[This Transcript is Unedited]

National Committee on Vital and Health Statistics

Subcommittee on Populations

September 8, 2005

National Center for Health Statistics/CDC
3311 Toledo Road
Hyattsville, MD 20782

Proceedings by:
CASET Associates, Ltd.
10201 Lee Highway, suite 180
Fairfax, Virginia 22030
(703) 352-0091

TABLE OF CONTENTS

  • Welcome and Introductions – Donald Steinwachs, Ph.D., Chair
  • Measuring Healthcare Quality and Disparities – Conversation with AHRQ Staff – Subcommittee
  • Follow-up Discussions on Subcommittee’s Work Plan – Subcommittee

P R O C E E D I N G S [3:40 p.m.]

Agenda Item: Welcome and Introduction – Dr. Steinwachs

DR. STEINWACHS: Why don’t we get started, I think that people are going to be having to leave so we want to try to use your time as well as we can. You have an agenda for this meeting, Nancy wanted to move up to the top of the agenda with an update on the final report on data collection on small populations dissemination, is that the same as, it’s not the same as this?

DR. BREEN: It is this, I wanted to know what the dissemination plans were too, maybe we know, maybe we don’t.

DR. STEINWACHS: Debbie was here, Debbie was just sharing that we have the mock-ups here but not enough for everyone so we’re passing that around. I know that there is a dissemination list being put together, Audrey, are there some things you could say about it or do you know —

MS. BURWELL: Debbie is going upstairs now to print off a copy for all of us but basically it’s all the suggestions that people made such as Nancy and the full committee made in addition to a number of organizational components around the country that have testified at these hearings in addition to all of the NIH, ICD directors, selected CDC centers and offices, in addition to a lot of different offices within the Secretary’s office, for instance ASPE and the Assistant Secretary for Health and a few other key locations. And also the authors and the Committee of the Eliminating Health Disparities from the IOM report and several other places like that. That’s all I can remember off the top of my head but Debbie is going to bring down the list, it’s going to go to the state Office of Minority Health and to all the regions, so we’ll see the entire list when it gets here.

DR. STEINWACHS: So we ought to, if there are suggestions for more on the list, more is better, also in terms of the schedule I guess hopefully this will be heading out to the printers near the end of the month, then is that a couple months process or so?

MS. BURWELL: Debbie had an estimate of it being out probably by the end of the year.

PARTICIPANT: Do you think it will be ready for APHA?

DR. STEINWACHS: APHA is now mid-December now, it’s been moved to Philadelphia —

MS. GREENBERG: Oh, I hadn’t heard what the result was.

DR. STEINWACHS: Yeah, they’re going to do —

MS. GREENBERG: I would think that would be a good goal, the thing is we could not send it to printing until the new fiscal year but I would hope so, let’s try to at least make that a goal, once Debbie gets back, I think Debbie is going to APHA also. I hadn’t heard about the rescheduling. Do you know what the actual dates are of APHA?

DR. STEINWACHS: 10th, 9th, maybe 10th, something like that.

[Multiple speakers.]

DR. STEINWACHS: While people check on the dates why don’t we try and move into the agenda for today —

MS. BURWELL: It’s the 10th to the 13th.

DR. STEINWACHS: 10th to the 13th then. We had come to sort of a consensus and that’s always open for further discussion that what would be great for the Population Subcommittee to do would be to take on issues around data collection, metrics, strategies, that underpin our capacity to talk about quality at the population level and recognizing —

PARTICIPANT: Is anyone on the phone?

MR. GREENWAY: Yeah, it’s Kirk Greenway.

DR. STEINWACHS: Kirk, can you hear?

MR. GREENWAY: Yes, I can hear you, there was some interruption due to a serious of telephone tones —

DR. STEINWACHS: That wasn’t us, someone else.

MR. GREENWAY: Okay, thanks, maybe someone trying to dial in.

Agenda Item: Measuring Healthcare Quality and Disparities – Conversation with AHRQ Staff – Subcommittee

DR. STEINWACHS: This is a secure line, not to worry.

So we had at least reached an initial consensus that we would explore the area of population level quality measures, particular focus on disparities which is a major quality issue and one that builds off of what we’ve been doing already. And so I was very happy that we were able to enlist the participation and support of several people, all known to you, who can talk about the issues as they saw them and lived through them in producing now what has been the first two rounds of reports on quality of care and also on disparities.

And so Anna is here representing AHRQ and Julia and Irma are here representing NCHS, and as I understand they all work together which is not unusual and was very productive. And so I don’t know who would like to start off first but the idea was to try and begin to talk about what are some of the issues with the hope that we would come out of this saying well what is it that this committee might take on, what’s of interest to us, what might we take on, and both looking for things that are short term where there could be recommendations that would be useful as well if there are longer term issues. Certainly as we go into this discussion we may come out and end up feeling differently about what are the priorities and directions but it’s a great chance for us to get going.

Just to say one follow-up thing because we aren’t going to probably time to do it today was I thought we’d try and schedule a conference call the next couple three weeks if we could to talk about next steps because if next steps are hearings or there are other next steps we need to start making those decisions soon and getting this moving along the pathway.

MS. POKER: Do you want me to just start and you guys, or you want to start —

DR. HOLMES: Well, I thought maybe we might point out that for further elucidation these medical care articles, there are two articles in there that are jointly authored by AHRQ and NCHS that deal with, specifically with some of the issues and reporting on, and gathering data on race and ethnicity that were encountered with the reports as well as problems in gathering data on and reporting on SES.

And I thought I just might start out maybe by mentioning some of the problems in terms of the gathering of data and reporting on race ethnicity and then, I don’t know Irma if you’d like to maybe talk about SES, the same issues, and then if you had any additional issues. But there are three issues actually I think in terms of reporting on race ethnicity, one is collecting the data and number two is who reports on reporting the data and then whether the data can actually be reported or not. And one of the main issues of course for us when the reports began when we began with the years ’99 and 2000 is that we had over, gosh I don’t know, whether it was 36 different data sources for the disparities report, that included surveys, that included administrative databases, it included surveillance systems such as the CDC HIV surveillance system.

And they differed in terms of where they were with adapting the OMB standards for reporting of race and ethnicity that is incumbent on federal reporting systems, that change was made in 1997, but federal data systems were given until I think to 2003 to fully implement them. And the main changes were number one, the ability to report multiple race and contrast to prior rules for reporting on federal systems, and also splitting Asian from the National Hawaiian/Other Pacific Islander. So those were the two major changes with respect to reporting of race.

In terms of ethnicity the change was from just reporting Hispanic to Hispanic or Latino was it? So at any rate all these 37 or 36 different data systems were at different places so that some of the data was able to collect data on Asians separately from the Native Hawaiian/Other Pacific Islanders, some data systems reported on multiple race. For example the National Health Interview Survey had reported on multiple race for a number of years, not just given the change.

