in the Medicare program, and basically what we will do is we will bring in a large, as large as we can study group of usually physicians and possibly other stakeholders in the process to advise us on the issues involved in this.
We will develop the tool. We will develop what the indicators that we are looking for are and a tool to collect those indicators, and I heard some discussion this morning about the issues of reliability, and we are very concerned about that.
We do extensive training of the abstractors on reliability, interrater reliability analysis as well.
After data are collected and analyzed, fed back to the providers we work intensively with them to share best practices and to try to help them improve their processes as well as doing some education.
So, the improvement activities are both oriented toward the process side like hospitals, implementing systems, for example, to assure appropriate medication subscribing and we, also, do education of individual physicians.
Some of our improvement projects, in fact, the three that I am going to talk about in a little bit more detail where we looked at race actually did not follow this model. They did not deal with data abstracted from medical records but rather with claims data, but I am going to get to that shortly. We do both national and local quality improvement projects.
Some of the members of the committee might be familiar with HCFA's cooperative cardiovascular project which has been going on now since 1994 in which all the PROs around the country have focused on care for acute myocardial infarction in terms of nine quality indicators.
In New York State we recently remeasured our results and did find significant improvement, and that has been a national project.
This I won't go into a lot of detail on this, but this is just to give you an idea. This is Page 1 of an idea of some of the local quality improvement projects, the scope of the kinds of things that we are involved with in IPRO, everything from H. pylori to prostate pathology, to community acquired pneumonia, pressure ulcer prevention and so forth.
We are doing quality improvement projects in all of these areas.
This is Page 2 of the list, and our projects I am going to speak about today are cardiac catheterization, radical prostatectomy and carotid endarterectomy where we have seen very dramatic influences on race.
Okay, radical prostatectomy was a quality improvement project and actually we did collect some clinical information on the scope and dimensions and pathology of the tumors, but in conjunction with that we kind of looked at comparison of race of prostate cancer and radical prostatectomy amongst blacks and whites in our Medicare population.
Basically the idea behind this project was that there has been a growing consensus that radical prostatectomy is generally not appropriate for males over age 75 and rarely appropriate for ages 70 to 74 because of competing mortality in that age group and the presence of a lot of comorbidities which make that surgery generally not indicated.
This was only a caveat of this particular project which was that we only used the inpatient claims and so the diagnosis of prostate cancer certainly can be made as an outpatient basis, and when the treatment is watchful waiting, for example, that wasn't included in what we looked at in this instance.
DR. STARFIELD: What was the denominator?
MS. STARR: The denominator we use is just the beneficiary population and in fact these data, this was one of our early projects; these are early data from 1991 to 1993, and so it doesn't reflect actually the growing identification of prostate cancer through PSA screening that is becoming more and more prevalent.
Okay, I hope that all of you can see these data and basically what these are are rates of both prostate cancer and radical prostatectomy per 100,000 male Medicare beneficiaries.
Basically there are a couple of things I just want to point out about this.
DR. STARFIELD: Again, this is just inpatient data, right?
MS. STARR: This is just inpatient data. That is correct.
A couple of things I did want to point out about the data, one is that the New York State rate for radical prostatectomy per 100,000 male Medicare beneficiaries in 1993 was 127, and although as you can see the rate had been increasing, still that was rather low compared to areas on the West Coast where in California I think it was double. I think the rate was three twenty something. I don't have it in front of me.
So, although it was low, one of the reasons why we undertook this project was because the rate was increasing. Actually the rate of increase in radical prostatectomies was not completely made up for by the rate of increase in prostate cancer, now, again, with the caveat that we are only talking about inpatient data here the rate of increase for prostate cancer was smaller than for radical prostatectomy.
So, the rate was growing although it was overall low. The thing that I really wanted to point out here is that there are significant differences between whites and blacks.
Every year the rate of prostate cancer is higher amongst the black population, and yet the rate of radical prostatectomy is lower, and those differences are at every year, and although the rate of increase in radical prostatectomies amongst blacks was higher than it was for whites, and actually I have the rates elsewhere, but it is still, the increase didn't catch up, so to speak in the time period that we looked at.
So, that is just one indication.
This is the breakdown by age, and these numbers I apologize might be even a little harder to see, but there are just a couple of things to point out here. Basically the same patterns hold for different age groups as they did for overall.
There is a couple of exceptions. One is that for the under 65 groups the numbers are not very stable because there is a very small number of actually blacks in the under 65 although this is Medicare data. The Medicare disabled data tend to be the people under 65. So, it is not the stable Medicare population.
The 65-to-69-year-old age group is the age group where this procedure is most likely to be indicated and, also, the age group where the rates are largest, and what you can see is that there are dramatic differences between whites and blacks, so that 1993, the rate of radical prostatectomy was 282 as opposed to 210, and again the same patterns of higher rates for prostate cancer, lower rates for radical prostatectomy can be seen at all age groups with the exception of the under 65.
Okay, carotid endarterectomy -- did you have a question?
DR. FRIEDMAN: Just a couple of minor questions that would help me understand. First of all, the race data, is that observational or self-reported?
MS. STARR: It is claims data, and in other words it comes from the hospital bill, and hospitals fill that out in a variety of ways. Sometimes it is in the admission office. I guess it generally would be.
The problem is that it is unstable and unreliable in general because when we go to the medical record, for example, in this instance when we go back to the medical record to do clinical abstraction we find the patient's race can be three different things within the medical record which is different from the claim which is different from the beneficiary file, and so, this comes from the individual claim of that record.
There is, also, no validation between claims. So, if the same patient is readmitted for something else they could possibly be coded with a different race code.
DR. FRIEDMAN: The other very quickly, the incidence data for prostate cancer is based on?
MS. STARR: That is, also, based on inpatient data. We just looked at those who had a diagnosis of prostate cancer.
DR. FRIEDMAN: The admission with --
MS. STARR: Yes.
DR. STARFIELD: That mustn't mean that the rate of some other procedure is higher for blacks because they were all hospitalized, and presumably they are hospitalized to have something done for the prostate cancer. Is that right?
MS. STARR: The prostate cancer could have been a comorbidity that they were admitted with. They didn't necessarily have to be admitted for the prostate cancer.
DR. STARFIELD: So, you really just don't know what you are dealing with.
MS. STARR: In that respect, right.
Carotid endarterectomy was not a procedure that is necessarily widespread in the hospital Medicare community. We did a what we call hospital report, comparison report on this which is just really a technical distinction. It is just rather than an improvement project, it is just data that we develop internally at IPRO and send out to the hospitals in this case.
We didn't break down hospital-specific data because the volumes were too small but just another way to help the hospitals have some information to improve on, but basically we looked at this because it can be effective.
Carotid endarterectomy as a procedure can be very effective in reducing mortality from stroke, but it can, also, bring on stroke, so that there are high risks in dealing with surgery as well as high benefits, and we wanted to look at what we can find out about mortality in this case.
That was really the purpose of the study was to look at both the utilization and the 30-day mortality, and here we combined in most of our analyses 3 years of data, 1993 through 1995, also, inpatient, Medicare inpatient data and this of course, is solely an inpatient procedure.
Basically just to tell you a little bit about what we did before we get to the race data, we did a logistic regression model to predict the 30-day mortality based on demographics, age, sex and race, type of admission; actually emergency and urgent admission were significant variables as were certain patient comorbidities which were CHF atrial fibrillation; aortic and mitral valve disorders were, also, significant comorbidities, and then we used that model to calculate, adjust for 30-day mortality rates, basically because in this case the volumes were too small not by hospital but by hospital peer groups and certain patient characteristics.
This is actually not related to race but just to give you a sense of our overall findings here what we found was that there is a significant relationship between hospital volume and there we groups hospitals into volume groups and mortality.
As you would expect as is known about many procedures there is an inverse relationship, so that hospitals doing small volumes had significantly higher mortality rates.
DR. STARFIELD: This is 30-day mortality both in and out of hospital?
MS. STARR: Yes, that is correct. Because we do get a beneficiary file of Medicare it is updated when the beneficiary dies. So, we are able to tell.
This was the breakdown. I am sorry that looks a little fuzzy but basically the important point here is that when we looked at race, yes, race has a different pattern in terms of mortality for this procedure, but basically the important point here is the volumes are so astonishingly low that we didn't even go the step of figuring out the rates per 100,000 Medicare beneficiaries because basically you had about 250 cases in the 3 years amongst blacks which is only about 2 percent of all the cases that were done.
So, it is obviously disproportionate, and I will have some conclusions later on about why we think that might be happening, but basically we see very astonishingly low rates in the use of carotid endarterectomy in black population, Medicare population in New York.
DR. STARFIELD: I am sorry I have to interrupt you again. What is the denominator for that rate?
MS. STARR: The 2 percent you mean?
DR. STARFIELD: The low rate.
MS. STARR: Actually I am just talking about the fact that the black population is about 10 percent of the population and represents only about 2 percent of this procedure. So, that is just a general comparison.
DR. STARFIELD: So, it is really the low numbers of procedures from what you would expect.
MS. STARR: That is correct.