Now that will change obviously as the different data systems reach the reporting year of 2003 and it’s incumbent upon them to report given the new OMB standards. On the other hand what’s important to remember is that non-federal data systems, it’s not incumbent upon them, the OMB 1997 standards are not incumbent upon those data systems so they can choose to either collect data on race and ethnicity at all, or they can choose to report on it and categorize it in their own way. So that’s a major problem in terms of talking about reporting on disparities.

A second issue then is in terms of who reports on race and ethnicity and there you have the major disconnect between whether you’re talking about a population person based survey or an establishment based survey, that is a provider based survey. Obviously in a population based survey, a person based survey, the individual is able to self report on their own race ethnicity and that is obviously the preferred method.

But a lot of the data that are in the National Health Care Quality and Disparities Reports are based on administrative data, say HCUP, or they’re the provider based surveys, as the surveys that we do in our division, the National Healthcare Hospital Discharge Survey, the ambulatory care surveys, and the long term care surveys, the nursing home survey and the home and hospice survey. And there the data are reported off of medical records where one is not sure whether the individual has provided information on race or ethnicity or whether the health care provider completing that form is providing the information.

So oftentimes the data are either not collected at all or one questions the accuracy of the reporting on race and ethnicity, particularly things like Hispanic, non-Hispanic and so on. So there are tremendous differences in the reports and our ability to report on disparities by race ethnicity.

In terms of then the final issue which is the actual ability to report we then get to an area that I’m sure you’re all familiar with which is small numbers, that is even if you collect the information on race and ethnicity that in terms of being able to report it, a reliable number on a number of different variables, that the numbers are unreliable, statistically unreliable, and as you all know too data systems, federal data systems usually have their own rules about how large a number you have to have in order to report it. And the usual default is if you have under 30 cases or if your relative standard error is greater then 30 percent you would not report a number.

In this book that I’ve referenced these two articles that we contributed you’ll be able to see, I’m trying to get the page, about particularly when we went to report on priority populations, that is African American women or Asian children, that for many you could only report out of the 150 quality measures and the 150 access measures, that we were probably only able to report for these certain groups maybe 30 percent, and very few could we present data on most of the majority of the measures. So that’s a real shortcoming of our ability to report on race ethnicity.

So I think that kind of summarizes the difficulties that we encountered and will continue to encounter in most cases with reporting race and ethnicity in the disparities and quality report. And I think it would be wonderful if this group could take on some strategies about how to deal with these issues. Some of them are pretty difficult to deal with but that’s not to say that we can’t make improvements in either the collecting of data, who provides the information, and then finally the reporting of data on race ethnicity.

Irma, do you want to —

MS. ARISPE: I think the only thing that I would add, since the report we have started to I guess maybe in part because of the report, the NIMS(?) and the HIMS have over time not reported ethnicity because of the high percentage of missing data. Now the missing data rate is much lower so for the first time in 2000, well this year in our emergency department report and in the physician office and hospital report we’ll be reporting by Hispanic ethnicity.

And Kathy Burke has been doing some analysis for APHA looking at injury rates for Hispanics and non-Hispanics and one of the difficult issues as Julia has mentioned is that when you are collecting data from the facility, from health care establishments, you don’t really know if what you’re seeing is true variation in injury patterns or an artifact of data collection. So to try to assess that she’s looked at different types of illnesses, illness visits and injury visits, to begin to try to sort out whether the patterns are the same which would suggest that it’s a reporting or data issue versus variation by condition.

The other issue, the difficulty for us with implementing the OMB, the 1997 reporting standards, is we changed the data collection form to collect from the provider the ethnicity but we often don’t have the denominator to use so it creates, there’s a real disconnect for us, so sorting out the rate is very difficult. Now for establishment of base surveys there aren’t that many missing, there aren’t that many multiple race cases, the samples are very small, there’s not a lot of data to be working with but over that’s going to be an issue.

As we redesign the hospital discharge survey and the ambulatory surgery survey we’ve talked about perhaps over sampling in states that have high minority populations and that’s a very, it’s a different thing for NCHS to do, we haven’t traditionally done state estimates, our goal is national estimates, we haven’t really had the funding to do the state level estimation that we need but the area of disparities is so important to the department and to CDC that it’s something that we certainly would like to think about doing in the future for those surveys.

The issues with socioeconomic status are somewhat similar, for a person based survey is probably the best reporting you’re going to get and even then there are some methodological issues. On vital statistics there’s education, which can be reported by anyone from a family of the person whose passed to the funeral director. I think to some extent the reports and in general people talk about race in the, the race variable the same way even though the quality of the race data is very different depending on the facility generated variable or a person generated variable.

So in the reports I think the general position particularly on SES is to go with the person based data source rather then something like education which is not really a very good indicator of SES. Or in the health care utilization surveys expected source of payment, we went around and around with what does that actually tell you about SES and unfortunately it doesn’t really tell you very much but everybody uses it all the time —

DR. BREEN: Could you talk a little bit about that measure, expected source of payment?

MS. ARISPE: Well, it’s generated at the time, well, at the time the patient presents to the health care facility, its collected by the registrar generally, and when the patient is in the encounter, say if it’s an emergency department, there’s some negotiation about who actually is responsible for covering what the patient’s illness is, if there’s a mental health issue versus a medical issue, to try to sort of out who’s providing coverage. In the hospital side it’s often over a year before the final payment source is adjudicated —

DR. STEINWACHS: And it’s only usually the principle payer, it’s not, it’s only the first payer so if you have multiple payers you don’t know who the multiple are.

DR. HOLMES: Right, right, so for example we have Medicare, well what does Medicare mean, what does that have to do at all with socioeconomic status.

MS. GREENBERG: You can’t collect Medicare —

MS. ARISPE: It’s just insurance places, it’s not out of pocket —

DR. STEINWACHS: There’s self pay but self pay may be people who can’t pay —

DR. BREEN: I was just wondering what the options are on the typical coding for this because I’m not familiar with this as a variable.

MS. ARISPE: I guess it depends on the survey, it’s Medicare and Medicaid, worker’s compensation, self pay —

DR. HOLMES: Private insurance.

MS. ARISPE: Private insurance —

DR. HOLMES: And then sometimes they’ll put Blue Cross/Blue Shield separate from private insurance which is not, they’re not mutually exclusive, and we have a great deal of trouble too for example with determining whether it’s managed care because of the fact that the information that we receive is very basic as to just whether it’s the Medicare or Medicaid and not whether it’s managed care. But the issue becomes how would you pursue that, how would you find that out out of a medical record.

MS. GREENBERG: I was just going to say that I don’t know if you’re familiar with the work that Amy Bernstein has been doing with the Public Health Data Standards Consortium, because this is a variable that used in those nasty minimum datasets or core datasets that states use for hospital discharge data, and it’s collected by all of I think the states.

Again, it hasn’t really been standardized very well and you don’t have necessarily, as you said you can’t tell managed care, you can’t tell whether if they had Medicare and Medicaid, what have you, and then if you looked at the standards development organizations the standard that they have for this item, because they do have it in their datasets, is a completely non-standard standard in that it’s totally redundant, it’s just, it’s a very poor set of categories. So because of the interest of particularly state hospital discharge groups systems and states, because they do use the data a lot, I mean it’s like a lot of data, it may not be that great but it’s better then nothing.

We have come up with this kind of standardized taxonomy that kind of starts at a very general level and then you can pull it out depending upon how much detail it’s possible for you to collect, like in some states they collect a lot of detail, and it’s being, we’re trying to promote it as a way to move towards greater standardization. The main impact here that could be quite relevant is when the plan ID comes out, that’s the plan, unique identifiers for plans, it’s one of the HIPAA standards, it has not yet come out as a proposed rule though it is supposed to be coming out as a proposed rule, it has been supposed to for some time and I don’t know whether they’re still staying it’s going to come out this year or not.

But that could improve this situation but it could make it almost even worse because there are some, I don’t know what’s going to be recommended because I’ve seen many different versions of this and none recently but to the extent that this information can be useful for SES or for other purposes there may be, it may be recommended that there’s a good taxonomy for it or they may actually recommend that it’s pretty much just up to the payer however they want to do it, so you could just do Blue Cross or you could do Blue Cross all the way down.

And that’s something that, I mean eventually I think the, when there is an NPRM, if there is one and I assume there will be, it’s a required HIPAA standard, this will be something that the committee will comment on and I would just recommend that it’s something that this subcommittee, the standards group will probably take the lead on it but it’s something that this committee should be thinking about to the extent that you consider it an important variable or useful variable, because it could influence quite a bit what the final decision is on that.

DR. HOLMES: And just as an addition to what Marjorie has said, in our division that does the provider surveys we are in the process of redesigning data collection for the National Hospital Discharge Survey and some of the other surveys so that it would be very helpful, I mean that’s another task that would be very helpful if we could get some direction about new ways and more accurate and reliable ways to gather information on expected source of payment. I hadn’t even known about that.

MS. GREENBERG: Amy was in health care but she’s in a different part but she got involved with the consortium.

DR. HOLMES: But this plan ID number is something that we should keep in mind.

MS. GREENBERG: We originally saw great hope for it and then, from a statistical research whatever point of view, and then it, like with many things it has to meet several needs and it’s a question of which needs take priority but I just alert you to that, it’s something to keep an eye on from a populations point of view.

I warned to ask Irma, you said that education wasn’t a very good socioeconomic variable, I don’t know, I guess I had thought that it was found to be quite highly correlated with —

MS. ARISPE: If it was self reported —

MS. GREENBERG: Oh, you’re saying if it isn’t self reported, okay, yes —

PARTICIPANT: If accurate it’s very useful.

MS. GREENBERG: Okay, you were saying if it isn’t self reported, thank you for clarifying that.

DR. HOLMES: And on vital statistics it’s not reported for I believe for those over 65 —

MS. GREENBERG: It’s only on the death certificate, it’s not on the birth.

DR. HOLMES: Well, I mean on the death certificate because of unreliable reports for the over 65 population.

DR. STEUERLE: I thought for whatever purposes, not this one, but that education is actually better then annual income —

MS. GREENBERG: I guess if it’s self reported.

DR. STEUERLE: — it’s a longer term measure, it’s more like a lifetime measure —

DR. VIGILANTE: As people retire from the work force, get married and one of the spouses decides not to work anymore, I mean there’s all these things that education corrects for.

DR. HOLMES: I think there’s some research that suggests that education may act differently, interact differently with income and SES for different race ethnicity groups too so that’s something to keep in mind.

DR. STEINWACHS: Gene, one time you were raising income and then other kinds of variables, and I guess I was mainly thinking about utility within maybe Medicare or Social Security —

DR. STEUERLE: The anecdotal example I gave of a great dataset I got to work with one time was a, Social Security was basically the SIPP, survey and income program participation merged with Social Security —

DR. VIGILANTE: This was a secondary data source?

[Multiple speakers.]

DR. STEUERLE: SIPP was merged with Social Security records and there was all sorts of fights over who could have access, it’s not available on the outside but within Social Security they finally decided that within the Social Security Administration they could have access. And then they went through another series of elaborations to allow some of us on the outside to collaborate with Social Security researchers which also got into issues because a lot of times Social Security doesn’t even want to issue things that might be politically sensitive, to finally actually not that we look at the data but we can look at the results.

But what was really useful about this is because SIPP had a lot of the socioeconomic variables and then Social Security had the lifetime earnings variables that ended up to be very useful and the question was, and I know there have been one or two of these merges but I’ve sort of lost track of it, is what happens if you take a dataset like that and you merge it with Medicare data. And admittedly that’s more end of life but you still, you have some measure of lifetime earnings, not that for most of the population it’s long enough, sometimes you have 20 years, or ten, you guys know what I’m saying, but you start getting some stuff in, in disability, when disability becomes available, you just get a lot of different ways of really analyzing data.

Of course the mortality data is there because you know when the people died, so it turns out to be this extraordinary source of data, I think that’s what you were referring to.

DR. STEINWACHS: Yeah —

DR. VIGILANTE: And it’s not generally available as of now —

DR. STEUERLE: I thought someone had told me at some point that in the past Medicare and Social Security records had been merged but I don’t know who, maybe just the actuaries do it for actuarial purposes —

DR. HOLMES: It’s not generally done.

DR. STEUERLE: I think it has been done once or twice for some select samples, and I don’t really, that’s as much as I remember on that. I know it would be a tremendous set of information. It also by the way provides something which you didn’t get into, to me I’m always also very interested in the financial side of this, how many dollars are going out too for different payments and of course in Medicare, you sort of know people have Medicare or insurance but you don’t know what their getting. If you actually have the Medicare you actually have, as opposed to the insurance value of Medicare you actually have an actual usage of —

DR. VIGILANTE: So does that have not just income but assets as well? In other words —

DR. STEUERLE: Well, the SIPP menus is limited, whether it was a sample of 2,000, 3,000, so it’s all the subpopulation problems you raised come up in spades there but nonetheless it does have a lot of socioeconomic data. SIPP is concentrated mainly on lower incomes, it’s weighted towards lower income population.

DR. VIGILANTE: And it is repeated —

DR. BREEN: And I think it does have asset information but it’s capped, is that what you meant by limited?

DR. VIGILANTE:: I mean by asset, home owner or not —

DR. STEUERLE: I meant it’s data limited, it’s not a good sample of higher income people because of the size if it’s weighted toward low income population but you’re right, the asset data is mixed.

DR. STEINWACHS: It’s often said that the health care disparities in quality are greatest by socioeconomic and then by race and ethnicity, so the challenge if you were to look at quality measures and understand whose at greatest risk, you really need both, otherwise race ethnicity becomes a proxy many times for different socioeconomic groupings and you fail to recognize that it’s not just race —

DR. VIGILANTE: And it also governs policy decisions about how to make interventions because it may be that the single best health intervention is not actually a health intervention, it may be an educational intervention —

MS. GREENBERG: Or housing.