The last project I am going to talk about is cardiac catheterization. Here the reason for doing this particular project is that the value and necessity of doing a right heart catheterization in conjunction with left heart catheterization for coronary artery disease has been increasingly questioned over time and when we analyzed the data now again this was Medicare inpatient claims data we did find that hospitals had very significant patterns where some hospitals were apparently performing right heart catheterizations or bilateral actually sometimes routinely, 90 percent of the time, 70 percent of the time, 80 percent of the time, and some hospitals were not performing it at all, and again, this is for coronary artery disease, not for procedures where a right heart catheter would be appropriate.
We took those procedures out of our analysis, and basically we wanted to look then at the trends in the use of catheterization where an inpatient claim by the way is generated even if the procedure is done at an outpatient cath lab.
We do get a Part A claim on those and in New York State at the time that we did this project there were only 53 licensed catheterization labs that were able to do this in the state.
So, we wanted to look at the trends in the use of it as well as the breakdowns by demographics, and basically we started -- this was, also, one of our earlier projects. We started this in 1993 where we had 1992 data.
We defined our baseline period in this case as the first 6 months of 1992, and we started doing our impact analysis in 1995, when we had the 1994 data and have looked at five impact periods since then, starting with 6-month periods starting in January 1994, and the intervention that we did for this was, well, it was a very focused intervention because there were only as I said out of the nearly 300 hospitals in New York there were only 53 licensed cath labs at the time.
We did a lot of focused educational interventions with those and particularly the ones that seemed to have routine rates of bilateral catheterization.
Okay, here I am just rather than showing all the data for the sake of simplicity basically am just showing the baseline and last impact period that we looked at which was the first 6 months of 1996.
Basically all of these differences are significant at the .01 level. The increase in, now again, the denominator here is just the number of Medicare beneficiaries during this period.
We had 293 catheterizations per 100,000 Medicare beneficiaries in the first 6 months of 1992, and in the first 6 months of 1997, that rate had increased to 372, whereas the corresponding rate of bilateral catheterizations went down from 89 to 65 per 100,000, and those trends, however, although they are replicated in the racial breakdowns the differences between the blacks and whites in this case were also significant so that whites were receiving overall more catheterizations per 100,000 beneficiaries but, also, higher rates of bilateral although they had declined similarly.
Okay, so by way of conclusion I am trying to indicate that we are doing a lot of quality improvement projects. These three that I picked out are the ones that we have either published some information on or we actually have some publications in process on these racial breakdowns which I thought the committee would be interested in, and basically we do see as has been shown in many other instances that there are different rates of usage in the use of these procedures amongst the Medicare population.
There is a couple of possible explanations, the obvious one perhaps being that there is differential access to both diagnosis and treatment, and part of this possibly there are financial considerations involved insofar as socioeconomic status overlaps with the use of Part B premiums, deductibles and so forth.
For radical prostatectomy, also, the literature supports that at the stage of diagnosis blacks are diagnosed at a later stage of prostate cancer, and so, they may be more inoperable cases at that stage which may explain some of the lower rates for radical prostatectomy, and in addition we know that in New York State, and I guess I am not sure of the numbers in other areas of the country but the expected life expectancy amongst blacks in New York State is 63 years old for males.
That was a year or two ago, and so many of them are not reaching the Medicare age population, and we see therefore lower, both lower volumes and lower rates that may be reflecting that very odd age distribution, and so, you know, part of our problem in looking at this factor is that the numbers, the volumes do get very small when we break it down both by age groups and for specific procedures, and so, I will answer any other questions you might have about this.
MS. WARD: Could you say, and you may have already said this, but I have lost it how many hospitals were included, how many different hospitals were included?
MS. STARR: For the cardiac catheterization there were 53 licensed catheterization labs. They are all connected with hospitals. They are not free standing in New York.
For the other projects we have close to 300 hospitals where we just used all the data.
MS. WARD: Did you find any differences among facilities? Is there anything that looks like the differences are facility related?
MS. STARR: Very definitely, we do look at hospital, interhospital differences all the time, and that is part of the data that we feed back to the hospitals as part of the improvement project.
MS. WARD: So, it wasn't as though the differences were -- did the differences exist across the board but were then worse in others or were there some places where in fact differences didn't occur?
MS. STARR: Are you talking about the racial differences? I really cannot answer that because the racial analysis that we did was really an adjunct to the main purpose of this study which was why I went into what the purpose of each of the studies was, and so, the idea of improving access really wasn't the focus of our study as much as it was improving care for the particular clinical condition.
DR. STARFIELD: You are doing some really interesting work, and it is nice to know that, and you are confronted with some data limitations of course that you have to deal with, but I guess what I get from the work you have presented is the need to do some more hard thinking about what you are using as denominators to try to combine inpatient and outpatient more than you have been able to do to take into account comorbidity.
MS. STARR: In the carotid endarterectomy we did do a specific analysis of the comorbidities that are involved. Cardiac catheterization we did in the sense that these are patients that were being treated for coronary artery disease.
A lot of the conditions that were indicative of right heart catheterization that would have been indicative for right heart catheterization were removed from the analysis.
I apologize if I presented this a little too briefly. We do take a lot of the comorbidities into effect in terms of either adjusting for them or removing them from our analysis.
DR. STARFIELD: But the big problem was with the prostatectomy one. You couldn't tell what people were hospitalized for.
MS. STARR: In that case right. That was one of our early studies where we were just looking at the volumes basically of prostatectomies in that case.
DR. IEZZONI: I think what is of concern to me is what you have talked about, the quality of the race coding. So, if there is so much variability when you go back to medical records to validate the coding of race on the Part A claims file, then your attribution of various observations to different racial categories is suspect.
MS. STARR: Right. That is one of the reasons why we haven't done a lot of work with this.
DR. IEZZONI: And so can you tell us a little bit more about the quality of the race coding? Are there any numbers that you can give us about when you go back to medical records and look at them what is reported on the claim?
MS. STARR: We really have not done that analysis to validate our medical record abstraction against the claim. It is more really just sort of anecdotal evidence where our abstracters are telling us that when they go back to the medical record they see that this is the information you have said is on the claim, and they see this is incorrect, and that does put a lot of caution to our use of that code, but we haven't gone back to do an analysis of what those numbers look like for any one of these projects.
DR. AMARO: I was wondering if, I mean the issue you are mentioning is very common. We have heard it everywhere we go about the inaccuracies of reporting, especially when it is an observer reporting or medical record and I was wondering if in the work you do going back to the providers you do any work around improving the quality of how they report race and ethnicity and if you have any lessons you have learned from that?
MS. STARR: In part of our case review program we do DRG validation, but the race code as you probably are aware is not really part of the DRG.
So, I guess the answer probably in terms of that specific code is no. In terms of overall improvement of the coding of the diagnostic and procedural information that would be relevant although you know in terms of the change of the scope of work in the Medicare program away from a lot of individual case review we are doing less of that over time, less DRG validation, but specifically to improve the coding of race we have not.
DR. FRIEDMAN: Just to expand on Hortensia's comments I would think that any observational race or ethnicity data needs to be regarded as uniformly suspect at a minimum.
MS. CARTER-PONGRAS: We were just a little puzzled over here. We are members of an internal working group on racial and ethnic data, and we were puzzled. You are talking about Medicare data?
MS. STARR: Yes, I am, right now.
MS. CARTER-PONGRAS: How come the encounter forms weren't linked to the enrollment or eligibility determination? As I understand you don't have the same kind of problems with the observed racial and ethnic identification on that because it is self-identification if they filled out the form.
MS. STARR: You are talking about the managed care data?
MS. CARTER-PONGRAS: I am talking about Medicare. You were talking about Medicare?
MS. STARR: Yes, right, but I am not talking about managed care.
MS. CATER-PONGRAS: You are talking about Medicare data?
DR. IEZZONI: Right. Those kinds of studies typically do not go back and link the individual Part A claims files with the original eligibility files from Medicare. They typically don't do that. They don't have access to those data, frankly. The IPRO wouldn't have access to that.
MS. CARTER-PONGRAS: I am just wondering. Mary Gornick had, also, done some studies on the Medicare population.
DR. IEZZONI: She is at central office. You can do that from there.
MS. CARTER-PONGRAS: So, it is a problem of access then, access to the original?
MS. RIME: It is a problem with several things. Let us take a frightful example using the Medicare beneficiary survey which is a self-reporting experience that uses or fails to use the revised OMB directive, but in the old race reporting and ethnicity people at least had the opportunity to report their own race and ethnicity, and now the variation from that in what I call the administrative files or the claims files was fairly significant. It ranged about 50 percent difference on how people actually reported themselves versus what someone in a Medicare office or someone in a hospital or someone who chose not to even fill out the blank, and the encounter data aren't really readily available.
I think you may be mixing the future anticipation versus the existing. The encounter data are based on some times assuming they ever see it, but I am not convinced --
MS. STARR: I guess one other comment I could make about that is that even the source document of the hospitalization, apart from the claim, the medical record itself contains contradictory information about race, and so, I am not sure if we would know what was the correct one even going back to the medical record, even comparing the claim with the beneficiary file, in addition to which we don't have any authority to make any corrections of either the claim in that regard unless we found data from the medical record, but going back comparing the beneficiary file with the claim file would be kind of interesting for us, and we would like probably to do that, but I think that it doesn't have any immediate action for us associated with it.
MS. RIME: There is a whole research question anytime anybody uses administrative files, whether they are HCFA claim files or any other files versus survey files or anything else.