DR. VIGILANTE: Or housing, or some combination thereof, and I think that’s the problem when health care people talk about and think about health disparities, you go to a baker and you get bread and you get solutions that are health care centric and when in reality the solutions are not, or may not be, and the folks outside the health care aren’t thinking about health disparities, they’re thinking about their thing. And so I really think that this is a forum and an opportunity actually when we’re talking about merging datasets or linking datasets or really to be looking at health disparities that’s really interdisciplinary, that informs policy making at a higher level because we’re not going to solve the health disparities just by improving access, it’s very important but it’s just not going to solve it.

DR. BREEN: I think that raises a question for Irma and Julia, our experts, in cancer there aren’t that many datasets that you can actually look at race by socioeconomic status or vice versa and that’s really important for the reasons that this discussion has just illuminated, and I wonder in health more generally, because I know you looked at nearly 100 indicators, were you able to parse that out and stratify your racial data by socioeconomic status or socioeconomic status by race?

DR. HOLMES: Maybe Jennifer might want to respond to that because there was, I mean obviously you want to know what the interaction effect is between race and SES and I mean we weren’t able to do it, we were only able to use two datasets, Health Interview Surveys and MEPS(?), I mean those were the only two that had a sufficient sample size that we could even begin to run analyses that looked at the interaction between race and ethnicity.

DR. BREEN: So that’s something that we really need to work on.

DR. HOLMES: Yes, I mean all the problems I stated about sample size and reporting and so on, as soon as we began to do those analyses they were magnified beyond belief and as I said we ended up with two datasets that we could do that type of analysis on it and that has not changed. I don’t know whether you have anything to add Jennifer —

MS. MADANS: Well, you remember the Health US, I don’t know how many years ago the chart book was about this particular thing, I think pretty much mined what we could do with available data, but it’s a sample size issue. With HIS did you combine your data?

DR. HOLMES: No —

MS. MADANS: Well if you’re doing something every year —

DR. HOLMES: Because now they’re looking at over time —

MS. MADANS: But one way to increase it is pull the data, and also the way we over sample race ethnicity does tend to reduce the variability on the socioeconomic status because you’re looking, you over sample in areas that are homogenous which tend to be of the same SES, so if we look at our distribution for African Americans they tend to be low income and you have this odd weighting, so you have very high weights for some and low weights, your effective sample size goes down. But it’s the only way to do it efficiently, so you can’t generally do a complete cross tab, you may be able to look at some differentials for some groups but not other groups, so it’s a challenge —

PARTICIPANT: A challenge that could be fixed with bigger samples.

MS. MADANS: That’s right.

DR. STEINWACHS: Anna, do you want to talk some about your perspective?

MS. POKER: I could just give a few updates, methodological updates, but I guess the biggest problem that reports have is data gaps and measure gaps. So any kind of reinforcement for data collection I’m sure would be very helpful for that.

One of the things that we started doing is to make the reports more meaningful, we take that 180 measures and developed the core measures, so right now we have 45 core measures that really we discuss in the reports and also started aggregating measures —

MS. GREENBERG: Are you talking about the quality report or both of them?

MS. POKER: Both of them.

DR. STEINWACHS: Both use the same measures I think —

MS. POKER: I’m sorry, thanks for bringing that up. Yeah, I’m talking about both reports, quality report and the disparities report. And we’re trying to look also from both reports if there’s things that we can aggregate.

We’ve also convened a technical expert panel to help summarize ways that we can measure disparities, because currently the way we’re measuring disparities in the reports is we’re looking for example on race about Hispanics and non-Hispanics and we compare Hispanics with non-Hispanics, or for example for ethnicity we usually compare everything to the whites. Now that isn’t, I mean there’s different ways of doing that. Healthy People 2010 has it by whoever the best performer is and they compare it to that. We can’t use that model because then we’d have a huge denominator so we can’t use that. But one of our goals is for technical expert panels to help us find a standard way for us to measure disparities.

And then another issue that we started is we just convened a group to help us identify standards on disability. Most data tables or databases have different definitions of disabilities and we really would like to record disabilities but it’s just so varied. And I think Ernie really wants, you guys may know more about this, but I think he also wants to include end of life in the disabilities which very personally being very interested in that subject I think that would be really good. I think that’s one of the subjects from a clinical perspective we’re hugely mismanaging and that is costing a lot of big bucks in the nation —

MS. GREENBERG: You’re saying end of life for disabled —

MS. POKER: No, I’m sorry, end of life care, end of life —

DR. STEINWACHS: Quality of end of life care.

MS. GREENBERG: And how did you relate that to disability?

MS. POKER: He’s thinking of capturing —

[Multiple speakers.]

MS. GREENBERG: That may be true too, this whole value proposition —

MS. POKER: Sorry about that, but at any rate, looking at disabilities and getting a standard definition for what is disability, that is what he’s hoping to get from this panel that he’s convening. And then we are going to —

MS. MADANS: Can I ask, there’s a huge amount going on on this right, now has he convened the panel?

MS. POKER: We’ve already started meeting, yeah, we’ve had one meeting.

DR. HOLMES: Are you talking about the technical expert panel?

MS. POKER: No, that’s a different, that’s on disparities. So there’s three issues that I brought forth, one of the things that we’re doing is aggregating summary measures and having core measures. The second issue is we’re trying to summarize definitions of disparities and there’s a technical expert panel that’s going to be helping us summarize that —

DR. HOLMES: We’re on that, that’s the one that Richard Klein and Ken Capple(?) are on —

MS. GREENBERG: Yeah, and we had a workshop here about that. But what about the disability one?

MS. POKER: And then there is another one, disabilities one, where we’re trying to get a standard definition —

MS. MADANS: I understand what you’re trying to do but it’s odd that, there’s something going on in the department, there’s something going on in the statistical community —

DR. HOLMES: Isn’t Barbara Patton(?) involved in your group.

MS. POKER: Yes, I think so, yeah —

MS. GREENBERG: Barbara Altman?

DR. HOLMES: She is part of the group?

MS. POKER: I don’t know some of these names because we only met once —

PARTICIPANT: Did they meet like two weeks ago?

MS. POKER: Yes.

[Multiple speakers.]

MS. POKER: And that’s the goal of that, to have some sort of across federal government what is disability.

MS. GREENBERG: Actually that’s what Barbara brought to this group, remember? We had a conference call with her —

DR. STEINWACHS: And we were discussing, that was one of the things we had on the list.

MS. GREENBERG: She wanted to engage you on that.

DR. STEINWACHS: Engage this committee on disability measurement, interagency working group I think.

MS. GREENBERG: Interagency Subcommittee on Disability Statistics.

MS. MADANS: And because the ACS, the new questions for the American community survey are going to be tested in those sets, so there’s a whole new battery to further confuse —

DR. STEINWACHS: Or to illuminate —

DR. STEUERLE: How do they get on there? Was there a mandate?

MS. MADANS: They were on there before, actually it’s the mandates from VA —

MS. POKER: So an example of a short term recommendation would be to do reinforcement of data collection that may continual, and maybe a long term one on disability, I mean I’m just throwing out ideas.