I spent some time going through the research on administrative files and what is there and what isn't and how shaky some of these data are. Perhaps I come from a Public Health Service viewpoint and ended up in HCFA and to them the administrative files are gold standard, and they really may not be.
DR. IEZZONI: Okay, thank you. We can talk about that a little more if you would like to learn because I can tell you how hard that would be to hook up with the original files. They just don't have access.
Harriet Starr, thank you very, very much for your presentation, and we have had a day of IPRO starting with the HEDIS quality measures earlier.
Now, we have an interesting panel which we will probably be hearing from for the next several years, frankly, while you work out the problems that are a result of the changes in the OMB directive on standards on race and ethnicity.
So, as the panelists are coming up to the speaker's table, I would like to hand around a written report or written comments from Harry Rosenberg who decided that what he would like to do is talk to us about the implications of the standards, the change in the standards around age reporting for coming up with mortality rates.
He preferred to give us his comments in writing. So, committee members, may I hand these out to you and then I will hand them out in this direction, and if there are leftovers we can send them around the room?
DR. STARFIELD: Lisa, while we are doing that I just wanted to point out that what was handed out this morning on the EU rules and regulations, as you know, they are now prohibiting any transmission of any data on race and ethnicity.
DR. IEZZONI: Oh, interesting. Okay.
Why don't we start? Katherine Wallman, do you want to take us off, and I am just going to excuse myself for a second. I will be right back. So, why don't you start?
MS. WALLMAN: You want me to start?
DR. IEZZONI: Yes, go ahead and start.
MS. WALLMAN: Thank you. I feel like I have seen you all before because I am talking to so many folks over the last 3 or 4 years and as was said we probably will be talking to all of you for all the years that we can foresee on this subject.
Some of you may have noticed a piece in the New York Times over the weekend from our paul Steve Holmes, and if some of you didn't see it, I have got a copy here that can be copied for people.
Steve's article was really written based principally on some meetings of the Census Advisory Committees over the last couple of weeks where the knotty question of once you have made a certain decision where does that take you was certainly in front of people, and I think that is in front of this group as well.
I did look back at the minutes of one of the subcommittee meetings, and I keep getting mixed up about which are your committees and subcommittees, and I should have that figured out after all these years, but I understood I get 1 day's grace on this because you guys have actually changed your subcommittee structure, and I learned that in a phone conversation the other day.
So, I will pretend that it is all because you changed your structure, but that wouldn't be true. Nonetheless, I was looking over the weekend at some minutes from a subcommittee meeting that was held back in January when I think I was not in the room at the time, but some comments were made by one of my colleagues who has been working on this issue about some of the issues, the difficulties, the problems and so on that we all face now, and I want to use that as a reason I guess to go back for just a minute about why we made the decision that was made in terms of collecting data because I have a feeling that that may get periodically lost in the discussions as they have now turned to the tabulation issues.
So, let me just reflect that as most of you know because we have met before that we have gone through a process that has now spanned more than 4 years trying to decide what we should do about our standards for the collection of data on race and ethnicity in the face of changes in the demographic make-up of our population that were brought about by a variety of factors and the two most often cited having to do with increasing intermarriage of people of different backgrounds and increasing immigration from particular areas, and the net result of that has been an increasing number of people in the population who have heritages that are of multiple origins.
Moreover, probably a small detail that gets lost occasionally is the fact that regardless of what we are asking for on federal forms and other reporting forms people are telling us that they are of multiple heritages, and I think we anticipate that this is a trend that will continue.
Now, I know in the case of the vital records there has been an open-ended question where people had that opportunity to report multiple heritages although only one of those then was carried forward in the federal uses of the information, at least as I understand it.
Please correct me where I am wrong because this is one of many, many cases of the uses of these data. So, I could well have my own data wrong at this point.
In any event there was as you know a lengthy set of work that went on including some considerable research that was undertaken about some of the options that people were presenting to us, and there are a number of people in the room including my pal, Clyde Tucker on my left who were involved in leading this activity, and we didn't let Clyde off the hook. He is now carrying through on some of these guidelines for tabulation work with us.
That being said, a decision was reached as you know at the end of October to accept the recommendations that our interagency committee made to us, and those recommendations were based on research. They were based on input from a variety of efforts to get to the public. They were based on work by a 30-agency group across government and probably other aspects that I am not thinking about right now, and the decision was made to accept the interagency's recommendation, the most dramatic recommendation that when reporting to the Federal Government respondents should be permitted if they wish to do so to report more than one race, and that is the principal shift. That is clearly the one that is getting all the attention.
Attendant to that are issues about how do you tabulate the information once you have this much richer, if you will set of information about people.
Moreover there are questions about what are the best ways to actually collect the information given that that decision has been made and there are a number of issues about how will one as I have come to say, build bridges to the past for working with the data collected under the new standards and looking back trying to do some things over a period of years, looking back at some of the data that were collected under previous standards.
Let me note that this is not unique to the collection of data on race and ethnicity. When I was talking with Steve Holmes the other day for some further background I reminded him of the case of industry classification where the woes that we are hearing right now are remarkably similar. When people throw their hands up, if you can envision that and say, "What are we going to do? The continuity with the past is lost." It is just a different set of folks who are saying the same thing in Business Week and the other set of magazines that do that kind of economic reporting.
They are saying basically, and what they said 2 years ago was, "What is the matter with those people in Washington? Don't they get it? The economy has changed. It is a service sector economy. How come they are collecting 80 gazillion categories of data on manufacturing industries and they have one industry that includes all of computing software and everything else that has happened in the last 10 or 20 years?
So, we got it. We sort of looked out the window and went through a major revision of the industry classification which has now been released. It has being implemented, and so what are the magazine saying now? They say, "Oh, my God, what are we going to do? We are not going to have any comparability with what was collected under this admittedly very outdated system for collecting data on industry."
I say that not to make light of the issue but rather to say that it is something that is not unique to the collection of data on race and ethnicity, just as some of the problems that we are talking about having now, and I was listening a little bit to the conversation that took place just before this panel about the vagaries in reporting of data on race and ethnicity that don't arise because on October 30, 1997, OMB issued a revised standard for how we collect this information.
Many, many of the concerns, not all but many of the concerns that we are hearing have to do with the collection of the data in the first instance, regardless of whether we collect it under an old standard or a new standard, and again, not to make light of any of the very real problems that we all face together in this, but I am just trying to put a little bit of perspective into the discussion.
What I think the expectation for today was to kind of continue to bring you up to date in where we are in aspects of this process, and we have indeed a panel who I believe are going to amongst them play three sort of different roles.
Clyde Tucker I have already introduced as the gentleman who has been leading a lot of the research work related initially to the decisions on standards for the collection of information and has now got the probably totally thankless job of leading off the guidelines effort as an overall initiative.
I am going to interrupt myself for just a second. I think it is important to recall that when we talked about the guidelines for the presentation of the information that we actually are talking about guidelines and not a single standard, not a one-size-fits-all kind of solution to things, not a regulation.
Those words mean different things to different people, but I think we have had this conversation a few places where people have perceived that there is going to be a single standard, and so, I want to get rid of the word "single" and I want to get rid of the standard and say that what we are talking about is issuing guidelines for ways that data might be presented, and there may be a few different kinds of guidelines depending upon the kinds of information that we are talking about.
That is an open question at this point. I don't have a conclusion for you one way or the other. Back to where I was, Clyde is going to talk a little bit about where we are in the guidelines process and the work that is going on.
Is that right?
MR. TUCKER: Right.
MS. WALLMAN: Nancy Gordon, my colleague to my right in red from the Census Bureau is going to talk in particular about where the Census Bureau is headed since that is the first big and very important implementation of the new standards in the 2000 census, and of course the denominator for a lot of the work that you folks do in the health area.
So, that was that. Jennifer is, I think going to talk about some of the, I don't know. I was told by a third party that Jennifer was going to talk somewhat about some of the knotty longitudinal kinds of issues that I have now set you up for so beautifully.
So, Clyde, and what is your overall time, and I think you want a chance to ask questions and stuff like that?
DR. IEZZONI: Yes, one-half hour until about four-thirty because we would like Ed Sondik, if you can stick around so we can talk at four-thirty about the workshop? Okay, great.
So, arrange yourselves and so maybe we can have some time for questions and maybe each of you could take 5 or 7 minutes or so. Does that sound reasonable? Clyde looks a little bit like he will do whatever we tell him to do.
MR. TUCKER: Let me say first of all that in dealing with the tabulation issues we have to be concerned not only with what comes out of the census but, also, with what comes out of national surveys, what comes out of vital statistics, what comes out of crime rates, all those kinds of things, how all of those fit into this.
So, it is not simply even though the census right now is getting the biggest play; it includes all of us, including everybody in this room.
In the October 30, notice we said in there that there were two types of users, those that have legislative mandates to carry out including bill of rights legislation and things like that, and there is also concern for those who are monitoring economic and social trends in the nation over time and that includes many of the statistical agencies and so, in the notice it was pointed out that for many purposes, particularly for Justice Department they might want to look at the full distribution of all of the possible combinations which turns out to be 31 if you don't have another category and 63, if you do, at different geographic levels, and then there is, also, the possibility of looking at what we called the all-inclusive which was to assign everyone to every race that they mentioned which gives full representation of the answers that we are given but adds up to more than 100 percent which is a complication, and then there was some discussion of the fact that these users could, also, collapse the categories using their best judgment for the particular uses they had for carrying out these mandates, but there was a whole class of users as I said before that we didn't really address that much, and that was the monitoring social and economic trends.
So, we have been starting to work on some issues on that, one being the bridge to the past as Kathy mentioned and trying to come up with different ways of considering how we might be able to produce something like that with the notion that what we were trying to do is to be able for our analysts to separate out what would have been true changes over time versus what has been methodologically induced change by changing the standard.
There is, also, concern about the ongoing collection of statistics where the denominator is the census figures like vital statistics and also crime rates, and also concern for our large national surveys in terms of the way controls are used for producing the final estimates from those surveys.
We put together some different groups. We put together a group that is working specifically on the bridge task. We put together a group that is working procedural design which is to look at different ways of us asking questions under different situations like whether you are doing it on telephone, whether you are doing self-administered, personal interviews, by mail, whatever.
Also, we are working in that group to look at how we would go about collecting aggregate data, such as school enrollment data.
Then we have a group that is looking at where we might be going in the future in terms of how we are looking at ongoing distributions. We have gotten to a certain point in society where we know we have become diverse and leading toward being more diverse still. So, we need to be considering where we think we are going to be going in terms of pursuing statistics over the long run and over the next 10 years between now and the next census.
If you looked at this article in the New York Times you know that we looked at a variety of different methods for tabulation, including not only full tabulation of all possible categories and combinations but, also, the all inclusive and, also, looking at a combining approach which uses information about people that are minorities and why you might get coded for instance into a minority classification based upon previous history, and then we have, also, some others we call an algorithmic approach which looks at different types of statistical means for producing a bridge to the past.
All of these are being considered. We are looking at them. We are currently in the middle of not only collecting information from different people about how they would fill out different ways of asking the question under different modes, but we are, also, working on statistical analyses of the data that we currently have and doing simulation research in order to try to determine which of these methods might be most appropriate under which conditions.
MS. GORDON: All right. The Census Bureau has a different set of problems than everybody else because we collect data from so many people. We have a great deal more information to tabulate and as Kathy mentioned it is used for a number of different purposes.
There is a handout coming around, and I could pretend that I am regrouping here but actually it is just in the wrong order. So, look at Page 2 first. Truth in advertising. We will bring you back to Page 1. So, what can I say?
Page 2 is the two questions on ethnicity and then race that are xeroxed from the form that is being used for the dress rehearsal which is going on now. The places where you see the little vertical lines, those are actually boxes on the real form, so that people can fill in specific extra detail.
So, you will see that there is a large number of boxes which individuals can mark which aggregate up to the five major categories that OMB identified plus one called some other race which the Census Bureau has an exemption to add for Census 2000.
There are three write-in lines. The middle write-in line is shared for other Asian categories that are not shown above and other Pacific Islander categories not shown discretely above.
So, that gives us 15 boxes and three write-in lines which is an awful lot of information. We are going to present information in a variety of different ways. I will summarize a few of them now, but the important thing to remember is that when someone reports whatever they report, that is the data that is kept, and so, researchers or interested parties will have the ability to specify their own tabulations in whatever way they want to that would not violate a respondent's confidentiality by providing back that information to the requester, and that is a whole other talk on how that is being arranged, but the point is that the products that the Census Bureau prepares for dissemination are by no means the only way that the data can be arrayed.
One of our most important products is in response to Public Law 94-171 which is used in the enforcement of t he Voting Rights Act, and that is what is described briefly on the first page of the handout.
Clyde was joking about how many categories are there. For us we have these six major categories which generate 57 combinations of two or more for a total of 63. Then the data have to be shown separately for everyone in the population and the non-Hispanic population and each of those groups are broken separately by all ages and by those 18 and over.
So that is multiplied by four. Then you have 14 totals. You get up to 266 possible cells, and these data are presented for every block in the country.
So, you have this huge array but what it shows you is the counts of people. It doesn't show you any characteristics of them, just how many people are in these different categories.
There may be some revision to showing all of those different 266 categories for every block because there is work still ongoing with the Justice Department to sort out the balance between protecting confidentiality and meeting the requirements of that particular law.
Now, that humongous full distribution as some people sometimes refer to it is perhaps not useful for many analytic purposes. Hence the term "collapsed distributions," and if you look at the, I believe third page of the handout you will see on the left hand side, the bottom half a stub that is titled race, and in that we are showing one way in which one might want to collapse the data.
The reason that this is a mixture of verb tenses is that we have to present the data from the dress rehearsal in some ways but final decisions about presenting data from Census 2000 await the guidelines which are in the process of being developed.
So, these are ways that we can present the data that seemed useful that don't in any way violate the standard as it was issued last October.
The collapsing that is shown there just to make it more concrete shows in this particular case just the number of people who reported exactly one race.
If one were using a stub like this to present characteristics such as education or occupation then one might not be able to show all the subcategories, for example, that are shown here under Asian.
So, that first part of the table shows the people who reported exactly one race, and in this particular case everyone who reported two or more is collapsed into a particular group.
Depending on the level of geography that you are looking at with census data, variations are possible, and one thing that we have been considering is that under two or more races one might be able to show data for the groups that had the largest number reporting.
For example, in Sacramento there may be a sufficiently large number of people who mark both white and Asian to be able to show some characteristics of that group from the dress rehearsal data, but again, preserving the confidentiality of the respondents is paramount and so it really is going to be a data-driven decision about how much information we can present in tables.
The bottom part of that table shows what Clyde was describing as the all inclusive group which shows the number of people in this particular case not available yet who marked a particular racial category regardless of whether or not they marked any more.
So, in essence it is a table of the number of races that were reported. If we have anyone taking up the option to report more than one race that will, of course, add to more than the number of people who reported.
One of the reasons that I think that the standard was adopted certainly that was discussed a great deal beforehand had to do with health research and the fact that knowing about the background of individuals is important to understanding more about some diseases that are linked to one's racial heritage.
Now, if that data is going to be useful to a researcher you need to know much more than just two races were marked or two or more. You need to know the specific races that were marked, and there I think one encounters a very important constraint which is called preserving confidentiality of the respondent, and so, you may want to consider developing some mechanisms to allow researchers to have access to data which does show the specific races that were reported but under strictly controlled circumstances.
The Census Bureau has developed an organizational structure called research data centers in which we allow under very strictly controlled circumstances researchers to have access to data that are protected by Title 13. They have to be sworn in. They are subject to not only financial penalties but jail terms if they divulge any of these data and so forth and so on, and if at some point you decide you want to pursue an avenue like that, I would be happy to tell you more about the problems we have encountered and the ways that we are attempting to resolve them.
MS. MADANS: Before I talk a little bit about what we are doing in NCHS in activities that really overlap what everyone has said, I am glad Nancy mentioned the data center because we are currently negotiating with the Census Bureau. Actually they are coming tomorrow to look at the space. We will have a census research data center at NCHS, hopefully this summer or sometime in the fall, and we are going to do one of the on-site inspections. There will, also, be a department data center separate but close by to do similar kinds of things with HHS data that we cannot release, but we can provide access to, but that is a topic for another meeting.
We really are undertaking two separate but related kinds of activities, one primarily involving vital statistics, and that is to ensure that what we are collecting for numerators is going to match somehow with what census is collecting for denominators so that we can continue to maintain rates and have meaningful rates.
The other issue there is being able to supply the Census Bureau with data that they can incorporate into the census counts to do the intercensal(?) estimates. So, there are very close ties here between what can be collected as part of the census and what is being collected as part of the vital statistics system so that we may be able to get somewhere else in the year 2010 that we cannot get to in the year 2000.
It is a fortuitous time because we are looking at re-evaluating the birth and death certificates. That process is ongoing, and the change in the classification is being considered as part of the re-evaluation of the certificate process, and one of the major objectives is to make the numerators and denominators as comparable as possible and to move forward on that avenue.
We, also, are concerned about how we are collecting race and ethnicity on the provider-based surveys and so the group that Clyde talked about in terms of developing better data collection strategies for that set of numerator data is of course, important, but I think the focus on vital statistics because of its importance in the whole process is really paramount right now.
The other activity is to analyze what has turned out to be kind of a gold mine, and I don't think we realized it when we were collecting it as part of the HIS.
The HIS has actually been collecting multiple race since 1982. We have not been publishing it, but we have been collecting it, and that information is stored, and we have been able to analyze it, and that is one of the data sets, if not the main data set that has been used for some of these modeling processes that Clyde talked about.
What would it look like if you did one of a variety of different ways, and we can look at change over time although I must emphasize the percent on HIS reporting more than one race is still relatively small, and so the numbers are relatively small, and it is not easy to look at time trends. You have to combine multiple years, sometimes 8 or 9 years to actually get some stable numbers, but there will be a set of analyses that will be presented in a report just to explain what happened over the time using this data collection approach and, also, a separate set of reports we hope that will actually look at some health outcomes to see is there variability, whether you look at just the single race or the multiple races.
HIS also asks the question if someone does report multiple races to kind elicit whether there is one race or another that someone would relate to more strongly.
Most people will answer that question, and that is the piece of information that has been reported on the data files I think through 1994, but I could be wrong on that.