DR. STEUERLE: Can I ask, I don’t even know quite how to pull this together but I’m trying to think of how the choice on what data to gather and how to fill, you talk about gaps, there’s huge gaps everywhere, how it’s related to something Kevin mentioned briefly and that’s sort of the likely either policy issues that are coming to the fore or the availability of policy solutions. So I’ll give you two examples, I know the current Social Security commission as well as Social Security advisory is heavily going to get into disability and that’s in part because of the continually expanding definition of disability. Mind you there’s disability as defined, as a way to think of a population being disabled but there’s disability for programmatic purposes, the two are highly correlated, consider yourself as disabled for a program you typically will think of yourself as disabled in a survey. So there’s that issue, the main reason they’re addressing it is not only are there these greatly expanding numbers they administratively can’t even deal with it and they have this major issue —

MS. GREENBERG: Who is this?

DR. STEUERLE: I’m talking about Social Security, this is not just elderly, non-elderly, because actually Social Security —

DR. STEINWACHS: Concerned about their insurance fund —

DR. STEUERLE: But it’s not just a cost issue, I mean basically there’s uniform agreement that there’s a total failure in that type of program, is once a lot of people are on disability they almost never get off. Now there’s policy reasons behind it including the fact that if you get off you can lose health care even if you’re capable of work, there’s all sorts of issues why you don’t want to get off. But I know there’s issues there that are major issues and so they’re trying to deal with experiments, possible alternatives for getting people to work, stuff like that. So that’s an example of is there some attempt to try to figure out issues like they’re dealing with, or maybe SR(?) or something, and the way the datasets inform what’s going on there.

And another one, let me give you another example and maybe I’m really switching topics here, is even some of the race income type measures, I think about okay well if I look at the data how is it going to inform me towards trying to solve a problem. And quite honestly knowing that there might be say a 30 percent differential between blacks and whites in care gives me less of a hook on a possible policy solution then knowing there’s a 30 percent differential between the poor and the rich because the difference between the poor and the rich I might do an income condition policy, I doubt seriously I would do a race condition policy, say I’m going to give blacks a bigger voucher then whites. So that might inform, in that case it’s very broad, but it might inform how I think about getting my data so that it could help people, our policy makers when they’re thinking about dealing with these disparities.

So it’s a very broad question but what —

MS. POKER: And I’m probably not the best person to answer it or at least not give you the best answer, you may want to ask Ernie when he comes tomorrow. But one of the attempts that we are trying to do with the report is to make it really meaningful to policy makers and we are trying to aggregate the data in ways that hopefully will be more meaningful for them. I don’t know if it’s not a good idea to start looking at aggregating maybe not just what’s in the quality report, I don’t know if that’s going to be feasible, what you’re talking about is not everything that we capture, certainly Social Security data is not something that we have. Are you kind of suggesting that maybe taking the quality and the disparities report data and maybe aggregating it over different types of datasets?

DR. STEUERLE: I suppose it’s a little different but thinking when we think about what we’re going to survey and the data we need to collect, it’s awful helpful to know what questions we want to answer with that data first, and not only what we want to answer but maybe what questions we can answer, maybe there are a lot of things we want to answer but in fact, this is all cost/benefit analysis, it’s implicit in everything else, but it’s just a different way of sort of addressing the issue.

MS. MCCALL: I think it’s a key question but we’re going to try, it has to do with the general issue of I guess how policy both informs and is informed by findings of anything but in particular if you think that there’s an issue, let’s say in this case on disability, how does the knowledge of the thought that there’s an issue inform work that’s being done and then how is the work that’s been done and the findings of that go back and then kind of re-inform policy. You could say that same thing around anything that we’ve been discussing or will discuss, could be quality metrics, there could be scientific discoveries, it could be new taxonomies that we would be talking about —

DR. VIGILANTE: Well it’s part of my hypothesis creation, it’s sort of this dynamic between what you study and how you study it.

MS. MCCALL: Right, and so I think it’s a clear concept but when you try to scale that and say how the heck do we do that across so many different dimensions and start thinking about how either the whole infrastructure around health information technology, around I hate to say it but secondary uses and all the policy that we set around that to just even begin to enable there’s another mechanism that needs to exist in order to really kick that concept into high gear. The reason I say high gear is that we will, when we realize our dream if we don’t have a mechanism for doing this we’re going to drown in the amount of information that we have. And so —

DR. STEINWACHS: Or at least in the amount of data.

MS. MCCALL: Okay, in the amount of data that we have, I think that actually that’s a very important distinction, I think that we will —

PARTICIPANT: It will be a new form of disability, we have a data disability —

MS. MCCALL: We will, and so we have to begin to literally architect and engineer ways to explicitly deal with it in ways that are good as opposed to just kind of letting it fall all over.

DR. STEINWACHS: I think partly there’s a set of issues that keep coming back into the policy arena, some of them keep cycling, quality is high these days, it disappeared, it was high a long time ago, cost, finance, the resource issues, I think part of this probably is a trick which is with the amount of surveys you’re trying to do is trying to figure out what are those recurrent issues and what are the key classification variables that you need and then I think HIS needs to reflect in that effort to try and add on modules which are more one time, can be recurrent. It seems to me part of that which is a key question is how do you make this really valuable —

MS. MCCALL: And I don’t know where that issue goes, it doesn’t necessarily fall into any one group that you maybe want on the subcommittee —

[Multiple speakers.]

DR. BREEN: I think the architecture is something that this committee should definitely take on because it’s not something that happens at DHHS.

PARTICIPANT: It probably doesn’t happen across agencies, even less so —

DR. BREEN: No, it doesn’t, Social Security is a good example, it’s no longer part of DHHS.

MS. MCCALL: There’s an issue out there waiting for all of us and is it that something that we do want to take on, if so where does it naturally flow, and maybe it doesn’t —

PARTICIPANT: The issue is what again?

MS. POKER: The issue, you just defined informatics sort of, it’s taking data and turning it into knowledge and how does that process flow and in order to get there what are the kind of questions we need to ask. Because right now the reports really collect the data that’s there, we don’t have a choice, I mean whatever national data there is that’s what —

MS. MCCALL: So while there’s certainly existing mechanisms today, so it’s not to say that this doesn’t happen, it does happen, but we are going to go into I think more then just a single order of magnitude, probably multiple orders of magnitude of change in quality and quantity of the type of things that we’re going to be dealing with and we literally will be in just a wash in all of it. And there’s not going to be any dearth of things that we’re going to want to know and things that we’re going to design that we say we want to know. But we have to manage that in order to get what we seek and I don’t hear that the existing mechanisms will suffice.

DR. HOLMES: Well, I think we should remember too that one of the overriding goals of both reports is to guide quality improvement, that is these national estimates for, other estimates provide benchmarks against which health plans or states or something can gauge how they’re doing, individual providers can look at national benchmarks in terms of how they’re doing with treating their patients. And then the other side of it is if there are differences, race, by race, SES, in terms of the quality of care provided or access to care, what policies can we enact to reduce those disparities. And what you had suggested too I think is very important, that one of the issues, important issues about getting better measures or measurements of race ethnicity over against SES is to determine what is the real underlying reasons for differences, is it race or is it SES. And whichever the answer is in whichever occasion would call for different policy responses obviously.