So, that additional piece of information I think will be very useful in this bridge to the past. Our concern as Kathy said is to be able to go through this bridge period very much as we do when we go through an ICD change as we are now where we can try to tabulate things both ways.
How do we think it would have been if we had asked the question the same way; how is it now and run for some of the major health indicators two sets of trend lines for several years and then pick up on the new line really depending on how quickly the society changes.
If we are going to stay at around 2 or 3 percent I think it doesn't make a whole lot of difference, but as that starts to take off we will start picking up a new set of tabulation specs but for a while we are probably going to have to work on both.
There are some interesting things coming out of this analysis. One is that there does seem to be a big age effect. Most of the multiracial are being reported among younger people and it will be interesting to see if we had some better longitudinal files whether that changes and that would be a very interesting area of research.
I hope that some of these reports coming out of NCHS, I don't know what the time line is for the committee reports but at least looking at some of the health effects might be available sometime in the fall or the winter.
DR. IEZZONI: Great. That was a very provocative and informative panel. Could I ask our subcommittee whether there are any comments or questions?
Dan?
DR. FRIEDMAN: Just a comment which is really directed more or less to Jennifer which is as you are trying to jibe the vitals data with the new census data it does seem to me that there is both a classification issue, but in addition to the classification issue there is the issue that we were talking about earlier which is the self-report versus observational, and it seems for example, if you want to live for a really long time in this country it is good to be Latino, and there is some question about the quality of mortality classification around race and ethnicity and there is no recommendation attached, but it seems as if this change in the nature of classification may even exacerbate that.
MS. MADANS: I couldn't agree with you more. I think that the easier-to-tackle problem is the birth certificates. I think we will be able to deal with that better. I have no confidence that anything will improve very soon on the mortality end, and I think you are right. This will make it worse. It is very unlikely you are going to pick up any kind of multiracial on anyone on the death certificate.
The only easy or relatively easy thing that we can think to do is to rely more heavily on matched files. The HIS now is routinely matched to the national death index, and so you can get some idea of what the misclassification is and look at some of the rates.
The CPS, that project is no longer ongoing. I don't think we are matching that any longer, but the HIS is large enough and as it keeps going we will have a large body of information to model as opposed to actually get estimates out of.
In infant mortality we only use the linked file really now to look at race differentials. I don't have a good idea for what to do for the mortality other than to continue to educate the folks who fill out those certificates and encourage them to ask that. Perhaps with the electronic record, and things are happening more in a different setting you might be able to improve the quality, but I think it is going to be an approach that tries to come from various directions, training, using alternate data sources, doing some modeling to try to get a better idea of what is going on.
It will not be straightforward.
DR. IEZZONI: Nancy, I wondered if you could just speak for a minute or two about some of the privacy issues that you kind of tantalized us with at the end that you are thinking about fines and various other punitive approaches and what are the concerns, and how are you thinking about making sure that privacy is protected for secondary users of the census data?
MS. GORDON: There is this thing called Title 13 which is the title of an act that governs the Census Bureau's protection of confidentiality of the data that we collect, and our obligation is not to divulge it, and so the concern is that if you are going beyond producing tables into producing micro data files there needs to be some mechanism to deal with the data so that people cannot be identified, and there are various techniques that we have used.
I mean the standard thing in micro data is to not describe very much geography. So, maybe you know the state the person lived in to top code the income amounts and so on. For the census itself that doesn't work terribly well because the whole point of the census, its greatest value is that it gives you information about small geographic areas.
So, then the question is how much detail can you provide about the people who reside in those small geographic areas without divulging information about them.
So, for example, if you live within an area that is not particularly diverse where there is someone who is observably black, and you look at the data set and you see that there is someone who is black and some other combination, white, you might then hypothesize that the person whom you know has this mixed racial heritage which was a piece of information you didn't have before. That would have been bad for us to do that.
We do deal in the census with some of this by swapping information across households so that even though you may think you found out something you may not have found it out, but again, there are limits to the extent that we can do that.
So, there are various things in which we protect the data that is released to the public, all of which is to say that the researchers did not like it. We take information from one household and information from another household, and we move it. We literally swap the information so that now you think you have identified this household which has four people in it, but in fact the information you are getting when you look up that household you think you have identified, it is really information reported by this other household because we took the records, and we literally swapped them.
MS. WALLMAN: Not on the original record but on what is released on the public use data.
MS. GORDON: On the public use file, right. Oh, I am sorry, I missed the point you were trying to get me to make.
MS. WALLMAN: That is why I was being so literal. I just don't want people walking away thinking that we are switching people's records.
MS. GORDON: No, the Census Bureau collects the stuff, and we keep it the way it was reported.
MS. WALLMAN: It is a little bit of jargon. That is all.
DR. IEZZONI: Do you swap it within geographic areas? I mean are the households both living in the same census tract?
MS. GORDON: We don't talk a helluva lot about what we do because if we did you could undo it. You can ask the question, it is just that we keep that a little quiet. We don't tell you how much we do it either. It is a choice between doing some things that people don't like and they especially don't like not knowing what we did and not being able to put out the data. So, there we go.
The research data centers allow researchers to get access to some data. We have put in a process with the National Science Foundation where universities primarily but it could be a government agency, a state agency or a non-profit organization submit a proposal to become a research data center. There is a competitive process. We are very gradually expanding the number of research data centers.
MS. WALLMAN: Is there anybody besides Nancy from Pittsburgh? The first live one of these at a non-government facility is up at Carnegie-Mellon. So, I thought if anybody was from Pittsburgh, they might know.
DR. IEZZONI: Are the data centers given non-swapped data.
MS. GORDON: Well, that is where I was going. Then the researcher wants to study a particular thing. They then have to submit a proposal, also, evaluated by NSF and the Census Bureau. The proposal has to be accepted. Not all proposals are accepted.
Among other things it has to be of benefit to the Census Bureau as well as to the greater understanding of phenomena in the world.
Then the researcher is give access to only the data necessary for that project. Sometimes what this is, it is a matter of we have a data that is a nationally representative sample, but the researcher wants to merge to that data information about specific neighborhoods.
That is fine but now we give that data back to the researcher where the researcher can now identify the neighborhood in which the sample cases reside. Right? So, that is only given back under the circumstances which is a secure site run by a Census Bureau employee; embedded in this particular example Carnegie-Mellon University we disabled the drive. So, you cannot copy the data on a floppy disk and think it would be so much more convenient to work at home.
We take very seriously the notion that data do not leave. The only thing you can take out are research results, and we strongly favor research proposals that are producing things like regression coefficients, not research proposals that are producing tabular results because again the confidentiality checking of tables is a very complicated process to get straight, but those are the real data.
MS. WALLMAN: In answer to the question, those are the real data although nobody has got the whole census, right? There is one project using decennial census data for one state. That is as much as has gone.
DR. IEZZONI: Hortensia, you had a question?
DR. AMARO: Yes, I had a question. I am thinking that all of this is going to be going on. I think, and maybe I am not right, that during the same time that the age adjustment changes are going on and which are going to produce some changes in estimates. So, you have got that causing a change in estimate, and you have this causing a change in estimate, and how are we going to keep it all straight? Tell me a little bit about how you are hoping to deal with that and how you see it laid out?
MS. MADANS: Actually the age adjustment goes into effect in 1999. It is the mortality, deaths in 1999, will change the age adjustment, and --
DR. AMARO: Data year 1999?
MS. MADANS: Data year, and we, the Census is not until 2000. So, there will be a year, but I don't think we are going to be able to disentangle a lot of this because there are certain parts of it that we will be able to because we can control it, but on the mortality side as Dan said, that is going to change in a very uncontrollable way, and we are not even sure it is worth doing for some of the race ethnic groups. So, to disentangle some of this is going to be very difficult.
MS. GORDON: Would you feel any better if you knew that occupational classification is going to change at the same time?
MS. MADANS: So, is ICD.
DR. IEZZONI: Ed Sondik had a question and then Barbara.
DR. SONDIK: No, just a point about the age adjustment. So, if one wanted to stay with the old one, with the old whatever standard you wanted, whether it was 1973 or whatever you could continue to do that but it is going to be quite a bouillabaisse.
DR. IEZZONI: Ed, why don't you come up to the table because we are going to want to have you here during the next discussion anyway, and one final question from Barbara?
DR. STARFIELD: I just wanted to make an observation of something that I think is going to really strain the system, for example, the state CHIP program which is 50 different programs because we have 50 different states where only certain areas are going to be targeted mostly because low-income kids live in certain areas. So, you are going to have to get down to a very small area where you are going to want to evaluate things, and everybody agrees that we are going to need population-based data to do that, not administrative data.
DR. IEZZONI: Nancy looks like she is creatively thinking about that and will probably have a solution for us, but thank you, panel for your forbearance in talking to us about this. We will probably have you back again to hear more in 6 months or so.
That is when, Jennifer, you said that you might have some of your first results?
MS. MADANS: I might have actual data, yes.
DR. IEZZONI: Okay, we will actually have data in 6 months. So, we should make sure to hear back.
Thank you very much.
MS. WALLMAN: Thank you for continuing to have us back.
DR. IEZZONI: We will.
MS. WALLMAN: And please keep your comments coming to us. We are still in that process in terms of as you can see, in terms of having some guidelines formulated. So, we welcome your cards, letters, thoughts.
DR. IEZZONI: Great, thanks, Kathy.
All right. I would like to move to something that is now happening because we are a fuller subcommittee than we were this morning when we first started and that is that we have Dan Friedman and Barbara's working group joining us that has been working with Ed Sondik on the workshop, I guess on the vision of health data needs for the 21st century.
So, this is the bridge going forward now that we have had the bridge going -- yes?
DR. STARFIELD: I just want to ask a question. We actually have a work plan for that working group, and I just want to know when you think we might deal with it?
DR. IEZZONI: That is tomorrow morning at 8 o'clock, basically reminding people that tomorrow morning at 8 o'clock we have a breakout session for our subcommittee during which we will try to figure out how we are going to do everything that is now on our plate and do it reasonably well.
Okay, so, Dan, do you want to lead this discussion then with Ed Sondik at the table?
DR. FRIEDMAN: Sure. I am not going to speak for Ed, obviously, but let me just say that we at Ed's invitation, we meaning Barbara and Paul Newachek and I have had a variety of discussions with Ed and Marjorie and Jerry Hendershot and their colleagues following up on a suggestion which Ed made last year about trying to institute a couple of workshops, one focusing on the vision for health statistics in the year 2005 and 2010, and the second the implications of that for NCHS, and as the discussions continued and evolved and became sometimes more confusing and sometimes more clear it really became evident to us that we needed to focus less on one or two workshops per se and much more on what were we trying to get out of the workshops, and what were we trying to essentially define as products of a process with workshops being part of that process, and what we did was put together a very, very draft work plan that focused on three or four goals, depending how we count the goals with the first one being developing a 21st century vision for health statistics, a second specifying that vision in terms of what role NCHS should play in implementing that, and then third, some specific, developing some specific operational steps for NCHS and the notion behind this I think was to really try to step back several steps and look at health statistics system and systems fairly generically because in public health surveillance there is a body of literature.
There are definitions for surveillance. There is discussion of what is surveillance and how we evaluate public health surveillance systems. There are operational definitions of surveillance and how we evaluate surveillance systems, but in fact there is really none of that for health statistics, and part of what we talked about in terms of developing a 21st century vision was trying to really among other things define health statistics in terms of its purpose, its roles and its boundaries through focus groups and then, also, trying to identify the criteria through which we could evaluate health statistics systems and in addition a process through which we could look at a variety of recommendations that have evolved in the past decade or so relating to health statistics and various health statistics systems and trying to assess where are we in implementing those recommendations.
So, why don't I stop and turn it over to Ed. I can either stop or continue with describing the work plan, whatever.
DR. IEZZONI: Why don't we hear from Ed?
DR. SONDIK: I think that was a very good summary. I want to go back just to make one point. In the current agenda book there is the paper on the national health information infrastructure, and I feel the need. I have said now a year ago that I have felt the need for NCHS to get advice because we have a lot of choices to make. We are very, very tight in budget.
By the way look at the health section today. There was an article on HANES today which talked about the budget and made it seem extremely tight. It is not quite that tight and, also, we didn't make significant cutbacks in it either.
I mean actually I think there has been over the last 6 months or so an increasing awareness of the issues or at least our issues in fact in terms of our budget, and I think there are more and more people aware of it. That doesn't mean that we have had any fiscal relief, but I think information is critical, and that article is just one of the steps, I think, and I was very pleased to see it come out.
We didn't plant it by any means, but it was very nice to see it come out.
In any case I think we have got a lot of choices to make, and we have received a lot. We, and health statistics in general have received a lot of advice. There have been a number of recommendations from CNSTAT, from this Committee over the last decade to maybe even two decades, and some of those recommendations I think are really out of date.
Some have not been acted on, but still others are out of date. They are out of date I think because of technology in particular and in some cases maybe they are out of date because some of the issues in health are going to deal with things that weren't even thought about 15 years ago.
I don't have time to have an extended discussion with this panel but it is interesting to think about health statistics and genetic technology and what that says about race and then couple that with SES and by the way Health US is going to have an SES chart book coming out in July, and I think that, also, is going to add to the mix, this bouillabaisse over here and I hope provide some clarity but I think, also, probably confusion as well, sort of what variables do you really need. I know what I think, but that is another story.
I think it is all going to get somewhat more confusing before it gets any better. Coming to that, I certainly think we need some guidance, and when we met, a small group met with people with CNSTAT and Don Detmer was there, and now I guess about 10 months ago or so, thereabouts, maybe 9 months or so, we felt that having these workshops and involving the health statistics community broadly would be very useful for us, and we are proceeding along the lines that Dan mentioned, and I feel that it is quite possible for us to have a workshop, a first workshop next January, and given El Nino it will be warm, no snow.
So, we have been discussing that, but we have, also, been having these discussions because I would like to see CNSTAT involved as well, and the sheet that I have handed out gives a kind of current conception at least from NCHS's point of view of what the different players in this might do toward this workshop, starting off here with the Committee on National Statistics and in particular one of their roles in this, and again, this is just totally draft here would be to either prepare or commission a synthesis of all of this advice that we have had over the past.
This committee, we felt could focus on the vision statement, a draft vision statement. Where should we be, say, in 2005 and 2010, health statistics in general, and this could be one of the principal, this vision statement plus commissioned papers that would look at other issues that would be critical in getting to that vision statement could be the focus of this workshop.
I think what I guess I feel is I would like a workshop that is as substantive and thoughtful as possible. In order to make it thoughtful and substantive, I think we have to go in with something for people to react to.
I think otherwise we get into a situation of creating lists of one thing or another which can be very useful, but I don't know why that cannot be done in subgroups ahead of time or in prepared papers and then have the workshop focus on that.
So, this was our thought. The key thing focused on establishing a vision for first 2005 and I would hope for the whole decade but certainly 2005 in identifying the most critical things that the health statistics system in general needs to be aware of and work toward, but I think this should be done in the context of this idea of the health information infrastructure. That paper provides a starting point for thinking about this.
It certainly says to me, I think the paper is very useful in kind of differentiating health statistics and the mission of these health statistics agencies within HHS and some of the allied agencies for that matter, such as census, differentiating that from some of the broader information issues.
So, what I would hope is that the committee can put this on its work agenda for this year and that we can target a date and with a subcommittee of this committee work toward that.
DR. IEZZONI: Don?
DR. DETMER: Just to comment briefly tomorrow we will be talking with Richard Knowles about the draft of the national information infrastructure paper. The second piece of that is, also, there and needs to be followed up on, and I think, also, candidly would be pretty useful prior to your effort because initially when we crafted this and then reviewed it at the Executive Committee meeting we had, also, a framework and concepts for where the country ought to be going relative to national health and health services to guide essentially what the health information infrastructure will look like.
I think there was good discussion and the feeling of let us tease those two things apart and try to deal with the HII piece but not walk away from also what the national health and health services view ought to look like, and I think we will find out tomorrow actually how the Committee comes out on that, but I hope that later this summer or fall we will continue to work on this other piece as well, and I would like to think that that would be useful to Ed's project.
DR. IEZZONI: In reading through what Dan had proposed which he sent me last week which jibes with what yo have got here it seemed like it was an effort that would get a substantial amount of input from a variety of different sectors of folks that you need to hear from, but it, also, seemed like a bit effort. It seemed like a lot of people, a considerable amount of resources, planning, putting it all together, and I am hearing you talk about the resource constraints for HANES and some of the other issues around resources in your department or center. Do we have a budget for this? Is there a target amount of dollars that is going to be spent on this initiative?
How is this going to be funded to do all these various different activities that are outlined?
DR. SONDIK: We would set aside the dollars. I mean we have a tentative budget for it. I don't think it is as extensive as you may feel it is.
DR. IEZZONI: Travel alone, you know, convening focus groups because you are going to want to bring in people from all over the country; you don't want an inside-the-beltway type of affair and just hearing Marjorie talk about travel budgets for our Committee here, you know, just even travel alone to get folks to the table is going to be expensive.
DR. SONDIK: I don't think we have an alternative. I am convinced that we need to do it, and I would hope that I get Department support from this. We have not discussed this extensively with the Data Council, but I have mentioned it several times. It is very difficult for us, actually to discuss dollars in the Data Council.
Have you noticed that, Jim? I don't understand that.
MR. SCANLON: But it sounds like this is turning into a fiscal 1999 expenditure.
DR. SONDIK: Principally, yes.
MR. SCANLON: So, we start with a new slate. Workshops like this are in the range of about 100 to 120 thousand dollars. They do this often in other areas. They are doing several others for us, but I think the Department would certainly kick in.
DR. STARFIELD: Certainly some foundations probably would kick in small amounts, and if you can get some of those you could handle it.
DR. FRIEDMAN: In terms of the timing I do think that it would be really essential for the commissioned papers and the focus groups to be conducted before the workshop. I don't know what your notion of that is.
DR. SONDIK: Absolutely. That would be in preparation for a workshop because as I said, I have been to too many of these that turn into lists and pieces of paper all around the room and people taking votes on the things that are there, and I find that troublesome. I don't think we need to do that sort of thing. In fact, I don't think it is terribly productive in many cases. In some cases it is great, but in many cases I think we have got a lot of background work that has been done in the past.
I think we need to review that, and I think we need to commission some thoughtful works on issues for the future.
MS. GREENBERG: If you look at the back of this, it is the preparatory activities leading to a workshop and then follow-up steps.