MS. POKER: Just to add to what you said, Julia, is the reports on a trend, to capture trends and benchmarking yes but hopefully they can also be used for implementation, how we can work better. And sometimes that’s where I think what Carol is saying is really coming into, how do we do that because then we have to change the kind of information we gather, the kind of information we would like. That’s what I hear at least from you, so that seemed to change.

MS. MCCALL: I was just kind of going more toward the, that’s I think even a third leg, one of them is we have a question, a hypothesis, and how do we get that hypothesis into this engine that’s going to do the research and then how do we make sure that all that happens. And it’s not just one question, one answer coming back to one audience, it’s the fact that there’s so many different people. You said, it was beautiful earlier, so there’s so much work going on in disabilities how do you know and so you go wow, take that and multiply it a hundred fold or maybe a thousand fold with all of the new information that will hopefully be coming online and then deal with that. Separate in this thing from once we have the answer how would we go about implementing it, I would hope that some of the information that we would gather would actually lend toward okay, here’s how we might go about it once we think we have a solution. But not necessarily because it could imply different data, if you have data that you know won’t answer it doesn’t mean that you got the data that will tell you how to fix it, it’s very different —

MS. POKER: And implementation is, could be different, I mean I might need to do different things —

MS. MCCALL: Absolutely, and you may not have captured it.

MS. MADANS: When the big national datasets, the surveys, not the administrative records which we think we deal with very differently, and then of course there’s a linkage which I think we really try to do to the extent that we can but we probably could do more. When those file datasets were originally designed they really were designed to try to capture the core indicators, the most important thing you’d want to know about it. And if we’re doing something in health we really would like to have datasets that capture the range, times change and you do modify the datasets but there’s always more to add rather then things to take out so I find it somewhat hard to believe that the datasets that we’ve known and loved all these years really are not capturing a large part of what you would want to know if you were going to sit down and do a quality, not all of it for sure.

But there are going to be gaps and so I guess one key is really what are the really the most essential gaps that we think we can change without overburdening the system in terms of content and is there anything you’re going to take off because there is a set amount of time. Or do we really need to kind of start all over again, if we went through the exercise of saying we’re going to do the report and what would we really want to put into it, and if we come up with very different sets from what we kind of have I’d be really upset so I’m hoping that wouldn’t happen.

But I think there is the gap and we want to kind of meet it to the extent that we can and then kind of the perpendicular thing is the sample, there’s content and sample and what do we need to do to get, to target the sample and we had a meeting in this building last week about getting race and ethnicity data on small groups and I have to say at the end of that meeting it was like we’re not going to do it, it’s too expensive. It’s not rocket science how to do it, it’s rocket science how to pay for it because of the way we go to people right now where they live and the way they’re distributed where they live makes it very hard to get information on any group that’s probably less then five percent of the population, it’s just very expensive.

MR. HUNGATE: I wonder if there’s an alternate way of going about that, populations like that also self identify as having special needs in some ways and if there were some way to take the techniques that are involved and make them available through foundation funding, to not use the classic methodologies of getting the data but testing the extremes of other avenues to gain the same data which then we would — [off microphone] —

MS. MADANS: Yeah, and there is talk about, and there have been examples, especially at a local level where you take the data, you get all the procedures and you give it to the groups. There’s always the assumption that they somehow could get the money, they’re unable to get the money and then it’s not cheap, aggregating up is a real problem. Maybe you don’t need to aggregate up, maybe you can deal with the larger database, it’s perfectly done, nationally representative, but isn’t deep, and then figure out a way to use this other information, not to combine it but to use it as aggregate. But even the basic data collection, everybody we talked to said well we can give you all the mechanisms, okay, but now I need $5 million dollars to do the survey.

MS. POKER: But isn’t there an ideal for data? Isn’t there if you get too granular then don’t you lose the robustness of that data because then you just don’t have enough data? For example if you’re digging out the Indian population, Indian Americans, to too many sectors, you’re not going to have enough data to make it meaningful, I mean you guys know that stuff, is that an issue? Because too much granularity can be a problem so what is the ideal amount that it’s meaningful and then can impact the quality of health care but it doesn’t get to the level of granularity that we’re just counterproductive.

DR. VIGILANTE: Are we talking about this in the context of what this committee ought to be focusing on?

Agenda Item: Follow-up Discussions on Subcommittee’s Work Plan – Subcommittee

DR. STEINWACHS: Yeah, let me just take one step back and then let’s talk about where we go, so I don’t know whether we have to, it’s 4:45, on task. The other backdrop here which we haven’t said much about, I understand its both reports, you’re using the IOM recommendations of the 22 conditions and so a lot of these are disease specific measures and then there’s some crosscutting measures —

MS. POKER: We might be switching over, there is that possibility they’re going to take the 20 IOM priority conditions, that is still something that, we just don’t have the data for it yet —

DR. HOLMES: The current reports don’t use the IOM priority conditions, they overlap to a certain extent but they don’t use all of them.

DR. STEINWACHS: Because that also could be an area for discussion here. Part of the roadmap if you were thinking about what the Quality Workgroup is doing, looking at specific conditions using more clinical data versus a national report that right now I think what you’re saying is in many cases you’re aggregating up probably and not much specificity, what’s the value in trying to get the numbers, I mean again this probably comes to the number issue in specificity —

MS. MCCALL: Actually I think that that’s a great topic because it came up in another context, which is that there are going to be some areas where the light is going to be bright, okay, it could be a condition, it could be a setting, it could be a type of data, there was a meeting earlier that talked about pharmacy data, and so when you think about trying to, think about all the issues where there’s gaps or thin data or sample issues or fine grain and say we’re not going to be able to do it all at once but if we could find just a delightful overlap on a couple of key things and we know it won’t be everything but because it’s not everything we can prove some, we can build some mechanisms that are maybe a little bit smaller and learn some new things and maybe different ways of surveying or maker it cheaper and getting it finer, and find those overlaps across a couple of key ideas.

It could be quality and population and electronic prescribing or whatever it is with respect to infrastructure and architecture, get some additional data elements captured in a different way, some standard taxonomies, find that river running through it, running through all of our work across NCVHS and see if we can make a couple of deep but more narrow things happen. I just think that there’s a real opportunity there.

DR. STEINWACHS: Other suggestions? I was just going to say that it seemed that we didn’t have a lot of time here and I was hoping that Ernie Moye(?) tomorrow is going to be talking about the survey too to the whole committee, I guess the National Quality Report and Disparities, and some of the issues there. I was hoping to schedule a conference call of the committee in a couple weeks, probably it’s as long away as it takes to find enough people who can meet at one time on the telephone, and try to talk about where do we try and focus our attention and making some choices. Because I think we need to have bubble up two, three, four ideas as do we think promising areas to focus attention on and then take the next steps to see if that really both valuable, and I think this group is looking for customers for what we do, so there ought to be agencies who have policy interests —

DR. STEUERLE: I think there may be a prior question, I really think we have to ask the question of process, I mean I’m just trying to think of what processes we want to set up for gathering, maybe what processes we would set up for gathering information ourselves and be, what I still, I think we need to define what our role is going to be.