DR. IEZZONI: If you are talking about January, the workshop and commissioned papers and focus groups, that is a very tight time line.
DR. SONDIK: It is. I think we have to be very clear where we are obviously by mid-August on it, but I would rather, I mean obviously the group, all of us have to go through this and fix a date, but certainly in my mind I would hope that we are not talking about something that is far off from that, a year from that, that sort of thing. I don't see this as a conference if you will with that kind of time line.
I see it as really we are aiming for a workshop and the purpose by the way of having this workshop is in order to be able to get recommendations from it, and this is a way that we can get recommendations without, frankly, having to go through a year and one-half to 2-year study if one were to use the Academy mechanism, and it would take at least that.
DR. IEZZONI: So, Dan, is there somebody at NCHS who you talk to? If Jerry Hendershot who is here, have people met Jerry? I didn't introduce him. He is at NCHS and one of our newly kind of anointed staff leaders. He has been working with the work group, but I meant to introduce him to the rest of the Committee who hasn't met him.
So, the two of you are joined on this.
Jim?
MR. SCANLON: Lisa, there are from virtually all of the agencies throughout HHS there are folks who have indicated they would want to participate in the planning in helping the planning from other agencies as well.
DR. IEZZONI: I would think the Data Council would have a big interest in this, too.
At this point then Dan and Ed, how can our subcommittee help you?
DR. SONDIK: You could tell me that it is on the work plan for the year because I thought this was a question as to whether or not you could fit this in for the year, and I would make my pitch that I think it is really important, not only for NCHS which I don't think should be the principal focus of this to begin with. I really think it should be health statistics aimed at policy decision making and research.
DR. IEZZONI: Dan, maybe what we could do is tomorrow morning you can come to us and tell us what that means to have it be on our subcommittee's work plan because I think obviously we want to be helpful, and you are absolutely right. This is a key thing for us to be participating in, but it would be good to hear exactly what it means in terms of what we would have to do and just so we can put it into our work plan.
Ed, I don't know if you know this but today is the first day of the newly reconstituted Subcommittee on Populations, and so we haven't even had a chance yet to talk among ourselves about where we are going or what we are going to be doing, but Dan will certainly, I am sure, shepherd us on this path tomorrow morning when we do talk about our work plan.
DR. SONDIK: If I could just say a word about at least my conception of what it would mean, I think it would mean identifying it as a priority which would mean that some subsection of this Committee would need to meet several times on it in the usual fashion for something like this and I think it would be a principal activity for those members for several months through the planning phase which I would hope would be by the beginning of September, be finished by the beginning of September which is tough, of course, given August, but I think we can certainly have a concept for this that is quite clear, and then there would be the less intensive shepherding of it and meeting periodically on it. I would see that the department staff and I am assuming here that CNSTAT would continue to be involved and the workshop itself would be held under their auspices in which case their staff would take a very significant role in moving it forward.
The critical part is really the planning phase of this in which there would be quite a bit of interaction with CNSTAT and obviously the commissioning of the papers and other activities.
DR. IEZZONI: Okay, so a summer-intensive commitment.
DR. FRIEDMAN: And the only thing I would add to that, I do think that sometime in the next oh, I don't know, month, we should, also, have a very serious discussion around time frames. I hate missing deadlines, and at the same time I hate sounding like Cassandra about deadlines, but I think we should be as realistic as we can about what is plausible in terms of papers, focus groups and so forth and products.
MS. RIME: And RFP development and other things maybe.
DR. FRIEDMAN: You can't just write out checks to people now.
DR. STARFIELD: There is one thing I don't see here on this paper and that is who is going to develop the topics for the papers. It says that we are responsible for supporting the process, but that you are going to commission the papers. Does that mean the topics, that basically you are going to decide on the topics?
MS. GREENBERG: That was proposed in the document. We will get everyone on this group, also, a copy of that document.
DR. SONDIK: Just to answer that, we would not obviously unilaterally choose the topics by any means. I would think that would come from the planning group.
DR. IEZZONI: I guess to the extent that we could have small planning groups that could conference call rather than actually physically meet I think you are going to find a lot more people willing to participate.
MS. GREENBERG: I think part of the concept, and I don't know if that is different, two-way or three-way conversations and even at NCHS we haven't all recently been in the same room to discuss this, but I think part of the thinking behind the document you passed out was that although the papers would obviously lead into the workshop, they wouldn't all be commissioned under the CNSTAT umbrella, so that some of these preliminary things that the National Committee would take the lead for would certainly benefit from comments from and input from the data council members, etc., but could move ahead kind of deciding on certain topics even before you had a fully planned workshop because of the importance of doing some of this preliminary work, and so, we have tried to kind of break it out so that everybody wasn't dependent on everybody else because that really slows down the process.
I guess we can talk about this a little more tomorrow morning.
DR. IEZZONI: Okay, yes, I think we will make every effort to do what you need, Ed, because I think it is in all of our best interests and it certainly is something that speaks to the goals of the NCVHS as a whole and our subcommittee in particular.
DR. SONDIK: Okay, thank you.
DR. IEZZONI: Okay, Kathy is not going to be with us tomorrow morning, and neither is Hortensia. Hortensia, I think you might be able to teleconference in, conference call in?
DR. AMARO: I have to check what time.
DR. IEZZONI: It is going to be at eight.
DR. AMARO: Eight o'clock.
DR. IEZZONI: Eight to ten. Hortensia is not sure whether she is available.
DR. AMARO: I have meetings that I didn't have when I mentioned that. That is the reason I cannot be here.
DR. IEZZONI: Because tomorrow morning is actually very important for all of us because we really need to regroup. We are going to have a slightly changed focus of our mission. I don't think the overall mission will change very much, but a lot of the specifics might be reorganized, reprioritized and so on, and we are going to really miss the input of Kathy because Kathy has this whole working group on quality that is now within our subcommittee.
Maybe, Kathy, you could just, since we do have a few minutes left, tell us if you can, if I can put you on the spot, just give us a sense of what your vision is for how the working group on quality might proceed so people can be thinking about it for tomorrow morning's breakout?
MS. COLTIN: I am thinking of the scope of work really for this work group as focusing mostly on the building blocks for measuring quality, meaning the data infrastructure that supports the ability to create measures of quality.
I think there are enough groups out there that are developing measures of quality that we don't need to go into competition with them, but they, I think are running up against problems and limitations that stem from the data that are available to work with and to draw upon in creating measures of quality, and I think that is an area where we have an opportunity to make a contribution.
So, one of the things that I would like to do is to learn from those who are developing measures of quality what they see as the major data limitations that they are coming up against. That would be one set of activities, you know, running whether it is hearings or just panel presentations, but I would like to hear from various groups that are trying to develop measures in different care settings for different populations around what the data limitations are that they are finding that are problems for them and that are obstacles to measuring things that they believe are important to measure.
Another thing is to really take a look at as we were doing earlier in the Committee the goals and objectives that were laid out in the President's Advisory Commission Report and see what are some of the data needs and data limitations that may affect their ability to move forward on some of those objectives, and I am thinking, you know, right down the line from the sort of global objective of setting priorities for measurement from information available on which to set those priorities.
Now, what are the different things that would be considered in determining what those priorities are, and I think that on the one hand we have some very good data about morbidity and mortality and prevalence of various conditions.
On the other hand I am more skeptical about what we know about where the problems are, you know and what data do we have that inform that.
So, that would be maybe a starting point to do some sort of an assessment of the adequacy of information to support prioritization.
Then once you have priorities and you start saying, "Okay, now, we are going to actually try to develop some measures in these priority areas," then what are the data limitations that you run up against when you are trying to develop some of the measures.
So, I think you can kind of go right down the line on a lot of those recommendations and say, "Are there data issues embodied in trying to achieve that particular objective and if so what are they and what can we learn about it?"
I guess another example would be that there is a clear objective to try to reduce errors in health care. What kinds of data do you need to develop an error tracking system; what do we have now; what works; what doesn't work; what would be necessary to put in place in order to have such a system?
So, those are the kinds of data issues, and that is really where I thought the focus of the work group would be more enabling of those who are trying to develop measures.
MS. GREENBERG: I just wonder to what extent you can build on the road map that was developed by NCQA.
MS. COLTIN: I think that will be a big part of it, but I think that we will, also need to look at some of the enabling and constraining forces around implementing the road map as well because there were a lot of those identified. Whether you are talking about safety net providers and their ability to implement computerized patient record systems or whether you are talking about confidentiality issues that relate to health plans having access to laboratory test results, you know, where are all the points at which moving full steam ahead on that are going to run into obstacles that we might be able to help address?
DR. IEZZONI: Okay, that sounds very, very rational and like an excellent plan. Do we have people signed up for doing this? Who have we got so far?
MS. GREENBERG: Elizabeth Ward and John Lumpkin have both signed up.
DR. IEZZONI: Excellent.
DR. STARFIELD: The paper hasn't been around because I haven't gotten the paper.
MS. GREENBERG: No, this is incomplete. Two means that we already have three.
DR. IEZZONI: Okay, and is there a quorum rule around working groups, Marjorie?
MS. GREENBERG: I was suggesting this, oh, a quorum rule.
DR. IEZZONI: Like if we wanted to hold hearings.
Can working groups hold hearings?