I’ll give you an example, we start off talking about populations, we often talk about, automatically I start thinking about big datasets and can I measure racial and socioeconomic disparities. But I’m not sure for me that answers a lot of questions whereas I can think of a 100 little questions where a survey of 100 people or 200 people, so we have the same debate where I used to advice the statistical income division of the IRS, do we get a much bigger sample so we get much better refinement on income tax or do we go off and measure something like charitable contributions and nobody has even one off and measured the data.

And I think we’ve got a lot of issues like that here, just stuff, I mean everybody I’m sure in the back of their mind has issues that are just really high that are on their mind, they’d love to be able to do something about, in some cases it might be nothing more then a very small survey, in some cases it might be do we have the ability to attach a tiny sample to a basic survey that’s already out there to answer that question. I don’t know, you’re saying what are the one or two or three questions we have, I don’t, I worry about this question of gaps and what’s not being done and it seems to me that’s where we have a role is where we can help more then anything else is identifying what’s sort of not being done or not being done well, not just because of aggregate size of resources —

DR. VIGILANTE: But that could be one of those two or three, I mean that could be in that short list of things that —

MS. MCCALL: Right, but actually finding where those gaps are, finding a process, a process to find them, so we have to decide how we’re going to decide. And I think that would be great because it’s really kind of a gathering in of all the things that it could be based on work that’s already been done and disparities that we know are there or have very strong reason to believe they’re there —

DR. STEUERLE: But I’d still like to know what exactly, I mean maybe you know this but what precisely, is our role of advantage as a committee in this vast penalopy of the health care sector, every time I get in it I’m just amazed at how big and vast and in some ways rich it is despite all our complaints, what is our relative advantage in getting in there and making a difference. And I don’t have, I’ve been on this committee now almost two years and I still haven’t figured that out.

DR. STEINWACHS: Well, if I figured it out we wouldn’t have a long discussion about it —

PARTICIPANT: When you leave they tell you —

DR. STEINWACHS: It seems to me there’s been a traditional role of the Population Subcommittee and that has been a very close set of advisory relationships with NCHS and other population large survey activities and so you say where is that, so some of the questions that you’re raising about where are the gaps in producing national population information, which could also come out of vital statistics, not just surveys, so your gap question, are there important gaps that we ought to be trying to fill, I think some of the metrics issues of are we really capturing well the variables we need to capture to make the classification so you can interpret it. The race and socioeconomic status it seems to me becomes a rather crucial issue in some of the quality debates and the point you made about solutions may not lie in the health care domain.

The other thing that we had sort of talked about which is wedded in the kind of firm agreement was that if we were trying to do something where the Population Subcommittee was complementing what was going on in the Quality Workgroup, and you’re not on both of those but a lot of the people here are, is one of those might be to be, and that was where I think we got to this agenda item on looking at population quality, might be to try and do some things that really probe more into the value, the interconnection between quality measures that come out of clinical settings and not at a population level in terms of talking about it versus those that you have the population level. Julia was saying well population level data is a benchmark, is that how you sort of measure against how you’re doing in clinical settings, is there some value in having something that says that the population level, this is what we see in quality and outcomes —

MS. MCCALL: Well, I think that and bringing in things that are not of the health care settings but are other variables that may say look, the answer lies outside this bright light of specific clinical measures, the answer is actually a root cause that’s way far away. So I think that that’s another —

DR. STEINWACHS: Exposure, environment, all those things that go into the health model.

MS. MCCALL: But it still begs the question is, because that work needs to be done, the question is what is our role, this group, to that work, and that’s what I hear you asking.

MR. HUNGATE: One of the questions that I keep asking myself is we talked a little bit about SES data here and for my understanding of health care outcomes, outcomes are ten percent off for certain levels of socioeconomic status, so if you’re going to judge a health care facility for its own quality you’ve got to be able to risk adjust for the socioeconomic status of that population served by that institution for procedures. So there’s an interlock between what happens with individuals and how its judged on a total basis and I think somehow that’s a piece of the crosswalk but I don’t know how that works out in a data sense.

MS. GREENBERG: It just is always an issue of whether you want to use those factors to risk adjust or you just want to stratify them and make the comparisons, because you can sort of risk adjust away potentially real disparities.

MR. HUNGATE: Well if I want to compare Boston City to Boston Mass General in Boston qualitatively, if I don’t make that adjustment I’ll make errors in judgment.

DR. STEINWACHS: At the same time you wouldn’t want to set quality standards for Medicaid recipients that are different then for Blue Cross/Blue Shield recipients even though the experience may be that Medicaid is not getting the same quality of care.

Before we have to vacate let’s talk about process, so it seems to me that in trying to address what this committee wants to take on and in great part I think we have a fair amount of latitude other then we have a title, and we can change our charge, we haven’t drawn that back out but essentially deals with population level health statistics I think is a large part of that charge. What’s a useful process, I think where we got to here in this was sort of a polling of people and sort of bubbling up this idea that maybe population quality, population outcomes measures, and thinking of that as an agenda where there was some synergism between that and the Quality Workgroup which is looking much more at the IT quality interface, EHR, and other IT sources feeding into quality.

The other certainly and I think we ought to be open to that is to say there are possibilities, it doesn’t have to be the population quality, but I do think we want to find some things where we think we can make some recommendations and get some useful information out in the short term and there may be a longer term part of it which in thinking through next steps and visions and so just like you were talking Gene about a process, could be a kind of visionary kind of thinking about what is the ongoing process that needs to go on but look at the health statistics function to say is it really answering, how does it collect the right questions and then try and process those. So to me that’s a kind of next step vision that at least I don’t have that would be useful.

DR. VIGILANTE: And the other thing is, and maybe this is, maybe when we think about having testimony at some point when we do that it might be sort of taking people who are dealing with difficult questions outside the area of health, people who do educational research, people who are doing sort of issues around crime and prevention, people who may be doing things around asset development to help socioeconomically depressed populations, I mean I think there’s a variety of folks out there who are probably saying God if I could only link this to that I could really, and if we get, so it goes back to what you said, if you got a couple of, got a convergence around a couple of these things where you did get that overlap and you heard from other disciplines that might be a worthwhile quest to go on.

MS. MCCALL: I think it would, and even a precursor set of work to that would be again to know who to invite into the room, where do we think, what are the strong theories that are out there, what are people grappling with, where are the big disparities, just where are we shown the light and seeing issues and then bring those folks in —

DR. VIGILANTE: We know there are employment disparities, we know there are incarceration disparities, we know there are health disparities, education disparities, I mean there are a list and there are huge overlaps in the populations that are afflicted with these and there are people who spend their lives working in that domain and trying to answer those questions and understanding where the overlaps, what the wish list is —

MS. MCCALL: What the wish list is, use that then to craft a picture, a vision if you will —

DR. VIGILANTE: Literal picture of geocoding —

MS. MCCALL: And including that, so craft a picture and use that to then ask the question how can this new world that’s coming, I don’t want to forget today, all right, because there may be some short term things in there as well, but how can this new world begin to create some of this convergence so that it is designed into its very fabric, my biggest fear is that it won’t be designed into the fabric, everybody’s going to assume that it will naturally occur, and it won’t because we made this gross and inaccurate assumption that well it’s just going to happen. And I don’t believe that it will unless it’s literally designed in, especially when you talk about convergence like we’re talking about.

DR. STEUERLE: Does this mean that we need to call in people, either health researchers or researchers who do other research for the health data are important, and call them in and ask them about, not just about what research they’re doing but where they really see, I would just say gaps but opportunities, I mean for me one big one is still the lack of ability to access administrative data and merge it which is sort of where we pay this huge cost to gather the data but then none of us can use it —

MS. MCCALL: I don’t want to hear only about their barriers to doing what they try to do today, I want to hear about where they think that there are opportunities and gaps in care or health or population health —

DR. STEUERLE: And I think we want to know actually what they see as the bureaucratic hurdles which are often as important as the natural ones, if the data is another agency I can’t get to it, maybe we can’t do anything about that but I’d at least like to start with some, it’s a little more of a white board approach to sort of getting some, then we can narrow it back down to sort of where —

MS. POKER: I’d like to make just a quick suggestions, just a thought, that if you’re looking from process, another way to look at it is also what is the biggest bang for the buck and one of the things that cuts across the nation and every human being is going to have to deal with these end of life issues. And that is one of the things, it’s not sexy and we’re really mismanaging that, and there’s two implications of it, not only the quality of care that people get at end of life that’s very poor but also the caretakers of, who are women, it’s a minority issue, and black women more so then white women actually go bankrupt through that process. There’s a rate of over 70 percent of black women who go bankrupt doing caretaking. So I mean I think there’d be a great business case to look at something, also what is the biggest need, the biggest gap that we are not addressing as a nation. So that’s just another process way of looking at things too.

MR. HITCHCOCK: That’s a policy question, not just a research question, which is good, I don’t think this group needs to get, there’s a distinction here, policy research and maybe research in general —

[Multiple speakers.]

MS. GREENBERG: Well, I’m not sure if this is consistent with, it seems a little more consistent maybe with what Anna said but this tension between doing something kind of narrowly that you can accomplish or make a difference and I think, I mean Carol is sort of talking about seeing what windows of opportunity are there, but not just going, and Anna is saying, you want it to be important, I mean there’s some things you can do but who cares, or it impacts on so few people —

DR. STEINWACHS: That must be the academic issues.

MS. GREENBERG: We’ll leave those to the academics, right, they can get their PhDs on it. But I just don’t want you to forget about and Barbara’s request to you because you also would like something for which there is a customer. I’ve said this before about the committee, you want to meet the needs of the department, who’s your main customer, but at the same time you don’t, you want to be ahead of them too or not just completely driven by that. So you have to get a balance in that. If there is this effort going to try to get more standardized definitions of disability and it’s obviously, I hadn’t even realized that it was going on with the quality reports but it’s going on as Jennifer said in many venues, and they’re looking for kind of a group that serves in an advisory capacity, could get input beyond the people they’re currently working with, and the committee has been asked if you’d help with that.

You know that, at least there’s someone who’s interested in your helping with it and it does tie in with some of the work you’ve done in the past and related to the work on disparities and all of that too. So I just, just because Barbara isn’t here I want to put in a good word for her request —

MR. HUNGATE: It’s a very health related thing, it does focus in that specific —

DR. STEUERLE: I take it by the way, we don’t have time to discuss it I guess but I take it things like all the health data that’s all sort of quickly going to be developed or that is available or that’s opportunity out of Katrina is something that CDC and other people are solving, it’s not something for us to worry about —

[Multiple speakers.]

DR. STEUERLE: It’s an example of something we’re not, we’ve cut off, and I’m just wondering if that has been explicit or just —

DR. STEINWACHS: We talked a little bit in the full meeting for a couple moments about what some of the issues are in trying to assess the impact on health of a Katrina and certainly the country’s good at measuring immediate deaths and some other things but the broader impact of the population —

DR. STEUERLE: I’m just even thinking of a health survey of 500 people for whatever, I can think of several questions, but CDC, I talked to Steve, apparently some people are already starting to bounce this around.

DR. STEINWACHS: Well what happened was in the room then people representing the different agencies responded about what each of the agencies are doing, I think of then it’s the kind of question of is there something that we would want to do and probably if you said what could we do in the kind of way we operate one would be to say well we could try and look at the situation and learn what’s being done and try and assess how adequate it is, what’s rolling out over the next six months or so —

MR. HITCHCOCK: CDC routinely does surveys on, that makes them disaster areas, they have sort of a blanket clearance from OMB where they can actually just go right into the field and explain to OMB after the fact what they did but there will be surveys going on.

DR. STEINWACHS: But all I wanted to say was it seems to me that could be the topic we want to pursue too and so I wouldn’t throw it out but I think it’s hard to be proactive on it now, I think now you’d really be saying what’s being done, how well is it being done, does it really address the important policy issues and so the next time there is they’ll be other strategies here, and to what extent is it coordinated, and I think each of the different agencies in the department seems to be taking on pieces that they see appropriate to themselves, I don’t have a sense that that’s a particularly coordinated strategy —

MS. MCCALL: One thing we could do, when the work is essentially over, because they’re going to do what they’re going to pursue right now, is to actually case study it and say okay, this is an example of kind of integration or knowledge transfer that will exist, it exists everywhere today but it’s really going to be increased by an order of magnitude, and to say okay, understanding what everybody found out, is there a different way to integrate that going forward or just enhance knowledge transfer. Not to say there’s going to be another Katrina but there’s always going to be some big pressing issue where A, B, C, and D are going to at something.

MS. MADANS: If you look at what happened after 9/11 where it was actually easier to do the data collection because it was concentrated very little came out of it, all of a sudden a lot of little surveys were done and there were a couple things published and most of them said well we don’t really understand how to interpret any of this. And I think that it would have been useful to actually look at that, most of the stuff that was useful I think was, I think they went into the first responders, some of that data I think is, some of the building stuff, but it was very, this is going to be so much harder to do because they’ve spread all these people around and we won’t be able to find them.

MS. GREENBERG: Maybe if you’re interested in pursuing this maybe you could get Jim to participate on your conference calls —

MS. MCCALL: I’m just trying to call the meta question, at some point I think it would be great to have a case study that’s very crisp and clear that everybody was focused on it and just shine a big light because it picks up on some of the things —

DR. STEUERLE: Jim is the type of person we would call in and say okay, do we have a role here you’re not able to do. And maybe we don’t.

DR. STEINWACHS: Well, let’s see if we can get Jim on the phone.

Okay, so we’re going to schedule a call, Audrey and I are going to try and summarize this discussion, you may not recognize the summary but we’ll do the best we can. And I want to thank Julia and Irma and Anna for coming and joining us on this, and seeing if we can make some decisions and next steps. Thank you very much.

[Whereupon at 5:00 p.m. the meeting was adjourned.]