MS. GREENBERG: We discussed that. I think it maybe is more effective for an actual hearing in which you are bringing in a lot of testimony to --
DR. IEZZONI: To do that under the subcommittee's auspices.
MS. GREENBERG: A subcommittee, but, also, you can make up a quorum with, it doesn't just have to be the members of -- let us say there are I don't know, eight people on the Subcommittee, and you need five people there, and they can be other members of the National Committee, too. So I think usually if you are going to bring in people and announce hearings, etc., you want enough committee members there to demonstrate the interest, but certainly you can have more working sessions.
DR. IEZZONI: I am beginning to wonder with what Kathy articulated this morning how we are going to pull off all of this because I really think that, Don, we might need to talk about using some of the 2 days that we have for full Committee meetings for maybe one-half day for a hearing or a panel presentation for something that is of great interest to a subcommittee whose members really overlap like John Lumpkin is chairing another subcommittee, and so, if we were to hold a breakout session where we had a panel talking about the issues John would probably want to be there, and so, I think that that is going to be the big challenge to be able to pull this off because I think probably most people on our Subcommittee are going to be interest in quality. How could you not be?
Dan, you had a question?
DR. FRIEDMAN: Is there a notion that this would go beyond administrative data?
DR. IEZZONI: Oh, absolutely, and I think clearly there will need to be some issues that we will want to ask the work group on the computerized patient record to address as well.
I think measuring functional status is going to be a really key issue here because clinical end points only go so far. People want to know how they are going to feel, and for quality measurement purposes, and that is going to be something that I think is really going to cut across our Subcommittee and the standards subcommittee.
Marjorie?
MS. GREENBERG: I know Barbara brought up earlier about the core data elements, and we had some e-mail exchange, which I hope sheds some light about where they, that they do currently live but in a different kind of maybe form than a few years ago, but I think one of the elements that there was a lot of discussion about, a lot of consideration was functional status, but a strong recommendation that really more work needed to be done on that, and I don't think the Committee has really had too much opportunity to return to that, and, also, I think, I mean there have been efforts to actually put functional status into the administrative transactions and then a kind of backing off of that, but it is something that that subcommittee, also, may have some future in.
DR. IEZZONI: It is very central to the postacute care data issues that I think we, also, really need to focus on because of the Balanced Budget Act.
All right, while Kathy is here, Elizabeth as you want to be part of her group do you have any questions for her specifically?
MS. WARD: No.
MS. COLTIN: One thing that I didn't mention explicitly but which I think Elizabeth and John would both be interested in has to do with some of the issues of linkage of public and private data as well and Dan may be interested in that area a little bit, too. So, that would certainly be something else that we would be interested in.
DR. STARFIELD: Could I just make a comment now that I have seen these papers? I think part of the problem is that the work plans of some of these new work groups are not very clear, and people may have difficulty deciding where they want to go with that, having a feel for what the work groups are going to do, and I think that is why you are getting some work groups without any volunteers, but we may have to revisit that.
DR. IEZZONI: Actually, Don, let me just let you know, I haven't had a chance to talk to all of our Subcommittee members, but I have talked to a few folks, and there is some sense that having a work group on population-based data within a Subcommittee on Populations where virtually all of us are interested in population-based data doesn't make any sense.
DR. DETMER: I assumed that you guys would probably decide to --
DR. IEZZONI: Right. That is being jettisoned. So, if that paper or if that work group doesn't have any people signed up for it yet, it is a moot issue because I think that we are going to take that on as a full Subcommittee, if you will and so, the only real work group within our Subcommittee will be the work group that Kathy is leading, but we will, also, have what we might want to call a work group or a task group or whatever to help Ed with what he wants our Committee's assistance with.
DR. SONDIK: Let me ask a question? One of the things that the full Committee talked about now sometime ago was working specifically with NCHS on advice, functioning as what I would call an advisory body for NCHS which I view this project, if you will in terms of this workshop or whatever form it would take, I view this as part of that activity.
So, I don't see it necessarily as specific to the population subcommittee as much as an NCVHS activity, and so I am suggesting that since you are going to meeting for reorganization and so forth I am just kind of putting that on the table to tell you that I don't see it as much as this may be an activity at this point or a specific task, but I see it as part of this continuing activity which has to do with NCHS and its programs.
DR. IEZZONI: I hear what you are saying, and, Don do you have any comments for Ed, on that point?
DR. DETMER: I don't think there are too many things that we could have as a Committee of the whole all along. Generally I think we have subcommittees that try to work to help us and some of those are partitioned out, and then there is always this imperfect splitting and clumping, and that we can talk about 9000 times, and it will still be there. I mean you have got to make decisions, and you have to deal with it. Clearly I think NCVHS is important enough that the full Committee will want to absolutely weigh in on the materials that come forward from Dan's activity and so forth.
It looks to me that if you look at the current subcommittees to bring things to the full Committee it is either standards and security or populations, it seems to me. So, it is much more so than privacy and confidentiality, it strikes me, but I think that your group needs to think about it. We clearly had anticipated that Dan would be relating to this group, but I think it is you folks' call, but that does not mean that it is not a full Committee interest and issue.
DR. IEZZONI: Right.
DR. SONDIK: My only point was that I hope that it is a continuing thread.
DR. DETMER: Sure.
DR. SONDIK: As opposed to a specific defined task.
DR. IEZZONI: I guess what I was doing was thinking very focused. We need people that are identified who can be responsible for you to do certain things immediately to help Jerry get the workshop up, but the role of giving advice over time is something that is ongoing.
Okay, I see folks fading around the table. Does anybody have anything interesting or insightful that they would like to say in conclusion or otherwise kind of just necessary?
MS. ARAKI: We forgot to acknowledge the Office of Minority Health for the support that they are giving to the Committee to travel and the representatives from the insular --
DR. IEZZONI: Wonderful
MS. ARAKI: Pacific insular areas.
DR. IEZZONI: Yes, thank you very much. I think that that is going to be a really exciting meeting. That is going to be fun.
DR. STARFIELD: Can we leave any materials here overnight?
DR. IEZZONI: We can leave stuff here, I think, probably.
Now, Richard, are you going to be joining us tomorrow morning or are you going to be at the privacy?
MR. HARDING: Privacy.
DR. IEZZONI: Okay, and you will, also, be with the privacy. Oh, dear, big problem. Kathy won't be here.
MS. ARAKI: Hortensia said that she could participate between nine and nine-thirty, for sure, and then she will try for earlier.
DR. IEZZONI: Okay, we are having a key example right now of why this is going to be hard to make it work.
DR. DETMER: It will always be that way. I don't know how you are ever going to be satisfied on it.
DR. IEZZONI: No, Don, all I am saying is I am just reiterating because I have just now heard but --
DR. DETMER: How do you solve it? I mean I don't see that you solve it.
DR. IEZZONI: Okay, tomorrow morning those of us who are available will meet at 8 o'clock, and we will try to figure out what we are going to do.
MS. GREENBERG: It is the discussion we had about the standards committees. You know, they try to inform everybody, but basically the people who show up want to make the recommendations, and then everybody kind of has a chance to vote on them or whatever.
MS. WARD: I would, also, comment that we are starting today with things that got planned. We are today writing up pieces of paper that say that we are going to be on some new activities. I still think with scheduling into the future we will not necessarily have two of us trying to go to two meetings at the same time. I think that is a problem for tomorrow morning. I don't think that means it will be impossible.
MS. GREENBERG: It is going to be challenging, particularly if we want to do as much in the context of full committee meetings so that we are not just asking you to come for a day or people from far distances, but I think ultimately there are going to be some priorities and people, groups will have work plans but they may not be executed in as short a time frame as other work plans.
There will be a series of hearings maybe decided upon and some will be in the next 6 months, and others might have to wait because I think to be fully operational and really fully active on all of these activities at the same time even with sufficient staff I don't think you are going to have sufficient members, but there seems to be a commitment to each of these areas. So, we will just try to schedule them as best we can.
MS. COLTIN: Doesn't that suggest that we ought to be thinking ahead to when our next meeting should be and how it should tie into the next meeting of the full Committee or whether it should be an independent meeting because we are not going to be together again as a full group until September.
MS. GREENBERG: As I see it we actually have September 14, 15, 16 and 17, kind of as a piece there. I mean 15 and 16 is a full Committee meeting. I think we have the 17th as a meeting for this Subcommittee.
DR. IEZZONI: To talk about Medicaid managed care.
MS. RIME: The other thing to follow up with Kathy's point, too, is we, also, need to talk about and Lisa had it on the agenda for tomorrow how we are going to communicate as a group.
DR. STARFIELD: It doesn't sound to me like we will have time for a meeting because if you have already got the agenda for the September 17 meeting, I understood Carolyn to say that the 17th is devoted to Medicaid managed care.
DR. IEZZONI: Assuming that the report is written and our subcontractor is a little bit late on deliverables. We will talk about this tomorrow because one of the things I want to talk about tomorrow, also, was how we are going to interact around the Medicaid managed care report but it doesn't look like anybody else on the Medicaid managed care is going to be there. I will be the only one.
Okay, we will deal with this tomorrow morning. Okay, so, we are adjourning for now and thank everybody and thanks, staff.
(Thereupon, at 5:17 p.m., a recess was taken until 8 a.m., the following day, Wednesday, June 17, 1998.)